Combinatorial methods involving data collection in multiparameter space allow a rapid identification of measured property trends as a function of system parameters. The technique has been applied with success to pharmaceutical, inorganic and organic materials synthesis, but not significantly to measurements of polymeric films and coatings. We demonstrate the use of 2-D combinatorial libraries to investigate thin-film dewetting. We have prepared libraries of thin films of polystyrene on silicon substrates containing orthogonal, continuous variations of thickness (h), and temperature (T) that represent about 1200 practical state points per library. The libraries were screened for dewetting behavior using automated optical microscopy. Dewetting trends were visibly apparent on the libraries, and a comprehensive map of the T, h, and time (t) dependence was generated in a few hours. The combinatorial libraries, spanning a large T, h, and t range, not only reproduced known dewetting structures and phenomena but also enabled a novel T, h superposition of the heterogeneous nucleated hole dewetting kinetics. We observed three hole nucleation regimes as a function of thickness: heterogeneously nucleated holes (h > 55 nm), a crossover regime where both heterogeneous and capillary instability nucleation compete (33 nm < h < 55 nm), and a regime of holes nucleated by capillary instability (16 nm < h < 33 nm).