This study aimed to characterise both the (V) over dotO(2) kinetics within constant heavy-intensity swimming exercise, and to assess the relationships between (V) over dotO(2) kinetics and other parameters of aerobic fitness, in well-trained swimmers. On separate days, 21 male swimmers completed: (1) an incremental swimming test to determine their maximal oxygen uptake d (V) over dotO(2max), first ventilatory threshold (VT), and the velocity associated with (V) over dotO(2max) (v(V) over dotO(2max)) and (2) two square-wave transitions from rest to heavy-intensity exercise, to determine their (V) over dotO(2) kinetics. All the tests involved breath-by-breath analysis of freestyle swimming using a swimming snorkel. (V) over dotO(2) kinetics was modelled with two exponential functions. The mean values for the incremental test were 56.0 +/- 6.0 ml min(-1) kg(-1), 1.45 +/- 0.08 m s(-1); and 42.1 +/- 5.7 ml min(-1) kg(-1) for (V) over dotO(2max), v(V) over dotO(2max) and VT, respectively. For the square-wave transition, the time constant of the primary phase (tau(p)) averaged 17.3 +/- 5.4 s and the relevant slow component (A'(sc)) averaged 4.8 +/- 2.9 ml min(-1) kg(-1) [representing 8.9% of the end-exercise (V) over dotO(2) (% A'(sc))]. tau(p) was correlated with v(V) over dotO(2max) (r = -0.55, P = 0.01), but not with either (V) over dotO(2max) (r = 0.05, ns) or VT (r = 0.14, ns). The % A'(sc) did not correlate with either (V) over dotO(2max) (r = -0.14, ns) or v(V) over dotO(2max) (r = 0.06, ns), but was inversely related with VT (r = -0.61, P < 0.01). This studywas the first to describe the (V) over dotO(2max) kinetics in heavy-intensity swimming using specific swimming exercise and appropriate methods. As has been demonstrated in cycling, faster (V) over dotO(2max) kinetics allow higher aerobic power outputs to be attained. The slowcomponent seems to be reduced in swimmers with higher ventilatory thresholds.