Here are the results of the quick poll I did last week related to whether figures should be placed in line or at the end of a manuscript. I prefer having the figures at the end of a manuscript (because this way I know where to find figures that are referred to multiple times), but I suspected I was in the minority. That suspicion was correct. Below, I also give results of where people want their figure legends placed: almost everyone wants the legend on the same page as the figure itself.
There was a preference for having figures in line for all of the career stage groupings (current PhD students, people who got their PhDs within the past 1 to 10 years, people who got their PhDs more than 10 years ago, and people who are not students and do not have PhDs). However, as I guessed would be the case, this preference was stronger for current grad students than for people who got their PhDs within the past 1 to 10 years, and least strong for those who got their PhD more than 10 years ago. It was highest for the group that didn’t have PhDs and weren’t students, but there were many fewer people in that group (15 total).
The number on the bar in white indicates the total number of respondents in that career stage grouping (meaning: those who said they prefer figures in line + those who prefer them at the end)
Based on these results, I thought that maybe I should put the figures in line for the manuscript I’m currently working on, rather than at the end as I usually do. But then when I went to look at the author guidelines for the journal I’m planning on submitting it to, they state that they want the figures in a separate file. This works for my personal reviewing preferences, but apparently is not preferred by more potential reviewers. So, now I will need to decide whether to follow the author guidelines, or whether to stick the figures in line and add a note indicating that, while I know the author guidelines requested the figures in a separate file, I’ve put them in line based on the most likely preferences of the reviewers and handling editors. (This latter idea comes from this comment by Chris Mebane on last week’s post.) I suspect I’ll follow the author guidelines.
This exercise also has me wondering whether journals could poll their reviewers to ask where they prefer to see the figures, with the idea that they could then potentially change their author guidelines to reflect what their reviewers prefer. There was some discussion in the comments last week about how the author guidelines reflect what is easiest for the typesetter — but I think that Stephen Heard’s point that we should be making things easier for the volunteer peer reviewers rather than the paid typesetters makes a lot of sense.
In terms of where people want figure legends: 95% of the poll respondents want those on the same page as the figure. That doesn’t surprise me!
Now that we’ve tackled this topic, it’s time to get on to the really controversial stuff. Expect polls on Oxford commas, one or two spaces after a period, and fonts/typefaces in the near future.*
*(I’m joking.)
Dynamic Ecology 
IMO, the worst part of reviewing a manuscript is the fact that Figures are at the end of the MS and the captions are on a completely separate page. My personal preference would be to review a fully (or almost) typeset and formatted MS.
Doesn’t hypertext make these concerns moot? If you’re printing to paper then obviously no.
I find a fully hyperlinked PDF is the best format for reading a paper, particularly for finding references but also tables and figures. This can be easily achieved by the author with most software options before typesetting happens.
I agree that captions separate from figures is awful (and negates the hyperlink).
As a reviewer I want Figures and Tables where I can see them in context, not at the end of the manuscript. Putting Figures and Tables at the end of the paper is a hang-over from the old days when you submitted paper versions an Figures were either on tracing paper (yep – tracing paper) or as photographic plates! Being paper you could rip them off the back and insert them into the right place in the manuscript you were reviewing or just have them to hand in a separate pile.
When I review a manuscript, I keep two copies open on separate monitors – one with the text that I’m reading, the other with the figures. This is the digital version of sleather2012’s tearing-the-figures-off method, I guess.
So I’m fine with the figures at the end, but I do wish the captions went along with the figures (or maybe I should get a third monitor).
Maybe this standard, odd, raw formatting of manuscripts is best explained by how layout programs (such as Adobe Indesign) work than by what reviewers prefer.