Secondary Logo

Share this article on:

Kinematics of Running at Different Slopes and Speeds

Padulo, Johnny1; Annino, Giuseppe1; Migliaccio, Gian M.2,3; D'Ottavio, Stefano1; Tihanyi, József4

Journal of Strength and Conditioning Research: May 2012 - Volume 26 - Issue 5 - p 1331–1339
doi: 10.1519/JSC.0b013e318231aafa
Original Research

Padulo, J, Annino, G, Migliaccio, GM, D'Ottavio, S, and Tihanyi, J. Kinematics of running at different slopes and speeds. J Strength Cond Res 26(5): 1331–1339, 2012—The aim of this study was to verify the influence of the combination of different running speeds and slopes based on main kinematic parameters in both groups of elite (RE) and amateur (RA) marathon runners. All subjects performed various tests on a treadmill at 0, 2, and 7% slopes at different speeds: 3.89, 4.17, 4.44, 4.72, and 5.00 m·s 1. A high speed digital camera, 210 Hz, has been used to record; Dartfish 5.5Pro has been used to perform a 2D video analysis. Step length (SL), step frequency (SF), flight time (FT), and contact time (CT) were determined and used for comparison. SL, SF, and FT parameters increased, and CT parameter decreased as speed increased. As slopes increased, SL and FT decreased and SF increased in both groups and only CT decreased in RE, whereas in RA, it increased. Data were fitted to the linear regression line (R 2 > 0.95). The 2 groups were significantly different (p < 0.05) in FT, SL, and SF at all speeds in level running. A significant difference between the 2 groups was found in FT at 2 and 7% slopes at all speeds (p < 0.05). Percentage alterations in all variables were greater in the RA group. In conclusion, the choice of optimum SL and SF, through efficient running can be maintained, is influenced not only by speed but also by slopes. Elite runners perform more efficiently than amateur runners who have less experience.

1Faculty of Medicine and Surgery, Motor Sciences, University “Tor Vergata” Rome, Italy

2LABFS, Laboratory of Sports Physiology, Department of Applied Sciences to Biosystems, University of Cagliari, Italy

3CONI, Italian Regional Olympic Committee, Sardinia, Italy

4Department of Biomechanics, Faculty of Physical Education and Sports Sciences, Semmelweis, University, Budapest, Hungary

Johnny Padulo and Giuseppe Annino contributed equally to this work.

Address correspondence to Johnny Padulo,

Back to Top | Article Outline


Biomechanical characteristics of human locomotion, that is, walking and running at different speeds and slopes have been well documented in literature together with mechanical efficiency (3). In competitions, the mechanical and physiological demands in marathon race change because of environmental and logistic variability.

In fact, elite marathon runners have similar physiological profiles with regard to maximum oxygen uptake (V[Combining Dot Above]O2max) (9,25), whereas in amateur runners, the V[Combining Dot Above]O2max is lower (6,18,32). However, various studies have shown that at constant maximum oxygen uptake, the main performance-limiting factor is the energetic cost (EC). The EC is defined by Di Prampero (10) as the metabolic demand in running while mechanical and neuromuscular profiles related to increased running speed to achieve high performance have not yet been established (21,26,32). It seems clear that finding a marathon runner–efficient biomechanical profile is an extraordinarily complex challenge. This complexity increases when analyzing the mechanical variables in races where the considered data vary with running speeds (2). Simple measures such as step length (SL) and step frequency have been investigated with conflicting results in marathon runners' performances (29).

Furthermore, other factors seem to play an important role in determining the best biomechanical profile of top-level performance in marathon races, that is, the speed and slope variations. In most studies, 1 or 2 submaximal running velocities have been analyzed.

It is well known that SL and step frequency increase as speed linearly increases at least at certain speeds (2,5,14,15,34). It is also well documented that contact time (CT) and flight time (FT) decrease with the increasing running speeds (2,5,14,15,34). However, as far as constant speed in marathon running is concerned during the race, the experienced athletes select individualized SL and step frequency, which provides the least energy cost (4,5,8,17,29) and the greatest mechanical efficiency (14,21,33); this was confirmed by Morgan's experiment (22).

