The effect of fear and relief from fear on heart rate (HR) was studied in partially restrained adult male rats. The emotional state was influenced behaviorally with the use of Pavlovian aversive conditioning procedure, and pharmacologically by injections of the selected anxiolytics and anxiogenics. A signal of danger (DS) - light or tone, preceded tail-shock (excitatory trials), while a signal of safety (SS), respectively tone or light, overlapped last 3s of 5s DS and predicted an omission of this expected aversive event (inhibitory trials). To assess the stability of HR changes to DS and SS we analyzed whether and how the experimental conditions (modality and position of signals, the ratio of numbers of excitatory to inhibitory trials in the session) influenced HR. HR changes to DS were different in pattern, stability and direction when compared to HR changes in response to SS. Reactions to DS, although accompanied mainly by conditioned bradycardia, were not consistent and depended on experimental conditions. However, the SS always evoked conditioned tachycardia. Anxiolitycs - benzodiazepines (diazepam and midazolam) and buspirone - influenced HR in nonconsistent manner. Anxiogenics - PTZ and FG7142 were without significant effects on HR. HR could not be trusted therefore as direct index of pharmacologically induced emotional states. The results are discussed in the context of possible biobehavioral meaning of HR changes in response to danger and safety and their reliability as fear/relief correlates.