The problem is that the usual boxplot generally can't give an indication of the number of modes. While in *some* circumstances, it is possible to get a clear indication that the smallest number of modes exceeds 1, in general a given boxplot is consistent with any number of modes.

Given the large sample it was based on, this one does indicate the presence of at least two modes (the data were generated so as to have exactly two):

$\qquad\qquad $ ![enter image description here][1]

but conversely, this one has two very clear modes in its distribution but you simply can't tell from the boxplot at all:

![enter image description here][2]

(Beware, however -- [histograms can have problems, too!](http://stats.stackexchange.com/questions/51718/assessing-approximate-distribution-of-data-based-on-a-histogram/51753#51753) -- and as Nick says, kernel density estimates may also affect the impression of the number of modes.)

There are modifications that can better indicate multimodality. In some situations they may be useful, but if I'm interested in finding modes I'll usually look at a different sort of display.

Boxplots are better when interest focuses on comparisons of location and spread (and perhaps to skewness) rather than the particulars of distributional shape. If multimodality is important to show, I'd suggest looking at displays that are better at showing that - the precise choice of display depends on what you most want it to show well. 


  [1]: https://i.sstatic.net/Eybbr.png
  [2]: https://i.sstatic.net/ZorKO.png