The purpose of this study was to analyze the relationship between the
intra-cycle variation of the horizontal velocity of displacement of the
center of mass (dV), the hand’s and feet’s velocity, as well as, to identify
the variables that most predict the dV’s, in butterfly stroke. The study
was divided in two parts. The aim of Part I was to investigate the behavior
of variables in study at slow swimming velocities and the purpose of
Part II was the same but at high swimming velocities. 3 male Portuguese
swimmers and 1 female swimmer, of international level were studied in
Part I. The swimmers were submitted to an incremental set of 200 m
butterfly swims. In the Part II, 7 Portuguese male swimmers of national
and international level were studied. Each swimmer performed two
maximal 25 m butterfly swims. Both protocols were recorded from four
different plans, allowing a 3D analysis. It was calculated the dV, the 3D
components (Vx, Vy, Vz) of the hand’s velocity and the 2D components
(Vx, Vy) of the feet’s velocity. Several variables presented significant
correlation coefficients with dV at all selected velocities (high velocity
ranged from r = 0.58 for Vx-out to r = 0.82 for Vy-1dwn; slow velocity
ranged from r = -0.45 for Vx-1dwn to r=0.73 for Vx-ups; overall velocity
ranged from r= 0.34 for Vz-ent to r = 0.82 for Vx-ins). It was also
computed a regression model to predict dV. For high velocity (up to
1.75 ± 0.09 m.s-1), the variables that best predict dV were Vy during the
first downbeat, Vx and Vy during the arm’s insweep (r2 = 0.93). At slow
velocity (up to 1.48 m.s-1), the variables included in the forward step-bystep
regression model were Vx during upsweep, Vy and Vx during
insweep (r2 = 0.69). For overall velocity, the variables that most fit the
regression model were Vx during upsweep, Vy during second downbeat
and Vz during entry (r2= 0.94). In order to reduce dV, butterfliers should
increase hand’s velocity in all orthogonal components at the end of the
underwater path, should increase the vertical velocity during
