Figures: the Art of Science

How to prepare figures to make an impression on editors, reviewers, and readers.

Figures are, perhaps not quite literally, worth a thousand words, but they are invaluable: try explaining in text all the details in anything but the most basic image. Yet a reasonable fraction of manuscripts submitted to Journal of Biogeography (JBI) contain too few or too many figures, figures that are difficult to interpret, figures that are poor quality, incorrectly sized, that don’t conform to JBI’s style requirements, are formulaic, and/or don’t convey the research in a compelling way. These are all things that can slow down review or publication of a manuscript or increase the likelihood it will be rejected, especially if there are also shortcomings in the text, so best to avoid them if at all possible.

By contrast, an appropriate number of nicely prepared, easily interpretable, information rich figures will emphasize the positive and can to some extent compensate for shortcomings elsewhere during review, because a series of good figures can alone tell most of the story in a paper, and reviewers and editors consider manuscripts holistically. (This is not to say other things will be ignored; to ensure the best possible chance of publication, all aspects of a paper should be prepared to publishable standard before submission.)

Moreover, creative and thoughtful illustrations will make your paper stand out from the crowd during review, when being considered for the Editors’ Choice or Cover Article, and when readers browse the journal or find your paper among tens of others in a literature search. As such, consistent with this column’s intention to help authors’ manuscripts be more successful at JBI, we here provide a primer on preparing figures for successful review and publication in JBI. This is intended as a quick reference guide to some key ‘good practices’ in preparing your figures and so it emphasizes aspects of, but does not replace, information provided in the journal’s author guidelines and Wiley instructions. We hope this is helpful!

1. Include a conceptual or synthesizing figure
We are all familiar with using the Introduction of a paper to establish the background for our study, the questions that background raises or leaves unanswered, and so the hypotheses our study is going to test. But far less often do we also present that information in a powerful visual graphic. Yet such graphics can quickly provide readers with an holistic view of your study and a visual expectation of what your data would look like if it was consistent with or refuted your hypotheses, which they can then pattern-match to your actual results. This approach can therefore also guide how you set up your Results and can simplify your Discussion. In a way, it’s really an element of good study design, showing the study has been thought all the way through from the start. An alternative (or addition), if you wish to provide a powerful take-home message that others will use in their talks and redraw in their papers, is to provide an integrative synthetic figure emphasizing the key points in the Discussion. Similarly, it has been noted that if your methods are complex, it can help to have a (numbered) flow-diagram which can both clarify and shorten your textual description..

2. Figure size and proportions
Create your figures at the size and proportions they’ll be printed. Figures are reproduced in the final article in three basic sizes: single column width (79 mm), two-thirds page width (110 mm), and full page width (168 mm). If a figure is designed for one of these sizes it’s usually obvious and the production office will make it look pretty. If it is not designed for one of these sizes, it will be shrunk or expanded leading to images with detail that is hard to read — in production, figures will be smaller than R output, so letters have to be proportionally bigger, lines thicker, and points larger to be read appropriately — or an overly simplistic and large kindergarten book kind of feel.

All graphics software these days allows sizes to be set and/or images scaled, so when designing figures, keep the final sizes in mind, and imagine what they’ll look like on the ‘printed’ page, either one full-page or fit-to-width view online or in a PDF; if you’re relying on people zooming in/out, you probably need another panel or figure. If you’ve multiple panels, consider how to arrange them so they fit down one column, or across the page, or in an # rows x # columns composite figure; if your panels are graphs and they have an axis in common, use that as your guide. Try to avoid arrangements or sizing that will lead to a block of empty ‘white space’ around the figure. Make sure to scale the text on your page to fit the final printed size (ideally use 8 pt after any reduction), don’t have too many different point sizes in one figure, and try to keep text size consistent across figures in your manuscript. N.B. The maximum height of figure is 230mm.

3. Resolution
Use vector graphics, if at all possible, which will ensure your image looks sharp at either one full-page or fit-to-width view on the screen and everything in-between and beyond. Vector graphics should be in .eps or .pdf format. If using photographic images save them in .tif format at 300 d.p.i. If figures combine line art and photographic images and cannot be saved in vector graphics format, use .tif format at high resolution (i.e. 600–800 d.p.i.).

