Modern tennis is a complex sport and competitive success requires excellent technical skills (e.g., stroke power and precision) and an optimal physiological condition (i.e., medium to high aerobic fitness), among other abilities (23,28,32,36). The technical skill of a tennis player can have a direct influence on the final outcome of a match. For example, the player with the fewest unforced errors has a higher probability of winning a match (15). In this regard, different studies have found relationships between technical skills (i.e., technical effectiveness, success rate, ball velocity, and precision of ball placement) and competitive performance at a range of competitive levels (4,39–41). However, given that fatigue reduces the performance of tennis players (e.g., technical skill performance by reducing ball velocity and hitting accuracy) (20,21,28,35,41) and that during tennis competition incomplete physiological recovery between points and matches occur (5), a good aerobic fitness level is recommended to avoid fatigue and ensure optimal recovery between points and matches (31). Thus, maximal oxygen uptake (V̇o2max) values > 50 ml·kg−1·min−1 have been considered optimal for competing at a high level (4,5,22,32), and competitive tennis players with higher aerobic fitness levels have been shown to be able to compete at lower intensities and fatigue levels (2).
One of the most practical and widely used on-field workload parameters for exercise training prescription is maximal aerobic velocity (MAV or vV̇o2max) or maximal aerobic speed (MAS). Maximal aerobic speed (MAS) is the minimum speed at which V̇o2max occurs during a graded exercise test (6,9,37) and incorporates both V̇o2max and running economy into one term (10,37). In running, the Université de Montréal Track Test (UMTT) (34) or short-distance time trials (e.g., 2,000 m) (6) may be used to predict MAS. This parameter has been widely used in continuous endurance sports (e.g., long-distance running) to determine different training intensity zones (8,9) allowing to use both continuous and interval training methods (6). In addition, MAS was related to the velocity sustained in middle and long-distance running (e.g., 1,500 m, 5,000 m, 10,000 m, half-marathon, and marathon) and has been proposed as a valuable information to analyze the runner's performance (33). However, because tennis requires players to accelerate, decelerate, change direction, and hit the ball using a proper technique, MAS is not an adequate training prescription parameter in this sport.
Up to now, different tennis-specific incremental field tests that consider the biomechanical specificity required in tennis have been proposed. This specific test can be acoustically controlled with real backhand and forehand stroke simulations (26,27) and can be conducted with a tennis ball machine (4,39). Among these, the SET-Test includes some elements of tennis (e.g., alternating backhands and forehands strokes in a prescribed pattern) and is based on the frequency of balls (Ballf) ejected from a tennis ball machine, that is the number of balls ejected by the ball machine per minute (shots·min−1). This new training load parameter makes possible to simultaneously evaluate technical parameters (e.g., % of successful hits) and physiological parameters (e.g., V̇o2max, ventilatory thresholds, and heart rate deflection point) (3,4). Because this specific endurance tennis test allows assessing V̇o2max using the Ballf to set the increasing workload, and its average duration (∼12 ± 1 minute) lies within the range considered appropriate for V̇o2max assessment (1,16), one could suggest that the Ballf at which V̇o2max occurs during the SET-Test would represent the maximal aerobic workload (“maximal aerobic frequency” of balls, or MAF) and could constitute a new specific training load parameter in tennis. That is to say, we propose to replace the running velocity with the Ballf ejected from a tennis ball machine, and this new parameter could be used by tennis coaches as a practical specific training load parameter to prescribe training intensity zones in terms of V̇o2max percentages to optimize both aerobic condition and technical parameters. Like the methodology used in the UMTT to provide an indirect value of MAS (33,34), we hypothesized that a Ballf sustained for 2 minutes below the last completed stage during the SET-Test could be identified as the workload eliciting V̇o2max and, consequently, be considered as the MAF. Therefore, the aims of this study were (a) to evaluate a new training load parameter in tennis based on the ball frequency at which V̇o2max occurs, and (b) to determine the accuracy of the SET-Test for predicting the MAF in competitive tennis players.
