#> 2 is a good divergence criterion for the general cubic...