1. The cardiac and respiratory responses shown by muskrats in both unrestrained and restrained dives have been compared with responses elicited by stimulation of a number of cardio-depressant receptor inputs, in an attempt to determine which are most important in initiating and maintaining diving bradycardia. 2. In unrestrained voluntary dives heart rate fell from 310 +/- 3 to 54 +/- 3 beats min-1 in 1 to 2 sec, which was significantly below that seen in dives by restrained unanaesthetized or anaesthetized animals. 3. Pouring water on the external nares during maintained artificial ventilation caused heart rate to decline to 76 +/- 12 beats min-1 after 1 sec. Flowing water through the internal nares caused apnoea, in the expiratory position, and bradycardia within one third of a second. Heart rate fell to 20 +/- 2 beats min-1, 1 sec after the start of water flow. Substituting saline for water reduced both the apnoeic and cardiac responses. Bilateral section of the maxillary branch of V and the inferior laryngeal (X) nerves completely abolished the cardiac and respiratory response to water flow. 4. Artificial ventilation throughout periods of nasal stimulation with water or saline reduced the bradycardia, although even the saline driven response could not be completely abolished. Lung deafferentation eliminated any direct effect of artificial ventilation on heart rate during nasal stimulation. 5. Lung deflation caused bradycardia within 0.97 +/- 0.17 sec, heart rate falling from 268 +/- 7 to 59 +/- 4 beats min-1. Bradycardia also occurred during maintained lung inflation but it was delayed for a period which varied from 6.8 +/- 1.8 sec at an inflation pressure of 0.5 kPa to 35 +/- 7 sec at 1.5 kPa. 6. Bradycardia caused by nasal water flow or lung deflation was unaffected by bilateral section of the sinus nerve. 7. Artificial ventilation of paralysed muskrats with 5% CO2 in N2 caused bradycardia when Pa, O2 reached 8.4 +/- 0.8 kPa and heart rate declined to 76 +/- 7 beats min-1 at 4 kPa. Bilateral section of the sinus nerve delayed bradycardia until Pa, O2 reached 4.5 +/- 0.5 kPa. 8. These results suggest that the cardiac response to submergence could be the expression of input from three groups of receptors, nasal, lung and carotid chemoreceptors, although it is not clear how they interact with one another to generate the cardiac responses displayed by unrestrained animals during submergence.