Introduction
Sprint performance is a direct result of the impulse (mean force multiplied by contact time) applied by the athlete against the ground (^{1} ). Therefore, one of the aims of training is to increase the specific strength of sprinters (^{12} ). This situation has led specialists to investigate new forms of training , with the objective of obtaining greater levels of adaptation. The benefits of the use of resisted sprint running is that it recruits more muscle fibers, requires more neural activation (^{2,7} ), and increases the load in hip extensor muscles (^{8} ). This mode of training is believed to develop specific strengths and to increase stride length (^{4,5,8,16,21} ). There have been relatively few longitudinal studies that have examined the effects of sled towing training on sprint performance (^{13,18,20,23} ); no studies have been found for trained sprinters.

The training principle of specificity states that for an exercise to be effective, it must maintain similar characteristics to the sport requirements (^{19,22} ). Different studies suggest that to maintain load specificity in sprints, the horizontal velocity should not fall bellow 90% of the athlete's maximum velocity (^{11,14,16} ). Furthermore, Lockie et al. (^{16} ) and Spinks et al. (^{20} ) have proposed an equation to calculate the adequate load required in sprint training with sled towing (Equation 1). Their study was developed with athletes from various field sports (e.g., field hockey, rugby, Australian football, soccer) in the acceleration phase (15 m) with sled towing.

where % velocity = the required training velocity as a percentage of maximum velocity (e.g., 90% of maximum).

This equation allows one to calculate the load for sled towing for the acceleration phase of a sprint. However, because the maximum velocity phase has different characteristics than the acceleration phase (^{3,17,22} ), the work of maximum speed required a different type of work and load. Therefore, the present paper has a double purpose. The first purpose was to verify whether the equation proposed by Lockie et al. (^{16} ) and Spinks et al. (^{20} ) works to establish the loads with sprint athletes in specific training conditions (on a synthetic track and wearing spiked sprint shoes). The second purpose was to develop an equation that accurately describes the relationship between towing loads and the resulting sprint velocity in the maximum velocity phase. Both studies have the final objective of helping coaches calculate the load of their sled tow training .

Methods
Experimental Approach to the Problem
To develop an equation using regression analysis that accurately described the relationship between towing loads and the resulting velocity on the maximum velocity phase, a cross-sectional analysis of sprints, jumps, and decathlon athletes was done. Different studies suggest that to maintain load specificity in sprints, the horizontal velocity should not fall below 90% of the athlete's maximum velocity (^{11,14,16} ). Therefore, Lockie et al. (^{16} ) and Spinks et al. (^{20} ) have proposed an equation to calculate the adequate load required in sprint training with sled towing (Equation 1). Their study was developed with athletes from various field sports (e.g., field hockey, rugby, Australian football, soccer) in the acceleration phase (15 m) with sled towing. This equation was developed on nonspecific athletic conditions. A pilot study was conducted with the aim of verifying whether this equation works well to establish the loads with sprint athletes in specific training conditions.

Pilot Study
Six Spanish national men's sprinters participated in the study (Table 1 ). The ethics committee of the Catholic University San Antonio of Murcia approved the study. The participants were informed of the protocol and procedures before their involvement, and their written consent to participate was obtained.

Table 1: General characteristics of participants (mean ± SD ).

The sprint trials were conducted on a synthetic track (Rekortan M99, APT Corp.) in an outdoor athletics stadium. Participants wore their own athletic training clothes and spiked sprint shoes. First, the sprinters were weighted to determine the load to use for losing 10% of maximum velocity, applying the equation described by Lockie et al. (^{14} ) (Equation 1).

