Most terrestrial animals massing several grams or more use more than two legs to move. Despite injury, aging, and disease; in the face of unknown and unstable substrates — animals often achieve their locomotion objectives. We seek a theory of animal motion that would allow us to design such robust locomotion into robotic systems. One candidate is oscillator theory. Viewed from its lens, rapid legged locomotion can be seen as a hybrid oscillator, or a collection of phase locked hybrid oscillators. Our recent theoretical advances allow us to analyze the stability of a large class of such hybrid oscillators, and demonstrate that by exploiting the hybrid structure they obtain robustness, and allow for modularity by requiring low communication complexity. Our new construction methods, based on modular robotics, have allowed us to rapidly build, test and evolve robot designs. Our quantitative analysis of the robot motions teaches us that we require better models for multi-contact slippage if the motions of our hexapedal robots are to be understood.