I have been using Delete Identical with only the shape filed checked in a model to delete stacked polygons after doing an intersect of census blocks and multiple overlapping buffers. I discovered that not all of the records with the same shape are being deleted. because of the spatial joins that happen after this intersect I thought a work-around would be to place the attributes onto centroid points. The duplicate points show no spatial diffrence zoomed in to 1:0. Even run out side the model Find Identical sees the records but Delete Identical does not. I only found this by doing a set of QC calculations checking that all the census numbers added up to the original block areas and populations, a few (not all) came up as high (so far) as 200% I need to process some 60000 census blocks this way so I need better reliability in the deleting of duplicate records. any suggestions. I am using ArcInfo desktop, features are stored in a File Geodatabase, and all the data is in the NAD_1983_UTM_Zone_11N projection.