The most commonly heard vocalizations of frogs are advertisement calls, which attract gravid females and mediate aggressive interactions between males. Frog vocalizations are energetically costly to produce, and body size often constrains the dominant frequency and intensity of vocalizations; propagation and degradation of these signals are affected by diverse physical and biotic factors. Behaviors and auditory mechanisms that mitigate these problems are discussed. With some exceptions, female preferences based on dominant frequency are intensity-dependent and mediate stabilizing selection within populations. Female preferences based on dynamic, gross-temporal properties typically mediate strong directional selection. The high values of these properties preferred by females increase a male's detectability in dense choruses and are a reliable predictor of his energetic investment in courtship. Female preferences based on fine-temporal properties (e.g. pulse rate) are often in tensity-independent and usually mediate stabilizing selection within populations. The overall attractiveness of a signal depends on variation in more than one of these acoustic properties; their relative importance differs between species. Parsimony analysis supports the idea that auditory biases preceded the evolutionary appearance of call elements that enhance the attractiveness of advertisement calls in one species group of neotropical frogs. A more specific claim that the bias has not been modified by selection after the establishment of the new signal has little empirical support. Indeed, the selective consequences of positive phonotaxis to any ''new'' stimulus, whether or not there is a sensory bias, must play a critical role in its establishment and maintenance as a mate-attraction signal and on the further evolution of the female preference. The hypothesis that present-day selective consequences of mate choice have also acted in the past evolution of call structure and preferences is supported by a few examples of reproductive character displacement. However, evolutionary divergence in signals and preferences will have multiple causes, most of which do not involve interactions between species. Phylogenetic analyses and studies of selection and other evolutionary forces in contemporary populations are complementary approaches to gaining insights about the evolution of frog vocalizations and animal communication in general.
