New test-particle simulations have been carried out to learn more about the secular, evolution and morphology of the gaseous and stellar distributions in barred galaxies. We verify the previous results Of P. Schwarz that gas clouds will tend to collect into ring-like patterns near major orbit resonances, owing to gravity torques. However, we improve on these results in several ways. Firstly, we use more gas clouds (10 000 vs 2000) than Schwarz and track individual clouds to determine when they collide, rather than using a collision box as did Schwarz. Secondly, besides Schwarz's isochrone rotation curve, we also use a flat rotation curve. Thirdly, we consider more bar pattern speeds and strengths than did Schwarz. Finally, unlike Schwarz and other simulators, we have a large database of images and color index maps of nearly 140 ringed galaxies that can be compared to the simulations to evaluate their significance. We confirm the two types of outer Lindblad resonance rings that Schwarz discovered, but find that their existence is not due so much to the initial density distribution of gas clouds as to pattern speed and the time interval since the bar potential was imposed. The simulations and the images lead us to suggest that we can divide barred galaxies according to the resonances which the bar pattern speed and rotation curve allow in the disk. We illustrate specific galaxies that we believe belong to fast, medium, and slow bar pattern speed ''domains'' and match them to particular simulation frames. We also discuss alternative hypotheses in which the pattern speed is such that all resonances are present but the gas has been depleted or was distributed differently in various galaxies to produce the three classification domains.