Marathon races are performed mostly on flat routes, but uphill and downhill routes cannot be excluded. Interestingly, relatively few studies can be found in literature investigating the effect of uphill and downhill running from the kinematic point of view. Some authors reported a decrease in oxygen consumption in constant speed running, which was related to an increase of the treadmill negative angle (8,11,16,20). Thus, running downhill on a treadmill may be energetically efficient with an overinvolvement of the leg joint structures in the eccentric work and causes changes in kinematic variables (11,16,20). Yokozawa et al. (34) studied ground reaction forces during downhill running under at different speeds and gradients and reported altered joint kinematic patterns on increasing slopes. In the study of Gottschall and Kram (12), the subjects run at 0, 3, 6, and 9% slope at a constant velocity at 3 m·s 1. He reported a slight decrease in SL and an increase in step frequency as the slope increased (12).

We assume that kinematic parameters, and metabolic factors, may contribute to the running performance. In fact, as confirmed by Hasegawa et al. (13), foot step strategy changes at higher velocity, influencing the running economy. Although kinematic parameters can shed a light on the athlete's efficiency, no previous study has investigated marathon runners' step characteristics during uphill running at different speeds.

Moreover, a specific training on slopes to improve cardiovascular conditions and to increase strength is used by distance runner (31) without knowing the biomechanical patterns of this training tool. To try and solve the problems related to biomechanical profile of marathon runners, this study analyzes, through the kinematic and kinetic parameters, the mechanical demand of Italian national-level athletes.

Back to Top | Article Outline


Experimental Approach to the Problem

After local ethical approval, 16 marathon runners were divided in relation to 2 groups according to the best performance (PB) in last year's marathon race: 8 elite runners (RE) 2h17'25'' PB (average speed 5.11 m·s 1) and 8 amateur runners (RA) 2h48'24'' PB (average speed 4.17 m·s 1). Elite runners' age was 21.17 ± 2.32 years, body mass 63.50 ± 9.90 kg, height 175 ± 5.50 cm, BMI 20.71 ± 1.88, training background of 8 ± 1.08 years, and had covered 141 ± 7.23 kilometers per week last year. Amateur runners' age was 21.83 ± 1.33 years, body mass 70.70 ± 3.88 kg, height 171 ± 4.50 cm, BMI 24.26 ± 0.38, training background of 7 ± 0.10 years, and had covered 127 ± 2.71 kilometers per week last year. The subjects were healthy without any muscular, neurological, and tendinous injuries and did not report that they were clear of any drug. The diet control in prestudy was designed to eliminate the risk of any major differences between diets in total protein, carbohydrates, saturated, and unsaturated fats.

Back to Top | Article Outline


Both the groups are homogeneous as regard to their training status, none of the subjects underwent any endurance strenuous activity and/or resistance training outside their normal endurance training protocol. All subjects did weight training for 1 month in October.

After being informed of the procedures, methods, benefits, and possible risks involved in the study, each subject reviewed and signed an informed consent to participate in the study. The studies met the requirements of the Declaration of Helsinki, and informed consent was obtained from all participants.

Back to Top | Article Outline


Starting tests were carried out on May 29, 2010, in the Human Performance Laboratory. All the participants were in good health conditions at the time of the study, and they carried out the test during that same period. The relevant data were acquired during 7 days, starting at 4:00 PM up to 6:00 PM under the following environmental conditions: average temperature 23°C (min 20°C, max 26°C). All participants after a standardized 10-minute warm-up, consisting of run at 10 km·h 1 on treadmill (19), wore running shoes categories A3, and after 5 minutes of active muscular stretching, they performed various tests on a motorized treadmill (Run Race Technogym, Run 500; Gambettola, Italy) validated and certified at 0, 2, and 7% slope for 5 minutes for each step at constant velocity. The treadmill was calibrated before each test according to the instructions of the manufacturer and regularly checked after the tests. The experiment started at 3.9 m·s 1 in level followed by 2 slopes (2–7%); the same protocol was applied at 4.17, 4.44, 4.72, and 5 m·s 1 velocity. Each session lasted 15 minutes for 3 slopes and 1 velocity, this is line with the protocol proposed by Cavanagh and Kram (5) (e.g., 5' at 3.89 m·s 1, 0 level + 5' at 3.89 m·s 1 – 2%, + 5' at 3.89 m·s−1 – 7%; after 5' recovery the session restarted for 5' at 4.17 m·s−1 – 0 level + 5' at 4.17 m·s−1 – 2%, + 5' at 4.17 m·s−1 – 7%; after 5' recovery the session restarted for 5' at 4.44 m·s−1 – 0 level + 5' at 4.44 m·s−1 – 2%, + 5' at 4.44 m·s−1 – 7%; after 5' recovery the session restarted for 5' at 4.72 m·s−1 – 0 level + 5' at 4.72 m·s−1 – 2%, + 5' at 4.72 m·s−1 – 7%; after 5' recovery the session restarted for 5' at 5 m·s−1 – 0 level + 5' at 5 m·s−1 – 2%, + 5' at 5 m·s−1 – 7%.) total 1h15' of run. During the tests, the procedure was never interrupted and the involved players were not hurt. The subjects' run on the treadmill was recorded with a high speed camera (Casio Exilim FH20: 210 fps sampling rate) located on a 1.5-m high tripod, at 6 m from the participant, so as to be perpendicular to his sagittal plane (1). The film sequences were analyzed off-line using Dartfish 5.5Pro motion analyzing software (Dartfish, Fribourg, Switzerland).

