The steeplechase requires a unique combination of endurance, power, and technique. In comparison with an open 3000 m race without water-jumps or barriers, runners are typically 30 s slower in a 3000 m steeplechase race (Popov, 1983). Steeplechase athletes must jump barriers 35 times during the 3000 m race. Seven of the barriers are followed by a 3.66 m water pit which gradually slopes upwards until it is even with the track surface from a depth of 0.70 m (Figure 1). The slower times in the 3000 m steeplechase compared with the open 3000 m race show the effect the barriers have on performance. While conditioning and physiological parameters are the main determinants of performance in steeplechasing (Kenney and Hodgson, 1985), certain adjustments in technique over the water-jump may benefit performance as has been shown in steeplechase and sprint hurdling (Hunter and Bushnell, 2006; McDonald and Dapena, 1991).

The barrier heights in the steeplechase are equal to the hurdle heights used in the 400m race (0.762 m and 0.914 m for women and men respectively). The differences in barrier height, body height, and approach velocity between men and women lead to differences in step lengths and body positioning in hurdling (Hunter, 2006; McDonald and Dapena, 1991). Although a different movement pattern is required for the water-jump compared with hurdling, some gender differences were expected in the water-jump for the same reasons described above with hurdling.

This study investigated the characteristics of technique that lead to maintaining velocity through the water-jump, and differences between men and women in various aspects of technique.

Eighteen men and 18 women were filmed from four different views during all seven water-jumps during two 3000 m steeplechase races (2006 USATF National Championships and 2006 Cardinal Invitational at Stanford University) with digital video cameras running at 60Hz with a shutter speed of 1/250 s (Canon Elura 60, Lake Success, NY). The cameras were placed 20-30 m from the water jump at four locations diagonal to the water jump. For each jump, two perpendicular camera views were chosen from the four camera positions. The two views were chosen based upon lighting and positioning of other athletes. The cameras were zoomed to include 6 m prior to and 4 m past the water jump. Athletes were digitized throughout this range once their entire body was completely in the field of view.

Prior to data collection, a survey pole calibration was performed using four 2.43 m poles with 36 locations digitized. The poles were positioned in a rectangle enclosing the area the athletes would be analyzed in. A theodolyte was used to determine the location of each pole.

Athletes had an average finish time of 8:38 ± 0:16 for men and 10:01 ± 0:09 for women. The University Institutional Review Board approved the study and waived the need for informed consent since the race was deemed a public event.

All jumps from all athletes were digitized using Peak Motus 8.2 (Colorado Springs, CO) using a 20-point spatial model. Since markers could not be placed upon the athletes, joint centers and endpoints of segments were determined by the researchers. Center of mass calculations were completed using body segment parameters adjusted from Winter, 1990. Following the application of the Direct Linear Transformation (Abdel-Aziz and Karara, 1971), three-dimensional coordinates were low-pass filtered at 6 Hz as determined optimal by the Peak Motus 8.2 program. Then the following variables were calculated (Figures 2 and 3">3):

Two stepwise multiple linear regressions were completed (one for men and one for women) with v/p as the dependent variable and all the variables listed above other than exit velocity as the independent variables.

Velocity divided by race pace was chosen as the dependent variable since runners have to return to race pace following the water jump. If they can keep from slowing too much during the water jump, there will be less acceleration required as they exit the water pit. However, coaches and athletes should realize that increasing this variable indefinitely is not advisable. Since seven jumps were measured for each athlete, the average value for each variable was inputinto the regression model. A repeated measure ANOVA was performed to determine differences between genders in all the variables listed above. Finally, the repeated measures ANOVA was performed again after normalizing the variables (other than crouching height and push-off angle) for race pace. This allowed us to determine whether any differences between genders could be accounted for by differences in race pace.

Velocity through the jump divided by race pace (v/p) was predicted very well by approach velocity and landing distance for men and women (Table 1). All other variables were removed as they did not make significant contributions to increasing multiple R² values.

The relationships of v/p with all parameters showed linear trends throughout the range of measured values. However, only three variables were found significantly different after normalizing by race pace. Variables that appear to be due to gender alone rather than just different race paces include: Push-off angle (greater for women), exit velocity (smaller for women), and loss of velocity (greater for women) (Table 2).

