Why do I get a false integral symbol when I do \bm{\int(2x}? I've just loaded the package bm in a usual way, and I'm not using any other different font in the preamble.

enter image description here

  • 1
    i'm not somewhere that i can check, but this looks to me very much like "poor man's bold" for symbols which have no "real" bold version. this mechanism sets three instances of the specified symbol, offset slightly from one another. i don't remember that a true bold integral exists in any usual tex font, so this approach makes sense, although the result is obviously terrible. Commented Apr 28, 2013 at 2:55
  • If you don't have any better alternative and need to fake bold, the best way is to change the width of the pdf stroke directly, as explained here.
    – Xavier
    Commented Apr 28, 2013 at 3:19
  • @Xavier when bm was written the majority of people were using bitmap fonts:-) Commented Apr 28, 2013 at 10:14
  • 2
    Please always supply a complete (small) document that shows the problem. bm tries several approaches to get bold and the poor man's bold you show is it's last alternative. But the code path depends on what other font packages you are using, and the order they are loaded Commented Apr 28, 2013 at 10:16
  • @DavidCarlisle Oh, it's really a poor man's bold :)
    – Xavier
    Commented Apr 28, 2013 at 13:44

1 Answer 1


The range of symbols available depends on the fonts used, and the classic Computer Modern collection doesn't have a bold version of the symbol extension font used for integral signs.

Assuming you don't want to change font setup globally there are a couple of choices. Use an extra set of {..} so \bm sees this as a complex expression and just uses \boldmath internally. This will make the other symbols bold but leave the integral sign from the standard font. or you could modify the poor man's bold to offset the symbols by a smaller amount so the gaps don't show. The original used an arbitrary value of .4mu here I use .2mu.

enter image description here








  • Thanks for this but it has caused the lower limit to appear vertically below the other rather than joined with the bottom of the integral. any idea how to fix it.
    – yolo
    Commented Jan 8, 2015 at 19:20
  • 1
    @yolo hard to do that in a clean way: hit it with a big stick: \bm{\int}_{\mkern-9mu 0}^n Commented Jan 8, 2015 at 20:25

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .