The laser-pulse sputtering of polymers should have two limiting cases. In the one there is ongoing release of particles from the target surface, the particles then form a Knudsen layer (KL), and there is finally an unsteady adiabatic expansion (UAE) ('effusion' model). In the other limit, bond-breakage occurs rapidly over a characteristic depth and the resulting gaslike particles then flow out directly in a UAE without a formal KL ('outflow' model). To test these idealized gas-dynamic descriptions, we discuss experiments in which approximately 20 ns excimer laser pulses are incident on polymethylmethacrylate in air at 193 or 248 nm and the release process is photographed with a approximately 1 ns probe pulse. The results not only give explicit support to the gas-dynamic description of the problem, but also indicate that the KL-UAE model is more appropriate. For example, only this model accommodates the observation that the release process continues for approximately 6-mu-s, which is approximately 500 times the laser pulse length.