The studied kinematical variables were CT (milliseconds), FT (milliseconds), SL (m), and step frequency (step per second) SF; for each speed in different slope, 400 steps were sampled (7,23). Because the speed of the treadmill was known, the SL and frequency were calculated as the ratio between time contact and FT of the right and left foot and cadence as the total number of right and left foot (ground contacts per second) (7). To evaluate if the sequence protocol was conditioned more by fatigue rather than a combination of slope and speed, a subsequent pilot study was carried out on 16 runners using an incremental and randomized protocol (Latin square design for 5 speeds and 3 slopes). The test-retest reliability of this testing procedure was demonstrated with an intraclass correlation coefficient (ICC) and SEM variables: (ICC: 0.97–0.98, SEM: 0.03–0.08 m in SL), CT variables (ICC: 0.96–0.98, SEM: 10–15 milliseconds), FT variables (ICC: 0.96–0.98, SEM: 10–15 milliseconds), and step frequency variables (ICC: 0.95–0.98, SEM: 0.10–0.12 Hz).

Back to Top | Article Outline

Statistical Analyses

Data are presented as the mean ± SD and primarily accomplished by using simple descriptive statistics. Linear regression analysis, using the Pearson's correlation coefficients (r), was used to indicate the strength of the relationship between slopes and (a) SL, (b) step frequency, (c) FT, and (d) CT at each velocity (i.e., 3.9, 4.17, 4.44, 4.72, and 5 m·s 1). After the assumption of normality, verified using the Kolmogorov-Smirnov test, the analysis of variance (2-way analysis of variance) was used to determine any significant difference in SL, SF, FT, and CT at each speed and slope. In case of significant F value, a Fisher's least significant difference post hoc test was applied. The level set for significance was p ≤0.05. The statistical analyses were performed using SPSS software (version 15; SPSS, Inc, Chicago, IL, USA).

Back to Top | Article Outline


Relationship Between Running Speed and Kinematic Variables

Means and SD are presented in Table 1 for each variable in each experimental condition. SL, SF increased, and CT decreased with the increasing running speed at all grades for both groups. FT increased in level running and decreased at 2 and 7% grades linearly as the speed increased. Significant differences were revealed between level running and 2 and 7% grade running at each speed. No significant difference was found between 2 and 7% grade running both in groups RE or RA. The relationship between the different constant running speed and the mean SL, SF, CT, and FT was significant, and the data could be fitted to the linear regression line (R 2 ranged between 0.95 and 0.99). The means of SL, CT, and FT were less at 2 and 7% grades compared with the means at level running. On the contrarily, SF was elevated at 2 and 7% grade at each running speed applied compared with 0% grade (Table 1). The percentile change because of the increasing grade was different with regard to the studied variables and in comparison to the 2 groups indicated in Figure 1.

Table 1

Table 1

Figure 1

Figure 1

Back to Top | Article Outline

Step Length

In RE group, the SL decreased linearly when speed increased: 4.4, 6.2, 7.9, 10.6, and 14.0% comparing 0 and 2% grades. SL in RA group also decreased linearly, but with lesser percentile change, that is, 0.4, 2.5, 4.7, 7.9, and 10.4%. The difference between the 2 groups was significant at 3.89 and 4.17 m·s 1, only. Comparing the changes at 0 and 7%, the decrease was 8.0, 10.2, 12.4, 15.6, and 18.7% for RE and 3.5, 5.1, 6.5, 10.2, and 14.5% for RA. Comparing the 2 groups, only RE showed significantly greater change at 3.89, 4.17, and 4.44 5 m·s 1.

