Running times are a function of reaction time, acceleration potential, maximal running velocity and the ability to maintain velocity as fatigue progresses (Ross et al., 2001). These factors are clearly influenced by metabolic (Allemeier et al. 1994; Barnett et al. 2004; Dawson et al. 1998; Jacobs et al., 1987; Mero et al., 1981) and neural factors (Casabona et al., 1990; Jönhagen et al. 1996; Mero et al., 1992; Nummela et al., 1994; Ross et al., 2001), however, anthropometric factors also play an important role (Mann et al., 1984; Mero and Komi, 1985).

A muscle’s maximal force is, generally speaking, proportional to its physiological cross-sectional area (Izquierdo et al., 2001; Powell et al., 1984). If we consider two geometrically and qualitatively similar individuals, we would expect all linear dimensions to be proportional. Accordingly, we would anticipate body mass (BM) to be proportional to height cubed (since mass equals density times volume), and muscle strength to be proportional to height squared (since area scales with height squared). In fact, it has been reported that muscle strength (measured as weight lifted) varied almost exactly with height squared in world weightlifting champions (Ford et al., 2000). Consequently, muscle strength relative to BM, i.e. the relative muscle strength (RMS), should be inversely proportional to height. In other words, we would expect RMS to decrease with increasing height.

In sprint running, the center of mass (COM) is accelerated during the early part of the race before reaching a plateau between 40 and 60 m into the race (Delecluse et al., 1995). During each step, the COM is accelerated both vertically and horizontally by the application of large forces during the stance phase, after the body is slowed by air resistance (horizontally) and negatively accelerated by gravity (vertically) prior to foot-ground contact. The acceleration of the body is proportional to the force produced but inversely proportional to the body mass, according to Newton’s second law. Therefore, in theory, the ability to accelerate the BM should be closely related to RMS. This implies an inverse relationship between height and performance in disciplines such as sprint running (Figure 1).

Although being tall may have disadvantages for sprinters, it may provide some advantages as well. A taller runner’s longer limbs will enable longer step length (Winter, 1990), which could be advantageous since running speed is a function of step frequency and step length. Accordingly, one may expect a smaller stature to be a disadvantage in sprint running. In sprint running, one may therefore expect a smaller proportion of world-class sprinters to be short in stature compared to the normal population.

While it is unclear whether there is an optimum body height, these theoretical considerations suggest that there should be an optimum height at which sprint performance is best. Therefore one would expect less height variability among sprint runners than among normal populations. The purpose of the present study was to describe differences in height and BM, and their variability, between world-class sprinters and normal populations in order to test the hypothesis that there is less height and BM variability among world-class sprinters than among normal populations.

Height and BM data (IAAF Statistics, 1980-2004) of 42 men and 44 women from the all-time 100m top-50 lists (International Association of Athletics Federations, http://www.iaaf.org/statistics/toplists) were compared to those of American (724 male and 663 female) and Danish (1336 male and 1306 female) 20-30 year-old residents. There are many Americans among the world-class sprinters. Therefore, it was considered appropriate to compare the sprinters with the American population. However, in order to validate the results against a second population, the sprinters were also compared to Danish normal population. Even supposing some ‘trained sprinters’ might be included in the ‘normal population’, the number would be very small and therefore its effect could be considered negligible. Accordingly, the two samples ‘sprinters’ and ‘normal population’ can be considered different.

The anthropometric data of the sprinters were obtained in part from numerous internet searches and other sources – too many to cite individually. Hence the height and BM data from the sprinters are presented in full in Appendix 1 (men) and 2 (women). The age range of 20-30 years was selected in order to maximize the match of the comparing groups. The American data are public-use data from National Health and Nutrition Examination Surveys (NHANES 1999-2002). The Danish data were part of a study conducted by the Danish National Institute of Public Health (formerly known as The Danish Institute for Clinical Epidemiology) (Kjøller and Rasmussen, 2002).

The body mass index (BMI= BM·height^-2) is the recommended screening tool for overweight and obesity. A large muscle mass can yield high BMIs even though body fat is not excessive, because the BMI fails to distinguish between the proportions of body fat and lean tissue. Nevertheless, BMI can be used to describe the compactness of a person. Thus it can be a useful anthropometric parameter. Since both height (determinant for limb lengths) and BM (determinant for muscle mass and hence strength) impact running speed, the interaction of body height and BM is important.

