Much is known about the complex network structure of the Web, and about behavioral dynamics on the Web. A number of studies address how behaviors on the Web are affected by different network topologies, whilst others address how the behavior of users on the Web alters network topology. These represent complementary directions of influence, but they are generally not combined within any one study. In network science, the study of the coupled interaction between topology and behavior, or state-topology coevolution, is known as 'adaptive networks', and is a rapidly developing area of research. In this paper, we review the case for considering the Web as an adaptive network and several examples of state-topology coevolution on the Web. We also review some abstract results from recent literature in adaptive networks and discuss their implications for Web Science. We conclude that adaptive networks provide a formal framework for characterizing processes acting 'on' and 'of' the Web, and offers potential for identifying general organizing principles that seem otherwise illusive in Web Science.