Back to Top | Article Outline

Step Frequency

The percentile change in step frequency increased for both groups. However, the change was greater for RE (4.1, 5.8, 7.2, 9.5, and 12.1%) than for RA (0.4, 2.4, 4.5, 7.3, and 9.5%) at 0 and 2%. The difference between the 2 groups was significant at 3.89 and 4.17 m·s 1 only. Comparing 0 and 7% running grade, the percentile change was greater for RE (7.3, 9.2, 11.0, 13.3, and 15.6%) than for RA (1.4, 3.3, 5.0, 8.5, and 11.8%) and the difference between the 2 groups was significant at 3.89, 4.17, and 4.44 m·s 1.

Back to Top | Article Outline

Contact Time

In RE group, the percentile change of CT was less pronounced compared with SL and SF at both 0–2% grade (4.4, 3.9, 4.3, 4.8, and 6.1%) and 0–7% grade (3.3, 1.9, 2.8, 3.2, and 3.7%). In RA group, the percentile alteration of CT in 0–2% was positive (0.6, 1.5, 1.2, 1.2, and 2.1%) similar to RE, indicating shorter CT at 2% grade than at 0%, increasing the gap between the groups (Figure 2). In fact, the difference between the 2 groups in mean percentile changes was not significant with 2 exceptions. Namely, the increase in CT was significantly greater for RA than the decrease for RE running on 7% grade at the speed of 4.44 and 4.72 m·s 1.

Figure 2

Figure 2

Back to Top | Article Outline

Flight Time

Fight time decreased almost with the same percentage in RE (0–2%: 4.3, 8.3, 11.1, 15.8, and 20.9%; 0–7%: 12.7, 18.6, 22.0, 27.7, and 33.3%) and in RA (0–2%: 3.5, 6.6, 11.5, 17.9, and 21.8%; 0–7%: 13.5,16.6, 21.3, 29.7, and 35.8%). There are significant differences between percentile changes at 0–2% and 0–7% in both groups. A significant difference was found between the 2 groups.

Back to Top | Article Outline


The study was aimed at focusing the footstep characteristic kinematic variables, that is, SL, step frequency (SF), CT, and FT, at 5 running speeds and considering 3 different slopes, that is, 0, 2, and 7%, in amateur and elite marathon runners.

The rationales carrying out the study were that marathon runners perform the distance on road with altering grades, and the speed has to be changed according to altered in a function of the different circumstances, which have not yet been considered in previous laboratory studies. It is expected that the international level athletes' running pattern (technique) is better adapted to the need of efficient work output than that of amateur runners who would react differently to the altered circumstances. The data of the present study demonstrated that the increased speed and slope gradient elicited distinct changes in SL, step frequency, CT, and FT, showing marked differences between elite and amateur marathon runners.

As it was expected, the SL, SF, FT increased, and CT decreased with the increased running speed in both groups. The greatest alteration occurred in SL and CT, indicating the significance of these 2 kinematic variables at increasing running speeds. Because there is no significant difference in the percentile alterations between the 2 groups, except for a longer FT and a higher SF found in RE (Figure 1), we may assume that the difference in marathon running performance could be attributed to the difference in power output rather than to the difference in kinematic pattern.

Similar to level running, SL and SF increased and CT decreased when running speed increased. In contrast to level running, FT decreased, more in RA than in RE, when running speed increased because of the belt inclination (i.e., the foot touches earlier the belt than in level running). At all speeds, the means of all variables were less than those of level running. The minimum alteration occurred in CT and the maximal in FT; we observed at the same time an increase of the gap in both parameters between RE and RA (Figure 1).

However, RE decreased in SL and CT and increased SF. In contrast, the RA maintained CT unchanged compared with the increased slope using a higher value with respect to RE. Probably, at 2% uphill slope, elite and amateurs runners adopt different mechanical strategies to increase velocity. In fact, while in RA the velocity increase was because of a higher as SF, RE achieve the same goal by increasing SL-SF and by reducing the CT, thus allowing them to run more efficiently (Figure 2C). However, the difference in percentile alterations between the 2 groups was significant in SL and SF at the 2 highest running speeds, whereas no difference was observed at a lower speed. On the contrary, considering the absolute value, the difference between RE and RA regarded the FT only because the gap in SL and SF observed at level in both groups is minimal. It seems that the RE group uses a biomechanical strategy aimed at better harmonize the SL, step frequency, CT, and FT during faster running.

