I mistyped: > I looks as if the tolerance should be 1.0E-7 in the calls to GRDFRQ. This is > close to the expected round-off error in computing the gridding error in > single-precision. 1.0E-7 * DF Chris Flatters cflatter at nrao.edu