To investigate how arterial baroreceptors affect the dynamic properties of short-term blood pressure control, we determined Lyapunov exponents and correlation dimensions of blood pressure. Two groups of conscious dogs were studied: a control group (n = 7) and a group subjected to total sinoaortic and cardiopulmonary baroreceptor denervation (n = 7). As a measure of variability, standard deviation was determined and power spectra were calculated. In the lower frequency range (f < 0.1 Hz) power density was inversely related to frequency in both groups, indicating "1/f noise." Estimating the correlation dimension via the Grassberger-Procaccia algorithm as a quantification of complexity revealed a decrease after baroreceptor denervation (1.74 +/- 0.2 vs. 3.05 +/- 0.23 control; P < 0.05). Determination of the largest Lyapunov exponents lambda 1, which indicates the sensitive dependence on initial conditions, a hallmark of chaos, also yielded a diminution after denervation (lambda 1 = 0.74 +/- 0.08 vs. 1.85 +/- 0.18, P < 0.01). The results were cross-checked with surrogate data statistics. The null hypothesis, that there is no nonlinear structure in arterial blood pressure time series, was rejected. This shows that after baroreceptor denervation, blood pressure control is less complex and less sensitive to initial conditions ("chaos"). In contrast, variability (standard deviation) is increased (22.2 +/- 3.1 denervation vs. 8.3 +/- 1.4 control; P < 0.05). It is concluded that under physiological conditions, arterial and cardiopulmonary baroreceptors reduce variability of blood pressure, however, at the cost of blood pressure being less predictable. Thus the regulation is more sensitive depending on initial conditions.(ABSTRACT TRUNCATED AT 250 WORDS)