4. Figure captions
Captions, a.k.a. legends, should be concise, comprehensive, and ‘stand-alone’ – i.e. the figure and its caption must be understandable without reference to the text. To this end both the geographical region/s and the taxon/taxa should be mentioned in each caption. Include definitions of any symbols, abbreviations, and units of measurement.

5. Colour
Use colour wisely. As colour images are free in the PDF and online (which is where everyone is going to read them) it may be tempting to use colour often and sometimes a lot of it. But unless colour actively increases readability of your figure (i.e. in figures with many components to be compared, or when >2 dimensions are needed) think twice about using it. Also, avoid heavily coloured base maps (e.g. captured from Google Earth) are a poor background on which to plot important data because they have a lot of superfluous information and don’t allow for sufficient contrast. Often it can be equally or more effective to combine different types of symbols (e.g. solid and dash lines or areas, or round versus square points) than adding many colors. This kind of presentation also works on a grey-scale printed page and for the many people who have different perceptions of colour. Colour figures should be saved in CYMK

6. Maps
(a) Include a map that provides regional context as well as details of your study. JBI has a global readership so many people might not immediately recognize your study area without broader context. Provide details of your study in your main figure and consider also an inset or additional panel(s) showing country-or-higher level context. In each panel include a scale bar in ‘km’ within the figure.
(b) Continental and global maps should usually use an equal area projection. All biogeography addresses life on an approximately spherical planet, but we draw them on a flat page. Much biogeography addresses processes influenced by area and/or distance, so patterns are best represented on a map that reproduces these attributes with as much fidelity as possible. For area, Mollweide or Aitoff’s projections are recommended choices. If distance is more important than area in your study perhaps choose an equidistant map instead (one pertinent to your area of study). If dealing with area and distance, consider different projections for different maps within the same study or a ‘compromise’ map. Regardless, maps based on, e.g. the Mercator projection provide misleading visual cues close-to or far-away from the equator and should not be used. Given the variety of maps and that it can be difficult to distinguish between some visually, it is important to state the projection employed and the reason (e.g. “equidistant” or “equal area”) in the figure caption. Two lists and descriptions of projections are available at https://en.wikipedia.org/wiki/List_of_map_projections and https://lib.bsu.edu/collections/gcmc/tutorials/pdfs/mapscartographymapprojections.pdf.
(c) Show the range of the species with which you’re working as well as the extent sampled if these are not the same. There is good reason to expect that biogeographic processes vary spatially and in regard to the relative position within a species range (e.g. periphery versus center, ‘leading’ versus ‘trailing’ edge), so to the extent it may be relevant to the goals of a study, sampling location(s) should be shown relative to the species’ range.
(d) An additional consideration. It can be more convenient to summarize sample sizes and other relevant background information in a map than in the Methods text; too often papers do neither and bury this important information in supplementary documentation.

7. Phylogenies
Must include edge/node support values. A list of names in the leaves is not always most helpful for readers. Consider complementing the basic tree with clade-level annotation, images, maps, and traits.

8. A few final tips from the Production Office
Ensure (i) all figures are cited, in order, in the main text of your article; (ii) each individual figure file less than 10 MB, and if not then remove excess white space surrounding figures for smaller file sizes; (iii) figure files are named with their figure number. A checklist for basic figure requirements is here for initial peer review as well as production.

A few examples:

Effective single column black and white figures https://onlinelibrary.wiley.com/doi/epdf/10.1111/jbi.13830

Nice map showing global location, regional sampling, and detailed study design https://onlinelibrary.wiley.com/doi/epdf/10.1111/jbi.13801

Good use of a conceptual figure https://onlinelibrary.wiley.com/doi/epdf/10.1111/jbi.13731

Effective arrangement in 1r * 3c to facilitate comparison of species richness; also appropriate use for colour (Fig. 2) https://onlinelibrary.wiley.com/doi/epdf/10.1111/jbi.13753

A variety of stacked, 2/3rds width, and multi-panel figures that also support adjacent placement of captions https://onlinelibrary.wiley.com/doi/epdf/10.1111/jbi.13755

Some additional resources:

McInerny et al. 2014. Information visualisation for science and policy: engaging users and avoiding bias. Trends in Ecology & Evolution, 29:148-157