Experimental Approach to the Problem
A group of high-level tennis players performed a modified version (4) of a progressive, tennis-specific field test (39), which was recently shown to be reliable (4). To validate the capacity of the SET-Test for estimating the MAF, physiological parameters (V̇o2, V̇co2, V̇E, and HR) and performance parameters (Ballf, time, stage achieved, and hits per test) were compared at MAF, last 2-min completed stage (LS), and maximal workload (MAX).
Thirty-five male competitive tennis players (mean ± SD: age, 18.2 ± 1.3 years; height, 180.5 ± 8.1 cm; body mass, 73.1 ± 8.7 kg; body mass index, 22.4 ± 1.4) with an International Tennis Number (ITN) ranging from 1 (elite) to 4 (advanced): ITN 1 = 7 players (corresponding to an Association of Professional Tennis Players ranking between 600 and 1,000); ITN 2 = 9 players; ITN 3 = 9 players; and ITN 4 = 10 players, volunteered to participate in this study. All subjects completed a short questionnaire to document their training background, the average training background of the players was 6.6 ± 2.0 years, during which time they focused on tennis-specific training (i.e., technical and tactical skills), aerobic and anaerobic training (i.e., on-court and off-court exercises), and strength training. Before their participation, the experimental procedures and potential risks were fully explained to the subjects, and they all provided written informed consent. The local university's research committee approved the study. The study was conformed to the recommendations of the Declaration of Helsinki. Parental written informed consent was obtained for the 5 subjects under 18 years of age.
Players were tested during noncompetition periods between February and April, and all tests were performed on an outdoor tennis court (GreenSet surface). The temperature ranged from 18° C to 23° C with stable environmental and wind conditions (air velocity, <2 m·s−1; relative humidity, 54–61% [Kestrel 4,000 Pocket Weather Tracker; Nielsen Kellerman, Boothwyn, PA, USA]). Before each test, all participants completed a 18-min standardized warm-up session consisting of 2 differentiated parts: (a) their usual daily warm-up consisting of 10 minutes of jogging round the court, dynamic flexibility, forward, sideways and backwards running, and acceleration runs; 5 minutes of ground strokes (players were asked to hit the balls to the center of the court); and (b) a specific test warm-up consisting of 3 minutes of test familiarization during which they performed the test protocol at the lowest work load (frequency of balls ejected from the ball machine (Ballf = 9 shots·min−1). Three to 5 minutes after the warm-up, the test began. Before baseline testing and within 5 days before the SET-Test, all participants attended 2 familiarization sessions to learn about the testing procedures and ensure that any learning effect was minimized in the study measurements. The subjects were instructed to maintain their normal lifestyle and dietary intake before and during the study. The subjects were told not to exercise the day before a test and to consume their last (caffeine-free) meal at least 3 hours before the scheduled test time.
Specific Endurance Tennis Test
A modified version of an incremental specific test conducted with a tennis ball machine and performed on a tennis court was used (39), the original test began with a Ballf of 12 shots·min−1 and increased by shots·min−1 2 every 3 minutes. To limit the duration of the test within limits favoring the determination of main cardiorespiratory parameters (i.e., V̇o2max and VTs), the SET-Test includes shorter stages (2 instead of 3 minutes) and begins with a lower Ballf (9 instead of 12 shots·min−1). Participants had to hit balls ejected from a ball machine (Pop-Lob Airmatic 104; Paris, France), starting on the right corner of the baseline (i.e., right-handed players starts with a forehand and left-handed starts with a backhand). Subjects had to hit the balls alternating between forehand and backhand and they could choose between cross-court or down the line in a prescribed pattern (e.g., drive, topspin). The chosen landing point for the balls was about 2 m in front of the baseline, and the balls were alternated to the right and the left corners (Figure 1). Slice strokes were not allowed because we assumed they would influence the positioning of the ball and, therefore, the physiological responses and test reliability (4). The test began with a Ballf of 9 shots·min−1, which was increased by 2 shots·min−1 every 2 minutes. The test ended at the player's request or was stopped by the researchers if the player was no longer able to fulfill the test criteria (i.e., he was no longer able to perform strokes with acceptable technique and precision), as determined by 1 single researcher through observation, and technical scores (i.e., hits-errors) were registered by a single experienced coach. In this regard, we acknowledge that there may be variability in the testing protocol depending on the coaches at hand. The ball machine was manually calibrated before each test and the ball velocity stayed constant during the test. The reliability of the device was assessed by manual timing (mean CV of Ballf = 3.5 ± 0.9%) and the use of a radar device (Stalker ATS 4.02; Plano, TX, USA) (mean Ballv = 68.6 ± 1.9 km·h−1; CV = 2.7%). A minimum of 40 new tennis balls (Babolat Team) was used for each test.
