BACKGROUND: The phrase complete vagal withdrawal is often used when discussing autonomic control of the heart during exercise. However, more recent studies have challenged this assumption. We hypothesized that cardiac vagal activity increases during exercise and maintains cardiac function via transmitters other than acetylcholine. METHODS: Chronic direct recordings of cardiac vagal nerve activity, cardiac output, coronary artery blood flow, and heart rate were recorded in conscious adult sheep during whole-body treadmill exercise. Cardiac innervation of the left cardiac vagal branch was confirmed with lipophilic tracer dyes (DiO). Sheep were exercised with pharmacological blockers of acetylcholine (atropine, 250 mg), VIP (vasoactive intestinal peptide; [4Cl-D-Phe6,Leu17]VIP 25 µg), or saline control, randomized on different days. In a subset of sheep, the left cardiac vagal branch was denervated. RESULTS: Neural innervation from the cardiac vagal branch is seen at major cardiac ganglionic plexi, and within the fat pads associated with the coronary arteries. Directly recorded cardiac vagal nerve activity increased during exercise. Left cardiac vagal branch denervation attenuated the maximum changes in coronary artery blood flow (maximum exercise, control: 63.5±5.9 mL/min, n=8; cardiac vagal denervated: 32.7±5.6 mL/min, n=6, 2.5×10), cardiac output, and heart rate during exercise. Atropine did not affect any cardiac parameters during exercise, but VIP antagonism significantly reduced coronary artery blood flow during exercise to a similar level to vagal denervation. CONCLUSIONS: Our study demonstrates that cardiac vagal nerve activity actually increases and is crucial for maintaining cardiac function during exercise. Furthermore, our findings show the dynamic modulation of coronary artery blood flow during exercise is mediated by VIP.