The second variant is to fill the area between the curve and a given axis, a horizontal or vertical line, or a point.
The third variant requires three columns of input data: the x coordinate and two y coordinates corresponding to two curves sampled at the same set of x coordinates; the area between the two curves is filled.
Syntax:
set style [data | function] filledcurves [option] plot ... with filledcurves [option]
where the option can be
[closed | {above | below} {x1 | x2 | y1 | y2}[=<a>] | xy=<x>,<y>]
The first two plot variants can be further modified by the options
filledcurves closed ... just filled closed curve, filledcurves x1 ... x1 axis, filledcurves x2 ... x2 axis, etc for y1 and y2 axes, filledcurves y1=0 ... line y=0 (at y1 axis) ie parallel to x1 axis, filledcurves y2=42 ... line y=42 (at y2 axis) ie parallel to x2, etc, filledcurves xy=10,20 ... point 10,20 of x1,y1 axes (arc-like shape).
Example of filling the area between two input curves.
http://www.gnuplot.info/demo/fillbetween.htmlfill between curves demo.
plot 'data' using 1:2:3 with filledcurves
The above and below options apply both to commands of the form
... filledcurves above {x1|x2|y1|y2}=<val>
... using 1:2:3 with filledcurves below
Note: Not all terminal types support this plotting mode.
Zoom of a filled curve drawn from a datafile may produce empty or incorrect area because gnuplot is clipping points and lines, and not areas.
If the values of 5#5a6#6, 5#5x6#6, 5#5y6#6 are out of the drawing boundary, then they are moved to the graph boundary. Then the actually filled area in the case of option xy=5#5x6#6,5#5y6#6 will depend on xrange and yrange.