Brady and colleagues have championed the importance of careful delineation and control of a subject's behavioral state. In this paper we develop the concept of visceral behavior from a physiological perspective. Sprague-Dawley rats were instrumented to record arterial blood pressure, renal sympathetic nerve activity, and respiration. The rats were restrained in a conical cloth sock. Rats that were well adapted to the sock restraint showed a regular respiratory pattern and consistent pressure recordings; they rested quietly in the sock and moved only occasionally to adjust their position. Fourier analysis of blood pressure and nerve activity showed a concentration of power below 1 Hz. The coherence between the two signals was strong (0.83 ± 0.03) at 0.42 Hz. Conversely, during their initial adjustment to the sock restraint, the rats tended to show large fluctuations in blood pressure associated with episodic apneic breathing; 1 animal displayed this pattern of visceral behavior throughout most of the experiment. Despite this instability in pressure, the rats' overt behavior was stable: They rested quietly in the sock with only occasional position shifts. Spectral analysis and coherence computations showed large shifts in the distribution of power and frequency range over which arterial pressure and sympathetic activity were tightly coupled. These data are consistent with the view that an animal's circulatory adjustments, as well as adjustments in other aspects of its physiological state, constitute an important aspect of behavior, and that this behavior can influence the interpretation of biobehavioral data.Key words: autonomic nervous system, blood pressure, respiration, power spectrum, rat J. V. Brady's professional career has emphasized the importance of the control and stability of an experimental subject's behavior. In contrast, physiologists tend to focus their primary attention upon visceral, rather than somatic, function. With the advent of implantable transducers, many physiologists are now studying organ function in unanesthetized subjects. These subjects obviously interact with their environment and display many different somatic behavior patterns. This has required the physiologist to understand and appreciate the possible influence of the subjects' behavior upon the biological systems of interest at the moment.This new appreciation has prompted novel ways to understand and classify behavior. For example, Engel and colleagues have pointed out that the cardiovascular responses of an organism may be considered to be an aspect of its overall behavior (Engel, 1986). More explicitly, they hold that "behavior is an inte-