Women and men both showed linear trends in the v/p regression model throughout the range of measured values. There was no apparent local maximum in the relationships between v/p and approach velocity or landing distance. Thus, increasing approach velocity or landing distance through the ranges of these athletes should increase v/p further. However, it is important to keeping in mind that a higher v/p value is not necessarily advantageous, because economy of movement must be considered in the steeplechase. None of the athletes in this study approached the water-jump barrier at their maximum velocity or attempted to maximize their landing distance. If they had, the current regression model predicts a higher v/p. While this may seem desirable at first, the large fluctuation in effort would likely result in a much greater energy cost (Billat et al., 2001).

While only seven water-jumps occur in a 3000 m steeplechase race, the obstacle must have some effect on running time since it takes athletes away from their normal running stride. The most important factor in steeplechase performance is physical conditioning (Kenney and Hodgson, 1985). With the water-jump making up only about 1% of the total race distance, even a weak correlation between v/p and race pace encourages us to believe that v/p is an appropriate variable to consider in steeplechase performance. This correlation was observed and was very small (R² = 0.02, F = 4.04, p = 0.046).

In order to complete a water-jump obstacle close to race pace (v/p close to 1. 00), one must obtain relatively high approach velocity. Average approach velocities were5.32 m/s (5:02 minutes/mi) and 6.16 m/s (4:21 minutes/mi) for women and men respectively. While many may think an increased v/p is desirable, an approach velocity could become too high. Since economy is so important in the steeplechase, it is possible to go too high above race pace. Another important factor to consider is landing distance. The athletes that are more successful at the water-jump land relatively close to the end of the water-pit. These athletes typically get only one foot wet with each jump (the second foot plant is beyond the water-pit). This matches with the average landing distance values found in this study (2.54 m and 2.85 m for women and men respectively, p < 0.001). Since the pit is 3.66 m long, those with a lower v/p are landing deeper and typically getting both feet wet before exiting the water pit.

Approach velocity and landing distance were expected to be correlated with v/p. Perhaps the more interesting finding of this study is the lack of significance of other variables. This may explain why many world-class steeplechasers appear awkward in their movements over the water jump. As long as their approach velocity is high and they obtain a relatively long landing distance, the other aspects of their technique do not relate to v/p.

The water-jump produces a greater disruption in running velocity in the women than in the men. Exit velocity was greater for men even after accounting for race pace while approach velocity was not. Demonstrating this idea further, loss of velocity was greater for women after accounting for differences in race pace. The lower exit velocities after accounting for race pace might be explained by the pit being the same length for men and women. Since women do not jump as far as men, they will be landing deeper in the water. They are also jumping from a lower height, which results in a decreased flight time compared with the men who jump from a greater height.

Women extend at the knee more than men as they push off the barrier. Pushing off the barrier through a greater extension may help women obtain a greater landing distance, partly overcoming their slower approach velocity and lower barrier height. Crouching height is no different even though body heights are typically different. Since women are taking off from a lower barrier height in the steeplechase water-jump, they may be crouching less to obtain a greater take-off height and increase flight time to get a longer jump.

One limitation to the current study is the lack of information about body height. Some of the gender differences may be due to body height rather than gender alone. Thus, the reason for some gender differences remains unknown.

While technique differences exist between men and women in the water-jump, the movement is similar. The same focus should be given to men and women in terms of what makes a successful water-jump. Increasing approach velocity leads to greater v/p. However, it should be realized that increasing v/p indefinitely is not desirable due to the required economy of a 3000 m race. Working towards an optimal v/p should be focus of steeplechasers. Coaches and athletes should realize that other small differences in technique occur between elite men and women, but they have little impact on the overall performance of the water-jump.

Women may need to be coached differently than men in technique for the water jump. With the lower approach velocity for women, women take off closer to the barrier so that they can land on top of it in a good position. The lower approach velocity also appears to lead to a shorter jump and as a consequence, results in an uphill step out of the water pit following landing. Thus, women's race paces are affected more by the water-jump than those of the men, as was found in this study, with loss of velocity divided by race pace being greater for women than men.

Success in completing the water-jump of the 3000m steeplechase without dropping from race pace dramatically can be accomplished by accelerating during the approach to the barrier and accomplishing a relatively long landing distance. There are obviously limits to how much acceleration and how far of a jump off the barrier should be attempted. However, with the athletes analyzed in this study, the larger the approach velocity and the longer the jump into the water, the better the athletes were able to keep their water-jump horizontal velocity close to their race pace. Thus, training for the water jump should include surging along with any technical work.