Despite over a century of inquiry, the mechanisms that achieve the close matching of oxygen supply to demand during exercise remain elusive. It has been proposed that in addition to its role as the primary oxygen carrier, the red blood cell (RBC) functions as a roving oxygen sensor, linking the oxygen demand at the muscle with oxygen delivery via the circulation (Ellsworth et al. 1995). It is hypothesised that the RBC would release adenosine triphosphate (ATP) in proportion to the number of unoccupied binding sites on the haemoglobin molecule as it traverses regions of high oxygen demand such as the microcirculation of active skeletal muscle. ATP would then stimulate the release of vasodilatory substances from the endothelium which would diffuse to neighbouring vascular smooth muscle resulting in vasodilation and an increase in blood flow in accordance with the oxygen demand set by the muscle. The first step in establishing a role for this mechanism during exercise in humans is to determine whether ATP increases in the venous blood draining an active muscle bed. Based on the handful of published studies, there is an increase in ATP concentration in the femoral vein during knee extensor exercise. However the response has not been studied in other vascular beds in humans. As such, the main aim of this thesis was to measure the ATP response to dynamic handgrip exercise. Secondary aims were to determine whether the response was modified by hypoxia, and to provide information about the timing of the changes in ATP concentration during a bout of handgrip exercise. These questions were addressed in Studies 3 and 4. Because blood flow is central to this hypothesis, a substantial portion of this thesis was also associated with the measurement of forearm blood flow (FBF) using venous occlusion strain gauge plethysmography (VOSGP), and this was conducted in Studies 1 and 2. VOSGP is based on the assumption that with venous outflow prevented, any increase in limb volume is proportional to the rate of arterial inflow. The rate of arterial inflow is determined as the slope of the change in limb volume over time. The slope must be calculated over the initial linear portion of this relationship, when arterial inflow is unaffected by the inevitable rise in venous pressure associated with venous occlusion. VOSGP was initially used to measure blood flow at rest and in response to pharmacological interventions which produced only modest increases in arterial inflow (Joyner et al. 2001). However, measurement of the high rates of arterial inflow that occur with exercise may challenge the limits of this technique. Tschakovsky et al. (1995) reported a marked reduction in arterial inflow over the first four cardiac cycles during venous occlusion following static handgrip exercise that elevated blood flow to 22-24 mL/min/100mL. Only during the first cardiac cycle was arterial inflow unaffected by cuff inflation. As such, the window for measuring high rates of arterial inflow may be very brief. Therefore Study 1 aimed to determine whether blood flow could be measured using VOSGP across the range of arterial inflows that occur with dynamic handgrip exercise. Participants (n = 7) completed four, five-minute bouts of dynamic handgrip exercise at 15, 30, 45, and 60% of maximum voluntary contraction (MVC). FBF was measured using VOSGP at rest, and following five minutes of dynamic handgrip exercise. The slope of the change in limb volume was measured over the first one, two, three, and four consecutive cardiac cycles following the onset of occlusion. FBF was 2.5 ± 0.5 at rest, and 16.5 ± 4.9, 24.9 ± 9.4, 44.1 ± 22.0, and 57.8 ± 14.9 mL/min/100mL following five minutes of exercise at 15, 30, 45, and 60% MVC, respectively. At rest, arterial inflow decreased across the four cardiac cycles (P = 0.017 for the main effect), however post-hoc pairwise comparisons revealed no significant differences between any of the cardiac cycles. In contrast, the inclusion of two, three, or four cardiac cycles at 30 and 60% MVC, and three or four cardiac cycles at 15 and 45% MVC resulted in reductions in calculated arterial inflow compared with using the first cardiac cycle alone (P > 0.05). The inclusion of just two cardiac cycles resulted in a 9-26% reduction in calculated arterial inflow depending on the workload. This reduction was even more pronounced when three (19-40%) or four (26-50%) cardiac cycles were included. In conclusion, resting FBF can be calculated over at least four cardiac cycles during venous occlusion