After a specific sprinting warm-up, athletes carried out 4 trials of 15 m, of which 2 trials each were completed with a sled and without a sled. The weighted sled (Power Systems Inc., Knoxville, Tenn) was attached to the athlete by a 3.6-m cord and waist harness. The sled traveled on 2 parallel metal tubes about 500 mm long and 30 mm in diameter. The sliding surfaces of the base of the sled were smooth and bare. The fastest trial was selected for the analysis. Electronic timing gates (Biomedics Inc, Barcelona, Spain) were used to record the participants' sprint times. The gates, accurate to 0.01 seconds, were placed at 1, 6, 11, and 16 m to record the participants' sprint times for every 5 m. The first gate was placed 1 m from the start line to avoid being influenced by the reaction time, and the last gate was set at 16 m (Figure 1 ). For the study, the time for the last 5 m (10-15 m) was used because velocity in this section was highest for each subject. The wind velocity for all trials was measured using a wind gauge (Cantabrian, Cambridge, England). Trials in which the wind was not between 2 and −2 m·s^{−1} were repeated.

Figure 1: Scenario of the 15-m acceleration sprints.

The data were analyzed using a nonparametric method for 2 related samples (Wilcoxon method; SPSS 13.0, SPSS Inc., Chicago, Ill) to study differences between estimated loss in velocity and real loss in velocity. The alpha level was set to p ≤ 0.05.

Significant differences were not found between estimated percentage in maximum velocity lost using the equation by Lockie et al. (^{16} ) and real lost percentage in maximum velocity (Table 2 ). However, the results are close to significant (p = 0.075). Consequently, caution must be used with this equation when applying it to experienced sprinters, jumpers, or decathletes.

Table 2: Estimated average maximum velocity, estimated and real percentages of maximum velocity, and difference between load and unloaded sprints in study 1 (n = 6).

Regarding load calculation for the acceleration phase, the results show that the equation proposed by Lockie et al. (^{16} ) and Spinks et al. (^{20} ) allows us to establish the load in sled towing with an error of ±2.2% in displacement velocity. This error in load calculation tends to involve coming up short rather than excess. Table 3 presents the reference values of loads for the different individual weights and work intensities for working acceleration sprints with sled towing obtained by the equation from Lockie et al. (^{16} ).

Table 3: Load (kg) required for sled towing in acceleration training , depending on individual body mass.

Subjects
Twenty-six men participants were recruited for the second study. The participants were active competitive athletes who specialized in sprints, jumps, or decathlon, and all had previously used resisted sprint devices in their training . They had been regularly performing bodybuilding-type resistance training (e.g., approximately 6-12 repetitions per set, 3 sets per exercise, twice a week) for at least the previous 2 months, they could produce a force close to twice their body weight during a dynamic squat lift (Table 1 ), and they had no recent injuries or medical conditions that would prevent maximal exertion. Testing was conducted midseason. The ethics committee of the Catholic University San Antonio of Murcia approved the study. The participants were informed of the protocol and procedures before their involvement, and their written consent to participate was obtained.

Procedures
The sprint trials were conducted on a Rekortan M99 synthetic track in an outdoor athletics stadium. Participants wore their own athletic training clothes and spiked sprint shoes. Before commencing the sprint trials, the subjects were weighed to determine the loads relative to 6 (sled mass), 10, and 15% of their body mass. After that, the participants performed a sprint-specific warm-up consisting of 8 minutes of running with a heart rate of 140 bpm, 8 minutes of active stretching, 10 minutes of running technique exercises, and 2-4 submaximal and maximal short sprints.

A Power Systems weighted sled was attached to each athlete by a 3.6-m cord and waist harness. When sprinting at maximum speed, the angle of the cord to the horizontal surface was between 12.5 and 15.5°, depending on the athlete's body dimensions. The sled traveled on 2 parallel metal tubes about 500 mm long and 30 mm in diameter. The sliding surfaces of the base of the sled were smooth and bare. The load required on the sled was calculated using the equation, Load = (body mass × % body mass) − sled weight, where % body mass was worked out as a decimal (e.g., 5% body mass = 0.05) and sled weight = 4.7 kg.

