A continuing challenge for preclinical research on anxiolytic drugs is to capture the affective dimension that characterizes anxiety and aggression, either in their adaptive forms or when they become of clinical concern. Experimental protocols for the preclinical study of anxiolytic drugs typically involve the suppression of conditioned or unconditioned social and exploratory behavior (e.g., punished drinking or social interactions) and demonstrate the reversal of this behavioral suppression by drugs acting on the benzodiazepine-GABAA complex. Less frequently, aversive events engender increases in conditioned or unconditioned behavior that are reversed by anxiolytic drugs (e.g., fear-potentiated startle). More recently, putative anxiolytics which target 5-HT receptor subtypes produced effects in these traditional protocols that often are not systematic and robust. We propose ethological studies of vocal expressions in rodents and primates during social confrontations, separation from social companions, or exposure to aversive environmental events as promising sources of information on the affective features of behavior. This approach focuses on vocal and other display behavior with clear functional validity and homology. Drugs with anxiolytic effects that act on the benzodiazepine-GABAA receptor complex and on 5-HT1A receptors systematically and potently alter specific vocalizations in rodents and primates in a pharmacologically reversible manner; the specificity of these effects on vocalizations is evident due to the effectiveness of low doses that do not compromise other physiological and behavioral processes. Antagonists at the benzodiazepine receptor reverse the effects of full agonists on vocalizations, particularly when these occur in threatening, startling and distressing contexts. With the development of antagonists at 5-HT receptor subtypes, it can be anticipated that similar receptor-specificity can be established for the effects of 5-HT anxiolytics.