The performance measurements considered were (a) the test duration until the player felt exhausted or failed to hit the ball twice in a row (Time), (b) the stage with a precision of 0.5 stages (i.e., considering the work rate sustained for at least 1 minute), (c) the total number of hits per test (Hits), and (d) the frequency of hitting balls (Ballf). Considering that performance variables were registered at 3 experimental time points: MAF, LS (work rate sustained for at least 1 minute at the final stage), and MAX (maximal workload or failure to hit the ball twice in a row). Thus, the variables obtained were Time (TimeMAF, TimeLS, and TimeMAX), Stage (StageMAF, StageLS, and StageMAX), Hits per test (HitsMAF, HitsLS, and HitsMAX), and Ballf (Ballf-MAF, Ballf-LS, and Ballf-MAX).
Ventilatory gas exchange and HR were continuously recorded, beginning 2 minutes before the familiarization phase and finishing 5 minutes after the end of the test (recovery phase). The HR was recorded at 15-s intervals. Expired air was analyzed continuously for gas volume (Triple digital-V1 turbine), oxygen concentration (zirconium analyzer), and carbon dioxide concentration (infrared analyzer) using a portable gas analyzer (K4b2; Cosmed, Rome, Italy). The portable measurement and transmission units were carried at the back of the player in the same way throughout all tests. The HR monitoring (Polar, Kempele, Finland) was used alongside the Cosmed K4b2 system. Gas and volume calibration of the measurement device was carried out on the morning of each test session. Ambient air calibration was performed before each test. The V̇o2max was determined by the highest 15-s average value for V̇o2 (1,11,12), through observation of a “plateau” or leveling off in V̇o2, or when the increase in V̇o2 in 2 successive stages was less than 150 ml·min−1 (2.1 ml·kg−1·min−1) (11–13,42). V̇o2 values exceeding 3 SD of the local mean were considered as aberrant values and eliminated (29). The HRmax was considered the highest value reached during the final 15 seconds of the test (27).
Physiological variables were registered at MAF, LS, and MAX (Figure 2). Maximal aerobic frequency (i.e., Ballf at V̇o2max in shots·min−1) was considered as the lowest Ballf eliciting a 15-s average V̇o2 value equal to V̇o2max, as long as this frequency was sustained for at least 1 minute, according to the criteria proposed by Billat et al. (11,12). Therefore, the physiological variables considered were HR (HRMAX, HRLS, and HRMAF), V̇o2 (V̇o2max, V̇o2-LS, and V̇o2-MAF), V̇co2 (V̇co2-max, V̇co2-LS, and V̇co2-MAF) and V̇E (V̇E-MAX, V̇E-LS, and V̇E-MAF).
The values presented are expressed as mean ± SD and 95% confidence intervals (95% CI). The normality of variable distribution was assessed by the Kolmogorov-Smirnov test. One-way analysis of variance and pairwise multiple post hoc comparisons (Tukey's honest significant difference) was used to test the significance of the differences between the performance and physiological variables corresponding to MAX, LS, and MAF. The Pearson product-moment correlation coefficient (r) checked for relationship between variables. When appropriate, a linear regression analysis was performed and the standard error of the estimation (SEE) was calculated. The magnitude of the differences in mean was quantified as effect size and interpreted according to the criteria used by Cohen (19): <0.2, trivial; 0.2–0.4, small; 0.5–0.7, moderate; >0.7, large. Mean differences in absolute and percent values and Bland-Altman difference plots for method agreement (14) were also used. Two-tailed statistical significance was set in advance at p ≤ 0.05. All statistical analyses were performed using SPSS for Windows 15.0 (SPSS Inc., Chicago, IL, USA).