In both groups, SL and SF increased, whereas CT and FT decreased when increasing speed. The means of all variables at each speed were less than at level or 2% slope running in both groups, except for the CT in RA that was longer than during level or 2% slope running. In both groups, the difference between 2 and 7% uphill running with respect to CT variable was not significant, whereas their gap increased further (Figure 1).

The means of SL decreased and SF increased, CT remained unchanged in RE. In contrast, in RA SL, SF and FT were almost the same as at 2% slope and CT was longer than that during level or 2% uphill running. The percentile change difference between the 2 groups was significant in SL, SF, FT, and CT at higher speeds. The gap between the 2 groups was reduced as to SL and SF while an increase in CT and FT was found. This result may indicate a better adaptation to the increased load for elite marathon runners, that is, the higher speed and greater incline for RE was not so unusual as for RA. Comparing RE and RA reactions with the altered slopes, it seems that RE were able to adapt better than RA. This assumption was based on variables' alteration in both groups because of the altered gradient. A significant change was found in SL and SF when increasing the gradient up 2% at all speeds in RE. RA did not significantly change in SL and SF at 3.89, 4.17, and 4.44 m·s 1 speed. Significant change occurred only at the 2 highest speeds. Therefore, the gradient increase did not result in significant change in either group except for the FT that showed a significant difference between the 2 groups and decreased significantly because of slope increase. This result may indicate that the alteration in FT is not only influenced by the gradient degree. In fact, considering the velocity increase, the longer FT observed in RE at each speed at level and on slope, with respect to RA, could be related to a more efficient propulsive phase performed in a shorter CT.

Interestingly, although minor changes were observed in the variables' means when increasing slope from 2 to 7%, the differences were not significant, indicating that efficient adaptation to 7% gradient is not running performance dependent. Considering that every speed needs optimum SL and frequency, we may conclude that RE are more capable than RA to find the optimum combination of SL and step frequency through a better propulsive phase during CT generating a longer FT, providing them a most efficient movement pattern (Figure 2).

Reviewing the relevant literature, the effect of running speed and uphill/downhill gradient on basic kinematic variables was studied from several points of view; however, none of the studies investigated uphill running with different gradients combined with different constant running speeds. During the last 50 years, authors reported that SL, step frequency and FT increase and CT decreases when increasing speed in level running (2,5,14,34). Högberg concluded that every speed needs optimum SL and step frequency to meet economical and biomechanical to meet requirements. Our results support this opinion, and it seems that it is true for uphill running too (14). In a unique experiment, Cavanagh and Williams estimated the optimum SL on the basis of oxygen uptake at a speed of 3.83 m·s 1 (8). They reported that the freely chosen SL (132.1 cm) was longer with 4.2 cm than that estimated (8). It is well demonstrated in the literature that to increase the running speed, the oxygen uptake should also be increased to provide energy to enhance muscular work. Because the running speed can be elevated by increasing the SL, it can be expected that optimum SL belongs to each speed. Also, it could be expected that elite runners can consume less energy to run with the same speed, and consequently, they use shorter SL relative to body size than amateur runners. Indeed, Cavanagh et al. reported 5.1% shorter SL for elite runners compared with amateur runners applying the same speed (7). The step frequency was higher and the FT longer. No significant difference was reported in CT.

In contrast, in our experiment, the elite runners had significantly longer SL, FT, and higher step frequency at 4.44, 4.72, and 5 m·s 1 compared with amateur runners. These results suggest that SL cannot be a differential factor alone between amateur without considering the step frequency also. In effect, this study showed that RA adapt their neuromechanical strategy to increase velocity at level and at 7% slope, increasing the step frequency (Figure 2B). As to uphill running, there are a few studies to be compared, and in these experiments, researchers applied only one speed for uphill slopes. Swanson and Caldwell (28), applying 4.5 m·s 1 speed at level and 30% grade running, reported SL reduction and step frequency increase. Similarly, Gottschall and Kram (12) found SL decrease and step frequency elevation at 3 m·s 1 speed and 3, 6, and 9% inclination with no change in CT. The results of the present study support the result of the aforementioned experiments. Although the SL, step frequency, and FT increased and CT decreased when increasing running speed, a significant reduction in SL was revealed when the inclination was increased to 2 and 7% with increasing step frequency at each speed (Figure 2A). All changes because of inclination increase are related to increased demand of work energy (24). Perhaps, the alterations trend of is similar for each runner, but the alteration ratio depends on the athletes training status and the individual runner motion efficiency.