at rest. However, exercising FBF should be calculated from the first cardiac cycle only following dynamic handgrip exercise across the range of intensities used in this study. This extends the findings of Tschakovsky et al. (1995) who demonstrated this effect following handgrip exercise at a single intensity. Study 2 was designed to establish the FBF response to dynamic handgrip exercise, whether the workloads produced different blood flow responses, and to establish the within- and between-day reproducibility of FBF measured using VOSGP. In Part A (within-day reproducibility), participants (n = 7) completed three trials of dynamic handgrip exercise at four intensities (15, 30, 45, and 60% MVC), with each exercise trial separated by 10 minutes of rest. In Part B (between-day reproducibility) participants (n = 7) completed three trials of dynamic handgrip exercise at 15, 30, and 45% MVC on three separate days within a two week period. FBF was measured at rest, and each minute of exercise during brief (5-7 second) pauses in contractions. FBF response. FBF increased from rest at all workloads (P > 0.05), and then plateaued between Minutes 1 to 5 at the 15 and 30% MVC workloads and between Minutes 2 and 5 at the 45% workload (P > 0.05 for each minute compared to Minute 5). Too few participants completed the 60% workload to permit any statistical analysis. FBF reached values of 13.0 ± 2.0, 26.8 ± 8.4, 44.8 ± 14.9, and 52.9 ± 5.1 mL/min/100mL in the final minute of exercise at the 15, 30, 45, and 60% MVC workloads. FBF was different between the 15, 30, and 45% workloads by Minute 3 (P > 0.05). Reproducibility. The within-day test-retest reliability of exercising FBF was poor to moderate (ICC = 0.375-0.624) with individual coefficients of variation (CVs) ranging from 6-25%, 9-23%, and 9-31% for the 15, 30, and 45% MVC workloads, respectively. The between-day test-retest reliability for resting FBF was moderate (ICC = 0.644, P > 0.05; individual CVs between 1 and 31%). Between-day test-retest reliability for exercising FBF was poor to moderate (ICC = 0.381-0.614), with individual CVs ranging from 14-24%, 8-23%, and 6-18% for the 15, 30, and 45% workloads, respectively. It was concluded from this study that VOSGP provides adequately reproducible measurements to detect changes in FBF of the magnitude seen between workloads in this study. However, the variability in the measurement precludes its use when smaller differences are of interest. Based on the previous findings reporting an increase in ATP concentration during dynamic knee extensor exercise in the leg (Gonzalez-Alonso et al. 2002; Yegutkin et al. 2007), Study 3 was designed to determine whether ATP concentration increased in the venous effluent during dynamic handgrip exercise in the forearm. Since the deoxygenation of haemoglobin is a primary stimulus for ATP release from red blood cells, a further aim was to determine whether this response was augmented by systemic hypoxia. Participants (n = 6) completed four, five-minute bouts of dynamic handgrip exercise at 30, 45, 65, and 85% MVC under normoxia (inspired oxygen fraction = 0.21) and hypoxia (inspired oxygen fraction = 0.12). Blood samples for the determination of ATP concentration were drawn at rest and 180 seconds after the onset of exercise at each workload from a catheter inserted into a forearm vein. Venous plasma ATP concentration at rest was 0.28 ± 0.11 μM/L and remained unchanged during exercise at workloads up to 85% MVC (P > 0.05). Systemic hypoxia, sufficient to reduce arterial oxygen saturation to 83 ± 2%, also failed to alter the plasma ATP concentration (P = 0.148). The lack of a change in ATP concentration was unexpected but there are several possible explanations. It is possible, although unlikely, that ATP was not released in the forearm microcirculation. The previous demonstration that ATP increased in response to static handgrip exercise (Forrester and Lind 1969) would suggest that this was probably not the case. When considered in the context of the findings from Study 4, the most plausible explanation is that a less than optimal blood sampling site may have hindered the measurement of a change in ATP. The blood flow response at the onset of dynamic exercise in the forearm is at least biphasic; Phase 1 describes the immediate, large increase in blood flow within 2 seconds of the onset of exercise and is believed to be governed by mechanical factors whereas Phase 2 has a latency of ~20 seconds and describes a further, slower increase until blood flow