Small dimensional changes in six parameters were experimentally studied for their effects on the performance of the bistable fluid amplifier. These dimensional parameters were offset, wall length, control width, power nozzle width, wall flat, and power nozzle shape. The investigation was conducted on a ten times actual size model of a typical, commercially available amplifier. Data were taken for operating conditions of 1 nozzle load and blocked outputs over a range of Reynolds numbers (based on the nominal nozzle width) of 2000 to 6000. Evidence in support of the hypothesis of the coexistence of two, different level, overlapping switch points was presented. The hypothesis was used to explain unusual behavior and to fit smooth and regular curves to the experimental data obtained. The center of the investigation was the six tables of data generated and the method used to obtain figures of the specific dimensional characteristics desired by the user. It was found that the overwhelming influence of operating conditions (Reynolds number and output loading) prevented specific discussion of dimension characteristics without having the user's specific needs or operating conditions specified. From a broad or general look at the dimensional characteristics, it was found that wall-length had the most severe characteristics. It was also found that the wider control width promoted a clean and well-behaved switch absent of two coexistent switch points. For reasons unknown, it was found that the shorter power nozzle shapes made the device extremely sensitive to bias. / Master of Science