The dissertation is a historical and theoretical study of competition between the Bell and independent telephone systems between 1894 and 1920. It is concerned with the historical origins of telephone monopoly in the U.S., and with the unique dynamics of competition between unconnected or incompatible communications networks. The study focuses on the competing networks' refusal to interconnect with each other, exploring the economic and communicative consequences of fragmented telephone communications. Two bodies of theory provided the foundation for the study's method: the "network externality" literature in Economics and the probabilistic models of interdependent demand developed by W. Brian Arthur. The dynamics of network competition are illustrated by means of an urn model. Unlike previous efforts, this urn model incorporates the possibility of nonuniform calling patterns and user duplication. In order to display the actual scope of telephone competition and to evaluate theories about the role of long distance connections in the competitive struggle, maps of the telephone access universes of three cities at various points in time were constructed. The conclusions of the study conflict with many standard assumptions about telephone history. Bell's refusal to connect with the independents stimulated and broadened the scope of competition rather than thwarting it. The concept of "universal service," first formulated at this time, denoted an end to competitive fragmentation rather than a telephone in every home. The universality of the U.S. telephone system had its roots in the competitive era rather than in subsequent regulatory policies. A telephone monopoly was created not because it realized supply-side economies of scale, but to achieve demand-side economies of scope. The decisive ingredient in Bell's success was not its ultra-long distance transmission technology but its ability to offer near-universal connections within a 100 mile region.