Table 1 shows the data obtained for all selected physiological and performance variables at the 3 established reference work rates (MAX, LS, and MAF). The values of performance parameters at LS (Ballf-LS, StageLS, TimeLS, and HitsLS) were higher than at MAF (p < 0.001) for all variables, with large effect sizes (1.20–1.39) and highly correlated (r = 0.72–0.77; p < 0.001).
Figure 3 shows the linear regression (a) and difference plot (B) between Ballf-LS and Ballf-MAF. The mean difference between these 2 performance variables was 2.1 shots·min−1 (95% CI: 1.7–2.5) corresponding to 1.1 stages (95% CI: 0.9–1.2).
As shown in Table 1, no significant differences were found between the physiological parameters at LS and MAF, except for V̇E (p < 0.001; ES = 1.09) and V̇co2 (p = 0.046; ES = 0.58). HR did not reach, but was close to, the preestablished significance level (p = 0.056; ES = 0.53). This physiological parameters (HR, V̇co2, V̇E, and V̇o2) were all strongly correlated, especially V̇o2 (r = 0.99; p < 0.005). Figure 4 shows the linear regression (a) and difference plot (b) between V̇o2-LS and V̇o2-MAF; the slope of the regression line was not significantly different from the line of identity, and the mean difference between these 2 parameters was less than 1% (95% CI: 0.3–1.3). Values of all performance (p < 0.001) (Ballf-MAX, StageMAX, TimeMAX, HitsMAX) and physiological (p < 0.05–0.001) (HRMAX, V̇co2-max, and V̇E-MAX) parameters at MAX were higher than the corresponding values at MAF (Table 1) except for V̇o2 both in absolute values and relative to body mass (0.968 and 0.920, respectively). No significant differences were found between the physiological and performance variables at MAX compared with LS.
This study aimed at evaluating a new specific training load parameter in tennis based on the Ballf at which V̇o2max occurs (MAF) during the SET-Test, and to determine if this incremental tennis-specific test would provide an accurate indirect assessment of MAF in competitive players. The present findings support the concept that MAF can be used as a practical performance parameter to prescribe specific on-court training, and that the SET-Test can be considered a valid method for determining MAF.
Most training conducted by tennis players is based on technique and tactics and is carried out on the court itself. Coaches would therefore find it very helpful if they could incorporate physiological and technical aspects into the same training session (24,38). However, there is currently no load parameter in tennis that incorporates both physiological load and technique, and coaches have no objective criteria concerning exercise duration (number of hits per work period), intensity, or volume (number of hits per exercise drill) to work on different physiological intensity zones (38). Using the Ballf (shots·min−1) as a training load parameter makes it possible to integrate both technical training stimuli (e.g., precision, depth, and direction of forehand and backhand strokes) and physiological training stimuli (e.g., %V̇o2max), and control the training load (e.g., number of hits per work period) more effectively (3,4).
To our knowledge, the present study is the first to adapt the MAS concept developed by Lacour et al. (33) to the specific needs of tennis. We reasoned that, in the same way that MAS can be estimated based on performance in an incremental running test (6,9,34), so could MAF could be predicted based on a specific progressive tennis test performed with sufficient accuracy. To apply it on the field, we replaced running velocity with the Ballf ejected from a tennis ball machine. Maximal aerobic frequency can be measured directly on the field using the criteria proposed by Billat et al. (11,12) (Figure 2), but this involves the use of portable gas analyzers, which are too expensive and complicated for most coaches. To validate the SET-Test for estimating MAF, we compared values measured directly on the field using a portable gas analyzer at MAF (i.e., when V̇o2max was first detected), at the time of maximal work rate during the test (MAX), and at the last 1-minute completed stage (LS). Our results show that the LS achieved through the test overestimated MAF by 2.1 ± 1.1 shots·min−1, and 1.1 ± 0.6 stages (Table 1). In fact, despite the trivial differences between V̇o2-LS and V̇o2-MAF (95% CI: 0.3–1.3%), the moderate and large significant differences observed between V̇co2-LS and V̇E-LS, and V̇co2-MAF and V̇E-MAF (95% CI: 9–14% and 12–20%, respectively) suggest that the anaerobic metabolism contributes substantially to energy supply when players reach MAF, and confirms the maximal nature of the test. Contrary to these results, it has been found that the velocity corresponding to the LS during an UMTT and the MAS determined on a treadmill were not significantly different (7,8,33), and the UMTT considers MAS as the maximal speed sustained for 2 minutes during the incremental running protocol (10). These differences are probably linked to the specific coordination and agility required by the SET-Test (i.e., acceleration, deceleration, stopping, and changing direction over short distances, time to react to the ball ejected from the machine, and technical stroke movement). In this regard, it has been reported that running with interruptions and changes in direction could be associated with a greater physiological load than continuous straight-line running as indicated by greater cardiorespiratory response, muscular V̇o2, blood lactate concentration, and rating of perceived exertion (17).
