GRID_CUBE

        GRID_CUBE

    This task makes a cube from a table  containing data  regularly  sampled
    on   a   2-d   grid, but possibly uncompletely sampled. Several gridding
    modes are possible.

    In mode GRID, since the input data is assumed to be  regularly  sampled,
    this  gridding  task  does not  use  any  convolution  or interpolation,
    but just fills a grid with values.

    In mode NATURAL,  a  convolution  kernel  is  used  to  interpolate  the
    irregularly sampled data using the weight of each individual point.

    In mode UNIFORM, the same convolution process is used but the individual
    points are assumed to have equal weight.


    The grid step is specified  by  the  user,  but  the  map  size  can  be
    determined by the task.  The input  table can be created by command GRID
    in CLASS, or using task TABLE from a formatted file.