The athletes performed four 30-m flying sprints (an unloaded sprint and sprints pulling resistances of 6, 10, and 15% of their body mass) at maximum intensity using a run-in distance of 20 m from a standing start to discover changes in maximum velocity (Figure 2 ). One trial was completed for each load (^{10} ). The order for the trials was randomized for each participant, and an unlimited rest period was given between trials to minimize the effects of fatigue on sprint performance. The rest period typically lasted about 6 minutes, which is sufficient for full recovery from repeated maximal sprints of short duration (^{9} ). The maximum velocity for the sprints was measured through the use of radar (StalkerPro Inc., Plano, Tex) with a record data frequency of 100 Hz. The wind velocity for all trials was measured using a wind gauge (Cantabrian, Cambridge, England), and trials in which the wind was not between −2 and 2 m·s^{−1} were repeated. Electronic timing gates (BioMedic) were placed at the start and finish of the 30-m flying sprint to record the sprint times. For wind velocities within this range, the wind produces a change in 30-m sprint time of less than ±1% from a zero-wind result (^{15} ).

Figure 2: Scenario of the 30-m flying sprints.

Statistical Analyses
Descriptive statistical methods were used to calculate mean and SD . Pearson correlation coefficients (SPSS 13.0) were used to determine the interrelationship among % body mass load and maximum velocity variables. The alpha level was set to p ≤ 0.05.

Results
Mean maximum velocity, 30-m sprint time, load, % body mass, and % velocity of both unloaded and loaded sprinting situations are shown in Table 4 . An increase in time and a reduction in speed were observed when the load increased.

Table 4: Maximum velocity, 30-m fly-sprint time, load, % body mass, and % velocity in unloaded and load 1, load 2, and load 3 sprints in study 2 (n = 26).

The resultant velocities produced from the loads were converted to a percentage of the maximum velocity for 30-m fly sprints. These data were plotted against each other to produce the regression equation (Equation 2, Figure 3 ). The R ^{2} value for the equation was 0.82. This value reflected a highly significant linear relationship (p < 0.001). The regression equation obtained is listed below.

Figure 3: Regression analysis of the effect of increases in load (as a percentage of body mass) on velocity (as a percentage of maximum velocity for 30 m).

where % velocity = the required training velocity as a percentage of maximum velocity, such as 90% of maximum.

Discussion
In relation to load calculation for the maximum velocity phase, the results allow one to calculate a regression equation (% body mass = (−0.8674 × % velocity) + 87.99. However, it should be noted that it has been calculated with men sprinters, jumpers, and decathletes with previous experience in resisted training methods and with spiked shoes on tartan surface.

In both types of work (acceleration and maximum velocity), the problem exists that it is not possible to work with light loads because the minimum weight of the sled is 4.0-4.5 kg. Therefore, this type of work should be done with other resisted training such as belts, vests, parachutes, etc.

The values presented in Tables 3 and 5 are only general reference values. They should be considered as estimates of work load. For example, timing during the season, shoes, and the surface used will affect the calculation. Coaches should control the running techniques of their athletes to avoid inadequate workouts. Working out with excessive loads in sled towing provokes significant increases in CG vertical oscillation and significant reductions in stride length (^{6} ).

Table 5: Load (kg) required for sled towing in maximum velocity training , depending on individual body mass.

Future research should study the effect of a 7-10% load on running technique and specific strength after a period of training (4-8 weeks) and the possible relationship between load and technique, anthropometry, or levels of strength.

Practical Aplications
This paper verifies previous tools and presents new ones to calculate the load for sled towing in the sprint training of acceleration and maximum velocity. The equations allow coaches and strength trainers to calculate the load for resisted training with sled towing. Tables 3 and 5 can help in quickly establishing work loads for athletes corresponding to their objectives. If you are training a 105-kg athlete and you want to work at 90% of his or her maximum velocity, you should use 13.22 kg when working the acceleration phase (0-30 m) and 10.42 kg when working the maximum velocity phase (30-60 m).

In must be kept in mind that the load applied to the athlete by a weighted sled depends on the coefficient of friction between the sled and the running surface and also on the weight of the sled. Therefore, the proposed equation is specific to the combination of sled and surface used in this study.

Acknowledgments
We would like to thank Dr. Nick Linthorne and Anthony J. Blazevich (Brunel University, London) for assistance throughout this project. We would also like to thank the athletes who have participated in the present study. This research was supported by Universidad Católica San Antonio de Murcia, Spain (FPI grant, code PMAFI-PI-05/1C/05)

References
1. Alexander, MJL. The relationship between muscle strength and sprint kinematics in elite sprinters.