Q-Q plots were made to test if data sets followed a normal distribution. In these cases, in which the assumption of normality was accepted, group variability (SD) was compared by using a two-tailed F-test, and group mean was compared by using one-way analysis of variance. For all variables, the sprinters and two normal populations were compared. In addition, male sprinters were compared to female sprinters. The p-values were adjusted for multiple testing (Bonferroni correction). Statistical significance was set at p < 0.05 for all analyses.

The results for the males are presented in Table 1. The male Americans were significantly shorter than male sprinters and male Danes (p < 0.01). But there was no significant difference in the mean height of male sprinters and male Danes. The male sprinters had less height variability (i.e. SD) than the male Danes (p < 0.01) who had less height variability than the male Americans (p < 0.01).

The male sprinters had lower mean BM than the male Danes (p < 0.01), and both these groups were lighter than the male Americans (p < 0.001). The male sprinters had less BM variability that male Danes (p < 0.001) who had less BM variability than male Americans (p < 0.001).

The male sprinters had a significantly lower mean BMI than male Americans (p > 0.001). But there was no significant difference in the mean BMI of male sprinters and male Danes (p = 0.11). The male sprinter showed less BMI variability than the male Danes (p < 0.001) who showed less BMI variability than the male Americans (p < 0.001). Thus, in general, the sprinters are lighter in BM and have lower BMI than normal populations and the height, BM and BMI variation among sprinters is less than the variation among normal populations.

The results for the females are presented in Table 2. The female Americans were significantly shorter than the female sprinters (p < 0.001) and female Danes (p < 0.001). But there was no significant difference in the mean height of the female sprinters and female Danes (p = 0.67). All female groups had similar height variability (p = 0.30-0.40).

The female sprinters were significantly lighter than the female Danes (p < 0.001) who were significantly lighter than the female Americans (p < 0.001). Likewise, the female sprinters had significantly lower BM variability than the female Danes (p < 0.001) who had significantly lower BM variability than the female Americans (p < 0.001).

The female sprinters had a significantly lower mean BMI than female Danes who had a significantly lower mean BMI than the female Americans. Likewise, the female sprinters showed less BMI variability than the female Danes who showed less BMI variability than the female Americans. Thus, in general, the sprinters are lighter in BM and have lower BMI than normal populations and the BM and BMI variation among sprinters is less than the variation among normal populations.

The male sprinters were significantly taller (p < 0.001) and heavier (p < 0.001) than the female sprinters, but there was no significant difference in the height (p = 0.18) or BM (p = 0.13) variability. The male sprinters had higher BMI (p < 0.001) than the female sprinters. There was, on the other hand, no difference in the BMI (p = 0.43) variability between the male and female sprinters.

The present analysis revealed a tendency toward less height, BM, and BMI variability among sprinters than among the normal populations. These results suggest the existence of a limited optimum range for height, BM and BMI for sprinters. The normal populations from both the U.S. and Denmark showed significantly higher BM than the sprinters. This result may be explained by higher body fat levels among the normal population given that sprinters often have lower body fat levels (Abe et al., 2001; Kumagai et al., 2000; WHO, 1998). However, in the present study, no fat measurement was performed.

The fact that sprinters have very low BM variability when compared to normal populations suggests that very low or very high BM could be a limiting factor in sprint running. High BM will handicap a sprinter because it takes higher force to accelerate a larger mass. On the other hand, strong sprinters should have more muscle mass and therefore be heavier than less strong sprinters. Hence, sprinters with a very low BM would probably have less muscle mass and thus be too weak. Moreover, the results of the present study point towards the sprinters’ BMIs being less variable when compared to the BMI of the normal populations. This result indicates that sprinters also have an optimum range for BMI that differs for males and females. Height, BM and BMI seem to be important anthropometric parameters for sprinters.

The male sprinters were heavier, taller and had a higher BMI than the female sprinters. Therefore, it may be concluded that there does not exist an optimum inter-gender height, BM or BMI for sprinters. In fact, regarding mean height, only the American normal population differed statistically from the sprinters. For a given height a male sprinter would, on average, be approximately 10 kg heavier than a female sprinter (see Figure 2). A significant correlation has been found between strength and muscle cross-sectional area although there does not seem to be a significant gender difference in the strength to cross-sectional area ratio (Miller et a1., 1993). These data suggest that the greater strength of the men primarily is caused by their larger muscles.