Definitely, during uphill running, the oxygen uptake should be increased. It is known that when increasing the running speed, the oxygen uptake is greater, and as a consequence, the SL should be increased (2). It seems that the increased step frequency and decreased CT can compensate the reduced SL because of the increased slope when runners have to maintain the increased constant velocity on the treadmill. Considering that training on slope is used by trainers to increase performance (27), these data suggest that using slopes around or more than 7% could alter, in a considerable way, the neuromechanical efficiency of athletes, whereas using slopes around 2% could influence positively the performance of the marathon runner (30). In fact, at constant speed on slopes, there is an increase of the propulsive phase and a decrease of ground reaction force (12); this is higher in RE than RA groups and produces a considerable SF increase and an SL decrease. In addition, increasing the slope could depress the best quality of RE represented by the capability of using the ground reaction force to increase SL and FT (Figure 2D) in a shorter stretch-shortening cycle. The question on how long the optimal SL in altered conditions should be remains still to be studied.

Back to Top | Article Outline

Practical Applications

Our results may allow us to conclude that not only running speed but also gradient influence the choice of optimum SL and frequency with which efficient running can be maintained. It seems that the optimum speed and gradient variations depend on marathon runners' training status. Namely, elite runners can adapt to alter circumstances more efficiently than amateur runners who have less experience. Whereas in a marathon race, the average number of steps in RE is 27,786 vs 52,661 in RA (calculated in relation to their race performance 5.11 vs. 4.17 m·s 1 and to step frequency). It is necessary for the coach to know the athlete's SL decrease on different slopes at constant speed, to estimate his efficiency level. For example, knowing the speed of the treadmill (the best performance marathon race speed) at the ground level and the number of steps per minute, the SL can be calculated (with a good approximation) with Equation 1. In addition, Table 1 could allow coaches to find the correct position of their athletes by knowing their SL both at level and on slope.

Furthermore, the coaches could train their athletes, by changing slope and step frequency on a treadmill. For example, knowing the speed and SL in level and slope, one can find the optimal frequency on slope through Training Equation 2. In this case, having defined slope and speed, the athlete shall be trained to perform a precise number of steps per minute so as to generate the desired SL.

Back to Top | Article Outline


The authors are grateful to the “Gruppo Sportivo Fiamme Gialle” in particular to Coachs: Andrea Ceccarelli and Nardino Degortes for supporting this research; Davide Viggiano PhD MD for the review, Dott. Napoleone Gasparo, Leonardo Beneduci and Maria Di Filippo for encouragement. The project was supported by Grant from CONI Italian Regional Olympic Committee, Sardinia, Italy.

