[R] How to plot a blank plot

Marc Schwartz MSchwartz at medanalytics.com
Sat Feb 14 05:20:33 CET 2004

On Fri, 2004-02-13 at 15:31, Don MacQueen wrote:
> And although frame() is obviously the better way, if you really 
> wanted to know what to put inside plot(), this will do it:
>     plot(1, type='n', xaxt='n', yaxt='n', xlab='', ylab='', bty='n')
> -Don

A slightly shorter version of the same thing is:

plot(1, ann = FALSE, axes = FALSE, type = "n")

This gives you a blank plot area with the following, perhaps most
important, par value:

> par("usr")
[1] 0.568 1.432 0.568 1.432

plot.new() and frame() give you the following:

> par("usr")
[1] -0.04  1.04 -0.04  1.04

Recall that par("usr") is c(x1, x2, y1, y2).

In either case, the above values can be overridden by a subsequent call
to par("usr"), such as the following:

par(usr = c(0, 1, 0, 1))

Thus setting the x and y ranges to known values if one needs them a
particular way.  

Another way of accomplishing the same thing with a single function call

plot(1, ann = FALSE, axes = FALSE, xlim = c(0, 1), ylim = c(0, 1), 
     type = "n", xaxs = "i", yaxs = "i")

The use of 'xaxs' and 'yaxs' sets the ranges of the x and y axes to
exactly the limits specified, rather than extending both by 4%, which is
the default (with values of "r"). This yields:

> par("usr")
[1] 0 1 0 1

instead of:

> par("usr")
[1] -0.04  1.04 -0.04  1.04

as seen above from a default call to plot.new() or frame().

So, ultimately, if you are not going to actually draw anything in the
particular plot region as is Adrian's situation, such that you are not
concerned with the coordinates and other details, plot.new() and frame()
are the quickest as Don, I and others have pointed out. 

Indeed, as ?frame points out:

"...('frame' is an alias for 'plot.new')..."

> frame
function () 
<environment: namespace:base>

So which one you pick is really a personal preference choice. As is the
case frequently with R, there is more than one way to get to the same
end point, some more 'elegant' than others.


Marc Schwartz

More information about the R-help mailing list