Conceptual models of polar motion have evolved over the past century, as improved data revealed signals over progressively wider frequency bands. In the 1890s, Chandler represented polar motion as a sum of discrete components, 14 month and annual terms, and this component model effectively summarized the observations, but did not provide a physical explanation for them. Over time both the search for a physical understanding of polar motion, and the ability to observe the broad band continuum outside the Chandler and annual bands have led to an understanding of polar motion as a continuum of variations, not adequately described by a few discrete components. The continuum concept is now the working model in most studies of polar motion. The transition from component to continuum conceptual models preceded the arrival of high quality data by several decades, and reflected significant contributions from Harold Jeffreys. With modern space geodetic observations and good global numerical models of the atmosphere, oceans, and other climate processes, it is clear that air and water motion and redistribution are the dominant contributors to the excitation continuum.