reaches steady state (Saunders et al. 2005b). The temporal characteristics of Phase 2, along with the fact that blood flow during this phase is closely related to the metabolic rate of the muscle, suggest regulation by metabolic factors. Currently there is scant evidence detailing the time course of vasodilator release, although it is important to demonstrate that the release of a vasodilatory substance precedes the blood flow response it is proposed to influence (Delp 1999). ATP is released from red blood cells in proportion to the offloading of oxygen and a reduction in the oxygen content of venous blood draining a muscle bed occurs within 10 seconds of the onset of exercise. Thus the release of ATP should follow soon thereafter. As such, Study 4 was designed to determine whether ATP increased in the venous effluent of the forearm following 30 and 180 seconds of dynamic handgrip exercise at 45% MVC; and whether this increase corresponded with a decrease in venous oxygen content. Participants (n = 10) completed two bouts of dynamic handgrip exercise at 45% MVC; the first was one minute in duration, and the second was four minutes in duration. Venous blood samples for the determination of ATP and venous oxygen content were drawn at rest and during exercise from a catheter inserted in a retrograde manner into the median cubital vein. Arterialised samples for the estimation of arterial blood gases and ATP concentration were obtained from the non-exercising hand. ATP concentration in arterialised blood from the non-exercising arm was 0.79 ± 0.30 μM/L at rest and remained unchanged at both time points during exercise (P > 0.05). ATP concentration in the venous blood of the exercising arm increased from 0.60 ± 0.17 μM/L at rest to 1.04 ± 0.33 μM/L 30 seconds after the onset of exercise (P > 0.05), and remained at this higher level after 180 seconds (0.92 ± 0.26 μM/L, P > 0.05 versus rest). This corresponded with a decrease in venous oxygen content from 103 ± 23 mL/L at rest to 68 ± 16 mL/L 30 seconds after the onset of exercise (P > 0.05) and 76 ± 15 mL/L (P > 0.05 versus rest) 180 seconds into exercise. Furthermore, at 180 seconds of exercise, ATP concentration was moderately and inversely related to venous oxygen content (r = -0.651, p > 0.05). In conclusion, this study provides the first evidence that ATP concentration is increased in the blood draining the exercising forearm muscles in response to dynamic handgrip exercise. The finding that ATP concentration was increased just 30 seconds after the onset of exercise is also novel, and particularly interesting in the context of the recently reported dynamic response characteristics of the forearm blood flow response. In conclusion, the work contained within this thesis provides several important findings. The first study has provided evidence that measuring high rates of arterial inflow using VOSGP is possible, but that the window for making these measurements is small, probably as brief as a single cardiac cycle. The second study demonstrated that while the reproducibility of forearm blood flow measurements using VOSGP is poor, it is adequate to detect the large changes that occurred between workloads. However, VOSGP cannot be used to detect more modest differences. Common to both Study 3 and 4 was the measurement of ATP at rest, and 180 seconds after the onset of dynamic handgrip exercise at 45% MVC. The primary difference was the position of the catheter which was inserted in an antegrade manner in Study 3, and in a retrograde manner in Study 4. Since ATP was unchanged in Study 3 but increased under similar conditions in Study 4, it is likely that ATP was also released during exercise in Study 3, but that a less than optimal blood sampling site precluded its measurement. This illustrates the necessity to sample blood from as close as possible to the probable site of ATP release, the muscle microcirculation. The most important and novel findings from this body of work come from Study 4. This is the first study to demonstrate an increase in ATP concentration in the forearm in response to dynamic handgrip exercise. However, the most novel finding was that ATP concentration was elevated just 30 seconds after the onset of exercise. Such an early increase has not previously been reported during dynamic exercise in any vascular bed. This is an important finding since establishing the time course for the release of vasodilatory substances is critical to our understanding of the mechanisms that regulate blood flow during exercise.