Back to Top | Article Outline


1. Belli A, Rey S, Bonnefoy R, Lacour JR. A Simple device for kinematic measurements of human movement. Ergonomics 35: 177–186, 1992.
2. Brisswalter J, Legros P. Use of energy cost and variability in stride length to assess an optimal running adaptation. Percept Mot Skills 80: 99–104, 1995.
3. Cavagna GA, Thys H, Zamboni A. The sources of external work in level walking and running. J Physiol 262: 639–657, 1976.
4. Cavagna GA, Willems PA, Franzetti P, Detrembleur C. The two power limits conditioning step frequency in human running. J Physiol 437: 95–108, 1991.
5. Cavanagh PR, Kram R. Stride length in distance running: Velocity, body dimensions, and added mass effects. Med Sci Sports Exerc 21: 467–479, 1989.
6. Cavanagh PR, Lafortune MA. Ground reaction forces in distance running. J Biomech 13: 397–406, 1980.
7. Cavanagh PR, Pollock ML, Landa J. A biomechanical comparison of elite and good distance runners. Ann N Y Acad Sci 301: 328–345, 1977.
8. Cavanagh PR, Williams KR. The effect of stride length variation on oxygen uptake during distance running. Med Sci Sports Exerc 14: 30–35, 1982.
9. Costill DL. The relationship between selected physiological variables and distance running performance. J Sports Med Phys Fitness 7: 61–66, 1967.
10. Di Prampero PE. The energy cost of human locomotion on land and in water. Int J Sports Med 7: 55–72, 1986.
11. Gehlsen GM, Stewart LB, Van NC, Bratz JS. Knee kinematics: The effects of running on cambers. Med Sci Sports Exerc 21: 463–466, 1989.
12. Gottschall JS, Kram R. Ground reaction forces during downhill and uphill running. J Biomech 38: 445–452, 2005.
13. Hasegawa H, Yamauchi T, Kraemer WJ. Foot strike patterns of runners at the 15-km point during an elite-level half marathon. J Strength Cond Res 21: 888–893, 2007.
14. Högberg P. How do stride length and stride frequency influence the energy-output during running? Arbeitsphysiologie 14: 437–441, 1952.
15. Ito A, Komi PV, Sjodin B, Bosco C, Karlsson J. Mechanical efficiency of positive work in running at different speeds. Med Sci Sports Exerc 15: 299–308, 1983.
16. Iversen JR, McMahon TA. Running on an incline. J Biomech Eng 114: 435–441, 1992.
17. Knuttgen HG. Oxygen uptake and pulse rate while running with undetermined and determined stride lengths at different speeds. Acta Physiol Scand 52: 366–371, 1961.
18. Lake MJ, Cavanagh PR. Six weeks of training does not change running mechanics or improve running economy. Med Sci Sports Exerc 28: 860–869, 1996.
19. Matsas A, Taylor N, McBurney H. Knee joint kinematics from familiarised treadmill walking can be generalised to overground walking in young unimpaired subjects. Gait Posture 11: 46–53, 2000.
20. Milliron MJ, Cavanagh PR. Sagittal plane kinematics of the lower extremity during distance running. Biomech Distance Running 65–105, 1990.
21. Morgan D, Martin P, Craib M, Caruso C, Clifton R, Hopewell R. Effect of step length optimization on the aerobic demand of running. J Appl Physiol 77: 245–251, 1994.
22. Morgan DW, Martin PE, Krahenbuhl GS. Factors affecting running economy. Sports Med 7: 310–330, 1989.
23. Owings TM, Grabiner MD. Measuring step kinematic variability on an instrumented treadmill: How many steps are enough? J Biomech 36: 1215–1218, 2003.
24. Roberts TJ, Belliveau RA. Sources of mechanical power for uphill running in humans. J Exp Biol 208: 1963–1970, 2005.
25. Saltin B, Astrand PO. Maximal oxygen uptake in athletes. J Appl Physiol 23: 353–358, 1967.
26. Saunders PU, Pyne DB, Telford RD, Hawley JA. Factors affecting running economy in trained distance runners. Sports Med 34: 465–485, 2004.
27. Staab JS, Agnew JW, Siconolfi SF. Metabolic and performance responses to uphill and downhill running in distance runners. Med Sci Sports Exerc 24: 124–127, 1992.
28. Swanson SC, Caldwell GE. An integrated biomechanical analysis of high speed incline and level treadmill running. Med Sci Sports Exerc 32: 1146–1155, 2000.
29. Taylor CR. Relating mechanics and energetics during exercise. Adv Vet Sci Comp Med 38A: 181–215, 1994.
30. Telhan G, Franz JR, Dicharry J, Wilder RP, Riley PO, Kerrigan DC. Lower limb joint kinetics during moderately sloped running. J Athl Train 45: 16–21, 2010.
31. Tulloh B. The role of cross-country in the development of a runner. New Stud Athl 13: 9–11, 1998.
32. Williams KR. Biomechanical factors contributing to marathon race success. Sports Med 37: 420–423, 2007.
33. Williams KR, Cavanagh PR, Ziff JL. Biomechanical studies of elite female distance runners. Int J Sports Med 8(suppl 2): 107–118, 1987.
34. Yokozawa T, Fujii N, Ae M. Kinetic Characteristics of Distance Running on Downhill Slope. Int. Journ. of Sport and Healt Science 3: 35–45, 2005.

uphill running; locomotion; kinematics

© 2012 National Strength and Conditioning Association