We presented a 2-D test case for sparse data interpolation and give a good PEF with which to do it. The test case renders the traditional GEE two-stage interpolation scheme inapplicable. Claerbout (2000) suggests a nonlinear iteration, where filter and model are taken as unknown, but the best solution is still a subject of discussion among many SEP researchers. Regardless of the chosen interpolation strategy, the ``correct'' PEF and model are both known in this test case, so it should prove a useful starting point.