It corrects for differences in total sample size ($n$).

It makes the bins look larger.

It forces the distribution to be normal.

It simplifies the calculation of the mean.