While an extensive literature on cardiovascular development exists for insects, almost all studies focus on in vitro preparations, and very few report on more than a single developmental stage. The present study examines in vivo cardiac performance in the intact, unanesthetized larvae, pupae and adults of the tobacco hornworm Manduca sexta. For all three stages, electrode pairs of fine steel wire were inserted subcuticularly at two dorsal abdominal locations. Impedance signals produced by contraction of the dorsal abdominal vessel (tube heart) were amplified and recorded. In addition to providing heart rate, a comparison of the relative timing of the signal from each electrode pair allowed the calculation of the propagation velocity and direction of heart contraction. Experimental treatments of intact animals included exposure to hypoxia and hyperoxia (21 %, 15 %, 10 %, 5 %, 0 % and 100 % O(2)), to hypercapnia (0 %, 4 %, 8 %, 16 %, 20 % and 24 % CO(2)), to temperature variation (10, 20 and 30 degrees C) and to 2 min periods of forced activity. The pattern of contraction of the dorsal abdominal vessel of M. sexta changed substantially with developmental stage. Larvae showed a relatively simple, invariably posterior-to-anterior pattern (mean rate 34.8+/-1.16 beats min(-)(1)). The heart rate pattern in pupal M. sexta displayed great variability in rate, amplitude and direction. Periods of regular heart beats (21.5+/-1.09 beats min(-)(1)) were frequently and irregularly interrupted by periods of cardiac arrests ranging from a few seconds to over 20 min. Adults showed a highly stereotypic but complex pattern, with periods of 'fast forward' (FF; rate 47.6+/-2.6 beats min(-)(1)), 'slow forward' (SL; 32.8+/-3.0 beats min(-)(1)) and 'reversed' (R; 32.2+/-2.4 beats min(-)(1)) beating. The contraction propagation velocity in larvae and pupae averaged 5. 52+/-0.36 and 2.03+/-0.11 cm s(-)(1), respectively. The SF, R and FF phases of the adults had average propagation velocities of 5.52+/-0. 51, 5.05+/-0.52 and 5.43+/-0.37 cm s(-)(1), respectively. Heart rate and contraction propagation velocity were remarkably resistant to ambient hypoxia and hypercapnia at all developmental stages, decreasing significantly only at 0 % O(2) or 24 % CO(2). As expected, the heart rates of all three developmental stages increased significantly with increasing temperature, with heart rate Q(10) values for larvae, pupae and adults of 2.33, 3.14 and 1.61, respectively, between 10 and 20 degrees C. Corresponding Q(10) values for these stages between 20 and 30 degrees C were 2.22, 2.03 and 2.29. Larval heart rates showed no significant response to forced activity induced by prodding. In contrast, adult heart rate increased nearly fivefold from 50.1 beats min(-)(1) during rest to 223.5 beats min(-)(1) after 1 min of prodding. The activity-induced tachycardia in adults ceased within 10-12 min. Patterns of cardiac contraction in larval, pupal and adult M. sexta were as dissimilar as their morphological appearances and revealed a gradation from simple to complex. These developmentally based distinctive cardiac patterns are undoubtedly related to developmental differences in both morphology and life-style. Larvae are anatomically 'homogeneous' compared with other stages, with no distinct head, thorax and abdominal region (or wings) that might require selective perfusion or drainage. The far more complex pattern of heart activity seen in pupae probably relates to the dramatic changes in internal morphology during this stage. Simultaneous degradation and synthesis of tissues throughout the body may expose the heart to numerous peptides or neurohormones that affect cardiac activity. In adult moths, the complex and repetitive pattern of cardiac activity is reflected in the previously described complexity of hemolymph movement, together with thermoregulatory capabilities in this species that depend on well-regulated hemolymph movements between the thorax, wings and abdomen.