The usefulness of the SET-Test when gas analysis is not available relies on the finding that players reach their V̇o2max (expressed by ml·kg−1·min−1) during the stage previous to LS (i.e., LS minus 1), which provides a good estimate of MAF (expressed by shots·min−1) (Figure 4). There was a fairly constant difference between Ballf-LS and Ballf-MAF, which represented a difference of approximately 1 stage (2 shots·min−1) (Figure 3). Thus, as a practical rule, MAF is reached at ∼2 shots·min−1 below the LS during the SET-Test. Once identified, the workload corresponding to MAF can be incorporated as a practical, specific training load parameter to determine optimal training intensity to improve V̇o2max and thus the aerobic capacity of tennis players.
As previously discussed (Methods), the technical scores and the criteria to fulfill the test were determined through observation by one single researcher and one single experienced coach, respectively. This fact could have introduced an unquantifiable interobserver variability in the testing results. Although we can speculate that the effect on the physiological and performance variables during the SET-Test is likely to be quantitatively small, this should be considered as a limitation of the study. When possible, particularly in longitudinal studies or evaluations, the same experienced observers should be appointed.
In conclusion, this study shows that the proposed training load parameter (i.e., MAF) corresponds to the intensity at which V̇o2max is attained during an incremental tennis-specific test and can be used to prescribe specific on-court training that incorporates both technical and physiological loads. The SET-Test seems to be a valid method for assessing MAF based on the estimation that the players reach their MAF during ∼1 stage (∼2 shots·min−1) before the last completed stage attained.
Aerobic fitness parameters are considered important training tools in tennis (22,30,32), and to optimize improvements in the aerobic condition, the individualization of intensity is a key factor. Because MAF can be expressed as Ballf, it may serve as a tool for coaches to define training intensity zones (e.g., %MAF). Table 2 shows a proposal for specific interval training in tennis by using the percentages of MAF (i.e., Ballf at V̇o2max) and first and second ventilatory thresholds (i.e., Ballf at VT1 and VT2) as reference intensity parameters, based on data reported by Baiget et al. (4), on the recommendations for the design of high-intensity protocols (18), and in specific interval training sessions reported in tennis players (24,25). It also provides the number of hits per work period as the work duration, and the total number of strokes per exercise as the volume parameter. Additionally, the SET-Test protocol involves alternation of forehand and backhand strokes with lateral displacements, and makes it possible to control technical parameters based on the precision and power of hits (e.g., percentage of successful hits or technical effectiveness) (4). The workload parameters were proposed based on the data obtained in competitive tennis players (i.e., ITN ranging from 1 [elite] to 4 [advanced]), and it should be adapted to the specific characteristics of tennis players (e.g., age, competitive level, or training background).
This study was supported by the Institut Nacional d’Educació Física de Catalunya (INEFC), Generalitat de Catalunya. The authors thank the Sánchez-Casal Academy, the Bruguera Tennis Academy Top Team, the Escola Balear de l’Esport and the Catalan Tennis Federation Centre Internacional de Tennis for their support and use of their facilities, and the players for their time and effort throughout the study. We also thank Lisímaco Vallejo, Pedro Zierof and Valery Kryvaruchka for their technical assistance during the experiments.
