Remember that figures are often the first thing that readers look atThe term ‘figures’, as used by biomedical journals, may include:
• graphs (e.g. line graphs, scatter plots)
• charts (e.g. vertical or horizontal bar graphs, pie charts)
• photographs (black and white or colour)
• micrographs (from electron microscopy)
• biological recordings (e.g. from ECG or EEG investigations)
• biological images (e.g. from CT, MRI or PET scanning)
• electropherograms (from automated DNA sequencing)
• electrophoretograms (photographs of gel-based electrophoresis)
• line drawings (e.g. flow charts or surgical procedures).
Photographs have their own separate list of tips (Chapter 17).
Remember that figures are often the first thing that readers look at
Readers (including reviewers) often look at figures first to make a preliminary assessment of the results, so they have to make a good first impression.
Avoid unnecessary figures
Do not use a figure if the data could be represented by a simple sentence or two. For example, a bar chart with only two bars may be acceptable as a slide, but if you included the same chart in a paper, many journals would reject it on the grounds of space. They would argue that the same data could have been more concisely represented in the text.
Be prepared to be selective
Even if you have half-a-dozen relevant, interesting figures, the journal may not be prepared to publish them all. Some journals set limits on the number of figures they will accept, others leave it to your discretion. As a rough guide, one figure or table per 1000 words of text will probably be accepted without quibbling. As mentioned in Chapter 14, you may be able to include extra figures in the online version of your paper.
Do not expect the journal to publish the same data in both figure and table form …
Submit your data in whichever form is most appropriate.
… But be prepared to submit backup data for figures
Many journals now request that figures be backed up with tables showing the raw data used to construct them. This is to save reviewers the trouble of trying to establish values for data points that are just represented as spots on a graph. Just because you have to supply the raw data, however, does not mean that it will be published. Look at the Instructions to Authors to see whether this requirement applies.
Provide figures in the format requested by the journal
Often on initial submission of the article and certainly after acceptance, figures (without legends) must be supplied in a specified format and as a separate file to the article text file. This is because they may be incorporated directly into the journal without redrawing, or will go through separate production processes. Check with the Instructions to Authors, as journals differ in their requirements. When combining multiple graphs or charts on one figure check what nomenclature is acceptable (e.g. Figure 1 a, b, c, d).
Do not try to fit too much information onto one graph
Think about how much information the readers will be able to take in. For example, line graphs with more than four lines are likely to be too difficult to read, especially if the lines overlap.
Make sure the lettering is large enough
Think about how much your figure will have to be reduced to fit into a single column. Then work out what the reduced height of the letters will be. Usually, the reduced size of a lower-case letter without ascenders or descenders should be no less than 2 mm, or it will be too difficult to read.
Make sure that symbols will still be distinguishable after reproduction
Remember that it may be difficult to tell the difference between microscopically small circles, squares or triangles. Make sure that they will be at least 2 mm high after reduction. The standard symbols used by most journals are circles (
/
), usually followed by triangles and squares.
Check where keys to symbols should be placed
The Instructions to Authors usually include advice on where to put the keys to symbols used on your graphs. Usually, the key is given on the graph itself, but a very few journals put them with the figure legend.
Emphasise the data, not the axes
The thickest lines should be used for curves or plots, thinner lines for axes and error bars. Most scientific graphics programs will take care of this for you.
Do not extend the axes too far
The X and Y axes should extend only to the next ‘tick mark’ on the axis after the maximum values for the data. Again, your graphics program will probably take care of this for you.
Indicate significant differences between points with asterisks
It is standard practice to draw attention to significant differences on graphs with asterisks, the meanings of which are defined in footnotes. A standard series is *p < 0.05, **p < 0.01, ***p < 0.001, but check with the journal.
The figure title and description – the legend – are usually placed below the figure
Each figure should have a legend (or caption) consisting of a title, description and definition of abbreviations/symbols if needed (see below). The legend should usually be placed below the figure. Legends are usually supplied separately from the figures themselves – check submission instructions carefully.
The figure and legend should be self-explanatory
The figure and legend together should be understandable without reference to the text.
Define abbreviations if used
Usually, abbreviations used in figures should be defined in the figure legend, even if you have already defined them in the text, but check with the Instructions to Authors. If you use the same abbreviations in several figures, it is common practice to define them in the first one and then cross refer, e.g. ‘For abbreviations, see Figure 1’, or ‘Abbreviations as in Figure 1’.
Make sure your figure is referred to in the text
Check that for every figure, there is a reference in the text (‘Figure 1’ or ‘Fig. 1’), and vice versa. You should also indicate approximately where the figure should appear, by a note in the text (clearly distinguished from the text to be printed).
If you use someone else’s figure, it is your responsibility to get written permission
If you want to reproduce a figure first published by another author in an original publication, you will have to get permission in writing. It is your responsibility to do this, not the publisher’s. You should write to the managing editor of the journal in which the figure first appeared, asking permission to reproduce it. Say where you plan to reuse it. The journal will usually respond giving you permission, and stating a set form of words to be used for acknowledgement, e.g. ‘Reproduced with permission from …’ Often, you will be asked to write to the original author as well. A few journals may ask for payment for reproduction of figures, and some impose rules on whether they can be redrawn or adapted in any way.
Apply this reviewer’s checklist to your figures
Do the figures:
✓ show data in the most appropriate and efficient way?
✓ show important data?
✓ show what the text says they show?
✓ with their legends provide sufficient information without needing to refer to the text?
✓ explain any abbreviations, symbols and shading used?
✓ avoid distracting extraneous detail?
✓ include error bars where appropriate?