We present and analyze a series of mathematical models for the emergence of resistance during antibiotic treatment of an infected host. The models consider the population dynamics of antibiotic-sensitive and -resistant bacteria during the course of treatment and addresses the following problems: (i) the probability of obtaining a resistant mutant during the course of treatment as a function of antibiotic exposure; (ii) the conditions under which high, infrequent doses of an antibiotic are predicted to succeed in preventing the emergence of resistance; (iii) the conditions for the success of multiple drug treatment in suppressing the emergence of resistance and the relationship between antibiotic synergism and suppression of resistance; and (iv) the conditions under which nonadherence to the prescribed treatment regimen is predicted to result in treatment failure due to resistance. We analyze the predictions of the model for interpreting and extrapolating existing experimental studies of treatment efficacy and for optimizing treatment protocols to prevent the emergence of resistance.