1. Astorino TA, Rietschel JC, Tam PA, Taylor K, Johnson SM, Freedman TP, Sakarya CE. Reinvestigation of optimal duration of VO2max
testing. J Exerc Physiol 7: 1–8, 2004.
2. Baiget E, Fernández-Fernández J, Iglesias X, Rodríguez FA. Tennis play intensity distribution and relation with aerobic fitness in competitive players. PLoS One 10: e0131304, 2015.
3. Baiget E, Fernández-Fernández J, Iglesias X, Rodríguez FA. Heart rate deflection point relates to second ventilatory threshold in a tennis test. J Strength Cond Res 29: 765–771, 2015.
4. Baiget E, Fernández-Fernández J, Iglesias X, Vallejo L, Rodríguez FA. On-court endurance and performance testing in competitive male tennis players. J Strength Cond Res 28: 256–264, 2014.
5. Banzer W, Thiel C, Rosenhagen A, Vogt L. Tennis ranking related to exercise capacity. Br J Sports Med 42: 152–154, 2008.
6. Bellenger CR, Fuller JT, Nelson MJ, Hartland M, Buckley JD, Debenedictis TA. Predicting maximal aerobic speed
through set distance time-trials. Eur J Appl Physiol 115: 2593–2598, 2015.
7. Berthoin S, Gerbeaux M, Turpin E, Guerrin F, Lensel-Corbeil G, Vandendorpe F. Comparison of two field tests to estimate maximum aerobic speed. J Sports Sci 12: 355–362, 1994.
8. Berthoin S, Pelayo P, Lensel-Corbeil G, Robin H, Gerbeaux M. Comparison of maximal aerobic speed
as assessed with laboratory and field measurements in moderately trained subjects. Int J Sports Med 17: 525–529, 1996.
9. Berthon P, Fellmann N, Bedu M, Beaune B, Dabonneville M, Coudert J, Chamoux A. A 5-min running field test as a measurement of maximal aerobic velocity. Eur J Appl Physiol Occup Physiol 75: 233–238, 1997.
10. Billat LV, Koralsztein JP. Significance of the velocity at VO2max
and time to exhaustion at this velocity. Sports Med 22: 90–108, 1996.
11. Billat V, Beillot J, Jan J, Rochcongar P, Carre F. Gender effect on the relationship of time limit at 100% VO2max
with other bioenergetic characteristics. Med Sci Sports Exerc 28: 1049–1055, 1996.
12. Billat V, Renoux JC, Pinoteau J, Petit B, Koralsztein JP. Reproducibility of running time to exhaustion at VO2max
in subelite runners. Med Sci Sports Exerc 26: 254–257, 1994.
13. Billat VL, Blondel N, Berthoin S. Determination of the velocity associated with the longest time to exhaustion at maximal oxygen uptake
. Eur J Appl Physiol Occup Physiol 80: 159–161, 1999.
14. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1: 307–310, 1986.
15. Bower R, Cross R. String tension effects on tennis ball rebound speed and accuracy during playing conditions. J Sports Sci 23: 765–771, 2005.
16. Buchfuhrer MJ, Hansen JE, Robinson TE, Sue DY, Wasserman K, Whipp BJ. Optimizing the exercise protocol for cardiopulmonary assessment. J Appl Physiol Respir Environ Exerc Physiol 55: 1558–1564, 1983.
17. Buchheit M, Haydar B, Hader K, Ufland P, Ahmaidi S. Assessing running economy during field running with changes of direction: Application to 20 m shuttle runs. Int J Sports Physiol Perform 6: 380–395, 2011.
18. Buchheit M, Laursen PB. High-intensity interval training, solutions to the programming puzzle. Part II: Anaerobic energy, neuromuscular load and practical applications. Sports Med 43: 927–954, 2013.
19. Cohen J. Statistical Power Analysis for the Behavioural Sciences.(2nd ed.). Hillsdale, NJ: Lawrence Erlbaum, 1988.
20. Davey PR, Thorpe RD, Williams C. Fatigue decreases skilled tennis performance. J Sports Sci 20: 311–318, 2002.
21. Davey PR, Thorpe RD, Williams C. Simulated tennis matchplay in a controlled environment. J Sports Sci 21: 459–467, 2003.
