Reduced heart rate variability (HRV) has a strong predictive value in terms of sudden death, as compared to other prognostic criteria, but it has never been previously studied in experimental models of cardiac hypertrophy and failure. However, it has been quantified in rats using Holter monitoring and the peak and trough method of analysis. In normal rats, as in humans, short and long oscillations, sensitive, respectively, to atropine and propranolol were detected. Both correlate with heart rate. Cardiothyrotoxicosis was characterized by tachycardia and, independently, by a pronounced alteration in long oscillations. HRV was normal in compensatory cardiac hypertrophy due to aortic stenosis, but in this model the normal correlation existing between HRV and heart rate had disappeared. It is suggested that the main determinants of the above modifications of HRV are the changes in the new myocardial phenotype observed in terms of beta-adrenoceptor and muscarinic densities.