The concept of critical speed has been proved to be a simple and non-invasive method for monitoring aerobic fitness in
swimming [2,3]. This method involves swimming at maximum effort over distances that usually range from 50 to 400
m, with subsequent determination of the linear regression equation between distance and time needed to cover it at
maximal speed. The slope of the regression line represents the "critical speed", which is an index of aerobic fitness,
while the y-intercept has been termed "anaerobic swimming capacity" and corresponds to the maximal distance that
could be performed anaerobically [1]. The purpose of the present study was to examine the validity of these two indices
by comparing the slopes and y-intercepts obtained from swimmers of different training and performance levels.

Forty four swimmers (20 males, 24 females, age: 22.2±1.9 y, Ht: 176.2±9.6 cm, arm span: 179.7±12.2 cm, Wt:
70.0±5.7 Kg, mean±SD) took part in this study. Of these swimmers, 19 were members of the Greek national team, 8
were competitive club level swimmers and 17 were Physical Education (P.E.) students specializing in swimming. To
ensure that all subjects were in their best physical condition, all measurements took place 10-20 days before the Greek
Championships (athletes) or before the practical examination period (students). All swimmers were instructed to swim
as fast as they could over four different distances (25, 50, 100 and 200 m) on four different occasions, 1-3 days apart.
Critical speed (CS) was estimated from the slope of the linear regression line between swimming distance and time,
while the y-intercept (y-INT) was assumed to represent anaerobic swimming capacity [1]. Means were compared using
a one-way or two-way ANOVA and Tukey post-hoc test for unequal number of subjects (Spjotvoll/Stoline test).

The group of P.E. students was slower over all swimming distances and had a lower CS compared to the national and
club level athletes (Table 1). No significant differences in any parameter were found between the national and club level
athletes. Contrary to what was expected, the y-INT was lower in the national level athletes compared with the P.E.
students, whereas no other significant differences in y-INT were found between groups. The r2 values (goodness of fit)
of the linear regressions equations were always >0.998 (P<0.001). When the regression equations were calculated from
only two swimming tests (100 and 200 m), the values of critical speed were lower (P<0.01) and y-intercept increased in
all groups (P<0.01). However, the y-intercept calculated from 100 m and 200 m tests was similar in all groups tested
(Table 1).

The results of this study show that critical speed calculated from the slope of the linear regression of swimming distance
and time differentiates swimmers of different training and performance levels, irrespective of the test distances used to
calculate it, and may therefore be used to monitor changes in aerobic fitness. However, the y-intercept of the regression
line, gave erroneous and inconsistent results and may not provide reliable information about the anaerobic swimming

[1]. Dekerle et al. (1992a). Int. J. Sports Med. 23: 93-98
[2]. Wakayoshi et al. (1992a). Eur. J. Appl. Physiol. 64: 153-157
[3]. Wakayoshi et al. (1992a). Eur. J. Appl. Physiol. 64: 153-157

NOTA: O texto com a iconografia está no anexo.