It has been reported that mechanical power output during a single short-term maximal exercise is greater in men than in women even when the differences in fat-free mass of whole body as well as BM are normalized (Froese and Houston, 1987; Gratas-Delamarche et al., 1994; Mayhew and Salm, 1990). When the results of the present study are taken into account it seems reasonable to speculate that there is a connection between the gender difference in BM and performance. Since high muscle strength (and therefore muscle mass) is required to perform well in the sprint events, the male sprinters may benefit both from their larger muscle mass (Cureton et al., 1988) and the influences of the principal muscle-building hormone, testosterone, on the development of anaerobic working potential (Kraemer et al., 1991). The male sprinters may therefore be heavier and perform better than the female sprinters simply because they have more muscle mass.

The present study demonstrated less height variability among male world-class sprinters than among American and Danish men. On the other hand, all female groups had similar height variability. Further research is required to disclose if this can be explained by gender differences in the degree of track running specialisation.

All of the world-class sprinters were in the height range of 1.68-1.91 m (men) or 1.52-1.82 cm (women). The fact that the sprinters’ height data was normally distributed indicates that both very short and very tall stature may be disadvantageous for sprinters. No mean height difference was observed between the female sprinters (1.68 m) and the Danish women (1.69 m), however, there was a significant height difference (p < 0.0001) between the American women (1.63 m) and the female sprinters. In other words, there might be very little likelihood that future top sprinters would be outside of this height range. Even though the present study indicates that very tall or short stature may reduce the chance of being a successful sprint runner, several world-class sprinters are taller than 1.87 m (i.e. Linford Christie, Carl Lewis, Joshua J. Johnson and Asafa Powell), which corresponds to the top 10% of the male U.S normal population and the top 23% of the male Danish normal population. Likewise several world-class sprinters were shorter than 1.71 m (i.e. Andre Cason, Leonard Myles-Mills and Coby Miller), which corresponds to the bottom 20% of the male U.S normal population and the bottom 5% of the male Danish normal population. Therefore, there is a small chance that a future male world-class sprinter may be taller than 1.91 m or shorter than 1.68 m.

The sprinters were compared to two populations at specific time points. Other time points might give different results, especially regarding BM, as people tend to be fatter (Hill and Melanson, 1999; Ogden et al., 2004). The mean height has also increased during the last decades. However, while among the American adult population, mean BM have increased more than 10 kg, mean height has increased less dramatically (2-3 cm) during the 40 years (Ogden et al., 2004). To test if the same happened among the sprinters, they were assigned to groups, old (men: 1983-1999; women: 1984-1994) and recent (men: 1999-2005; women: 1995-2005), of approximately equal size according to the date of their best performance. No statistical height, BM or BMI difference was observed between the groups (men: BM_old/recent= 75.7/77.7 kg, p = 0.18, height_old/recent = 179.2/181.4 m, p=0.20, BMI_old/recent = 23.6/23.8 kg·m^-2, p=0.57; women: BM_old/recent= 57.6/58.7 kg, p = 0.52, height_old/recent = 168.5/167.0 m, p = 0.47, BMI_old/recent= 20.3/21.0 kg·m^-2, p = 0.10). Therefore, no data was excluded due to the date of the performance even though some of the athletes performed in the 1980s and early 1990s, which is earlier than the normal populations (the U.S. data were from 1999-2002).

Although the U.S. has the greatest representations in the group of sprinters, comparisons against another population might turn out differently. However, comparison against the Danes who are completely unrepresented gave relatively similar results, thus suggesting that the findings are valid.

It is likely that there is no single optimal height for sprinters, but instead there is an optimal range that differs for males and females. This range in height appears to exclude people who are very tall and very short in stature. Sprinters are generally lighter in body mass than normal populations. Also, the body mass variation among sprinters is less than the variation among non-athletic populations. These anthropometric characteristics typical of world-class sprinters might be explained, in part, by the influence the anthropometric characteristics have on relative muscle strength and stride length.