22. Fernández J, Méndez-Villanueva A, Pluim BM. Intensity of tennis match play. Br J Sports Med 40: 387–391, 2006.
23. Fernández-Fernández J, Sanz-Rivas D, Méndez-Villanueva A. A Review of the activity profile and physiological demands of tennis match play. Strength Cond J 31: 15–26, 2009.
24. Fernández-Fernández J, Sanz-Rivas D, Sánchez-Muñoz C, de la Aleja Tellez JG, Buchheit M, Méndez-Villanueva A. Physiological responses to on-court vs running interval training in competitive tennis players. J Sports Sci Med 10: 540–545, 2011.
25. Fernández-Fernández J, Sanz-Rivas D, Sarabia JM, Moya M. Preseason Training: The effects of a 17-day high-intensity shock microcycle in elite tennis players. J Sports Sci Med 14: 783–791, 2015.
26. Ferrauti A, Kinner V, Fernández-Fernández J. The Hit & Turn tennis test: An acoustically controlled endurance test for tennis players. J Sports Sci 29: 485–494, 2011.
27. Girard O, Chevalier R, Leveque F, Micallef JP, Millet GP. Specific incremental field test for aerobic fitness in tennis. Br J Sports Med 40: 791–796, 2006.
28. Hornery DJ, Farrow D, Mujika I, Young W. An integrated physiological and performance profile of professional tennis. Br J Sports Med 41: 531–536, 2007.
29. Jones AM, Poole DC. Oxygen uptake dynamics: From muscle to mouth–an introduction to the symposium. Med Sci Sports Exerc 37: 1542–1550, 2005.
30. Kovacs MS. Energy system-specific training for tennis. Strength Cond J 26: 10–13, 2004.
31. Kovacs MS. Applied physiology of tennis performance. Br J Sports Med 40: 381–385, 2006.
32. Kovacs MS. Tennis physiology: Training the competitive athlete. Sports Med 37: 189–198, 2007.
33. Lacour JR, Padilla-Magunacelaya S, Chatard JC, Arsac L, Barthelemy JC. Assessment of running velocity at maximal oxygen uptake
. Eur J Appl Physiol Occup Physiol 62: 77–82, 1991.
34. Léger L, Boucher R. An indirect continuous running multistage field test: The Université de Montréal track test. Can J Appl Sport Sci 5: 77–84, 1980.
35. Lyons M, Al-Nakeeb Y, Hankey J, Nevill A. The effect of moderate and high-intensity fatigue on groundstroke accuracy in expert and non-expert tennis players. J Sports Sci Med 12: 298–308, 2013.
36. Méndez-Villanueva A, Fernández-Fernández J, Bishop D, Fernández-Garcia B, Terrados N. Activity patterns, blood lactate concentrations and ratings of perceived exertion during a professional singles tennis tournament. Br J Sports Med 41: 296–300, 2007.
37. di Prampero PE, Atchou G, Brückner JC, Moia C. The energetics of endurance running. Eur J Appl Physiol Occup Physiol 55: 259–266, 1986.
38. Reid M, Duffield R, Dawson B, Baker J, Crespo M. Quantification of the physiological and performance characteristics of on-court tennis drills. Br J Sports Med 42: 146–151, 2008.
39. Smekal G, Pokan R, von Duvillard SP, Baron R, Tschan H, Bachl N. Comparison of laboratory and “on-court” endurance testing in tennis. Int J Sports Med 21: 242–249, 2000.
40. Vergauwen L, Madou B, Behets D. Authentic evaluation of forehand groundstrokes in young low- to intermediate-level tennis players. Med Sci Sport Exerc 36: 2099–2106, 2004.
41. Vergauwen L, Spaepen AJ, Lefevre J, Hespel P. Evaluation of stroke performance in tennis. Med Sci Sports Exerc 30: 1281–1288, 1998.
42. Wasserman K, Hansen JE, Sue DY, Stringer WW, Whipp B. Principles of Exercise Testing and Interpretation. Including Pathophysiology and Clinical Applications.(4th. ed.). Philadelphia, PA: Lippincott, Williams & Wilkins, 2005.