

   HHiissttooggrraammss oorr DDeennssiittyy PPlloottss ooff MMuullttiippllee GGrroouuppss

        ldahist(data, g, nbins=25, h, x0= - h/1000, breaks,
                xlim=range(breaks), ymax=0, width,
                type=c("histogram", "density", "both"),
                sep=(type != "density"),
                col=if(is.trellis()) trellis.par.get("bar.fill")$col else 2,
                xlab=deparse(substitute(data)), bty="n", ...)

   AArrgguummeennttss::

       data: vector of data. Missing values (`NA's) are allowed
             and omitted.

          g: factor or vector giving groups, of the same length
             as `data'.

      nbins: Suggested number of bins to cover the whole range
             of the data.

          h: The bin width (takes precedence over `nbins').

         x0: Shift for the bins - the breaks are at `x0 + h *
             (...{}, -1, 0, 1, ...{})'

     breaks: The set of breakpoints to be used. (Usually omit-
             ted, takes precedence over `h' and `nbins').

       xlim: The limits for the x-axis.

       ymax: The upper limit for the y-axis.

      width: Bandwidth for density estimates. If missing, the
             Sheather-Jones selector is used for each group
             separately.

       type: Type of plot.

        sep: Whether there is a separate plot for each group,
             or one combined plot.

        col: The colour number for the bar fill.

       xlab: label for the plot x-axis. By default, this will
             be the name of `data'.

        bty: The box type for the plot - defaults to none.

        ...: additional arguments to `polygon'.

   DDeessccrriippttiioonn::

        ~brief description

   VVaalluuee::

        None.

   SSiiddee EEffffeeccttss::

        Histogram and.or denisty plots are plotted on the cur-
        rent device.

   SSeeee AAllssoo::

        `plot.lda'.