Can J Sport Sci 14: 148-157, 1989.

2. Cissik, JM. Means and methods of speed

training , part I.

Strength Cond J 26(4): 24-29, 2004.

3. Cissik, JM. Means and methods of speed

training : Part II.

Strength Cond J 27(1): 18-25, 2005.

4. Costello, F.

Training for speed using resisted and assisted methods.

NSCA J 5: 74-75, 1985.

5. Donati, A. The association between the development of strength and speed.

N Stud Athl 11: 51-58, 1996.

6. Elvira, JLL, Alcaraz, PE, and Palao, JM. Effects of different resisted sprint running methods on stride length, stride frequency, and CG vertical oscillation. In:

XXIV ISBS Symposium 2006 . S. Hermann, ed. Salzburg: International Society of Sports Biomechanics, 2006.

7. Faccioni, A. Assisted and resisted methods for speed development: part 1.

Mod Athl Coach 32: 3-6, 1994.

8. Faccioni, A. Assisted and resisted methods for speed development: part 2.

Mod Athl Coach 32: 8-12, 1994.

9. Harris, RC, Edwards, RH, Hultman, E, Nordesjo, LO, and Nylind, B. The time course of phosphorylcreatine resynthesis during recovery of the quadriceps muscle in man.

Pflugers Arch 367: 137-142, 1976.

10. Hunter, J, Marshall, R, and McNair, P. Reliability of biomechanical variables of sprint running.

Med Sci Sports Exerc 36: 850-861, 2004.

11. Jakalski, K. The pros and cons of using resisted and assisted

training methods with high school sprinters. Parachutes, tubing and towing.

Track Coach 144: 4585-4589, 4612, 1998.

12. Korchemny, R. Evaluation of sprinters.

NSCA J 7: 38-42, 1985.

13. Kristensen, GO, Van Den Tillaar, R, and Ettema, GJC. Velocity specificity in early-phase sprint

training .

J Strength Cond Res 20: 833-837, 2006.

14. Letzelter, M, Sauerwein, G, and Burger, R. Resistance runs in speed development.

Mod Athl Coach 33: 7-12, 1995.

15. Linthorne, N. The effect of wind on 100-m sprint times.

J Appl Biomech 10: 110-131, 1994.

16. Lockie, RG, Murphy, AJ, and Spinks, CD. Effects of resisted sled towing on sprint kinematics in field-sport athletes.

J Strength Cond Res 17: 760-767, 2003.

17. Murphy, AJ, Lockie, RG, and Coutts, A. Kinematic determination of early acceleration in field sport athletes.

J Sports Sci Med 2: 144-150, 2003.

18. Paradisis, GP and Cooke, CB. The effects of sprint running

training on sloping surfaces.

J Strength Cond Res 20: 767-777, 2006.

19. Sale, DG. Neural adaptation to strength

training . In:

Strength and Power in Sport . Komi, PV, ed. Oxford: Blackwell Science, 2003. pp. 281-314.

20. Spinks, CD, Murphy, AJ, Spinks, WL, and Lockie, RG. The effects of resisted sprint

training on acceleration performance and kinematics in soccer, rugby union, and Australian football players.

J Strength Cond Res 21: 77-85, 2007.

21. Tabachnik, B. The speed chute.

NSCA J 14: 75-80, 1992.

22. Young, W, Benton, D, Duthie, G, and Pryor, J. Resistance

training for short sprints and maximum-speed sprints.

Strength Cond J 23(2): 7-13, 2001.

23. Zafeiridis, A, Saraslanidis, P, Monou, V, Ioakimidis, P, Dipla, K, and Kellis, S. The effects of resisted sled-pulling sprint

training on acceleration and maximum speed performance.

J Sports Med Phys Fitness 45: 284-290, 2005.

Keywords: acceleration; athletics resisted; maximum velocity; sprint; training

© 2009 National Strength and Conditioning Association