Hmmm, this is tricky because to make the scales equal, I'm currently
extending the shorter scale to the length of the longer - for this
example it pushes the limits in to the negatives, which is problematic
for log axes. Thanks for the reproducible example - I'll look at it
when I'm re-looking at coord_equal.
Hadley