Secondary Logo

Journal Logo

Cross training: indices of training stress and performance


Medicine & Science in Sports & Exercise: February 1998 - Volume 30 - Issue 2 - p 294-300
Applied Sciences: Physical Fitness and Performance

Twenty well-trained runners (˙VO2max 4.6 ± 0.5 L·min-1) were age and ability matched and assigned to either a cross training (CT) or run only group (RT). All subjects maintained normal running distance and intensity for 6 wk and reported for three additional training sessions per week. These workouts were performed outdoors on a 400-m track or measured road course (RT) or on a bicycle ergometer (CT). The sessions were as follows: (work·rest-1 ratio = 1): 5 × 5 min at >95% ˙VO2max/peak (Monday), 50-60 min at 70%˙VO2max/peak (Wednesday), and 3 × 2.5 min at >105%˙VO2max/peak, plus 6 × 1.25 min at >115%˙VO2max/peak (Friday). Subjects were tested before (PRE), after 3 wk (MID), and after 6 wk (POST) of intensified training. Blood samples were obtained from RT, CT, and ten controls (CON) at each time point (0600 h). Runners also completed a 10-min submaximal run at the same absolute intensity(velocity to elicit 75% of initial ˙VO2max) during which heart rate, RPE, and ˙VO2 were measured. Each runner then completed a simulated 5-km race (time trial) on a treadmill. Total testosterone (TT), free testosterone (FT), cortisol (C), and creatine kinase activity (CK) were determined. Running economy was similar between RT and CT; however, RPE decreased significantly at MID and POST compared with that at PRE (P< 0.05; time effect). There were no significant differences among groups for TT, FT, or CK, but C was significantly lower in CON than in RT and CT. Performance was significantly faster (P < 0.05; time effect) in the 5-km race at MID (1076.1 ± 81.4 s) and POST (1068.6 ± 83.9) compared with PRE (1096.6 ± 79.5) but was not different between CT and RT. In conclusion, RT and CT responded similarly to 6 wk of increased training, and both groups improved 5-km performance to a similar extent.

Wastl Human Performance Laboratory, Purdue University, West Lafayette, IN 47907; Exercise Physiology Laboratories, University of Toledo; and Northwest Ohio Center for Sports Medicine, The Toledo Hospital, Toledo OH 43606

Submitted for publication April 1997.

Accepted for publication August 1997.

The authors would like to acknowledge the assistance of Cindy Hodgson, James Fritz, and Laura Lewis. We would also like to thank our subjects for their dedication to this investigation. This study funded in part by a United States Olympic Committee Sports Science grant.

Address for correspondence: Michael G. Flynn, Ph.D., Purdue University, Wastl Human Performance Laboratory, Department of HKLS, Lambert 107C, West Lafayette, IN 47907. E-mail:

Cross training has been recommended as an adjunct to sport specific training for athletes wishing to improve performance or reduce the risk of injury (3,19). However, little scientific evidence is available to support these contentions. We recently found that male distance runners responded similarly to 10 d of hard training (200% of normal training) when the training consisted of either all running or equal amounts of cycling and running (5,20,21). Specifically, when using this extreme training model, there were no differences in global mood state (5), leukocyte subsets(21), endocrine markers of overtraining(5), muscle soreness (20), or performance parameters (20) between sport specific training and cross training. The subject number and the incidence of injury in this study were not sufficient to determine whether cross training provided less musculoskeletal stress or protected runners from injury; however, subjective ratings of training stress were similar for the two training methods (5). Triathletes have been reported to have a higher rate of injury than runners, cyclists, and swimmers(17); however, since triathlon is considered triple task specific (24), the total volume of training is likely the most important factor in injury rate.

Mutton et al. (18) found that running or a combination of running and cycling resulted in similar improvements in maximal oxygen consumption and running performance in “moderately fit” male runners. The authors acknowledged that the subjects were not competitive runners which was evidenced by their 5-km run times (>22 min), body fat(>19%), and training volume (16-30 km·wk-1 for previous 2 months). In addition, these subjects improved their 5-km run time by 1.7 min after 5 wk of training. Although these data may be generalized to higher level competition, it is important to determine whether cross training has similar effects on the performance of competitive runners.

Runners may increase their training volume abruptly when preparing for an important competition. Rapid increases in training volume have been reported to increase global mood state (16), alter serum levels of testosterone (4,6), cortisol(2), testosterone-to cortisol ratio(1,25), and performance (6). Increased training using a different exercise mode, i.e., cross training, may reduce the stress associated with rapid increases in sport specific training. As described above, we found similar changes in indices of training stress when runners doubled training volume with either additional running or an equal amount of cycling (5,20,21). However, we wanted to examine responses to increases in training more representative of those employed by competitive runners. The purpose of this investigation was two-fold. First, we sought to examine the effects of 6 wk of increased training, using “cross training” (cycling and running) or mode specific training (running only), on indices of training stress or overtraining. Second, we attempted to determine whether the 6-wk intensified training period employing either running or cross training would have similar effects on selected performance parameters.

Back to Top | Article Outline


Subjects and preliminary testing. The subjects in this investigation were 20 well-trained male distance runners. In addition, ten age-matched recreationally active subjects were used to control for potential seasonal variations in blood parameters. This study was approved by the Human Subjects Research Review Committee at the University of Toledo and written informed consent was obtained from all subjects before participation. Runners were required to have trained at least 32 km·wk-1 for the previous 6 months and were currently competing at 5-, 10-, or 42-km distances. The descriptive data for the runners are provided in Table 1. Each individual was tested for ˙VO2max/peak on a motor driven treadmill and a Monark (Varberg, Sweden) bicycle ergometer. The initial workload on the bicycle ergometer was 100 W and was increased by 50 W every 3 min for the first four stages. Thereafter, the workload was increased 25 W each min until volitional fatigue. The initial speed on the treadmill was set 53.5 m·min-1 below each runner's self-reported 10-km race pace. Velocity was increased 26.8 m·min-1 every 3 min until they were running 26.8 m·min-1 faster than race pace. Thereafter, the treadmill grade was increased 2% every 2 min until volitional fatigue. Past experience has shown that this protocol will provide at least four submaximal data points for each runner, rarely results in protocols > 16 min, and allows each individual to run fast but familiar speeds before the grade is increased.

Each subject kept detailed records of training distance and time for 2 wk before the start of the experimental period. In addition, they were provided with a notebook containing training logs, profile of mood state questionnaries and forms for subjective rating of training effort, muscle soreness, and leg fatigue (0-4 scale) during this 2-wk period. They were also required to report to the lab for two treadmill acclimation trials to become accustomed to the equipment used for the 5-km time trial. Photocells at the front and rear of the treadmill and interfaced with the treadmill control panel allowed the runner to control treadmill velocity. A third photocell, tripped by a retroflective strip on the treadmill belt, allowed speed and distance traveled to be monitored via microcomputer. This information was continuously displayed before the runner on a computer monitor. During the first acclimation trial the subjects were given a demonstration of the treadmill system, i.e., they were shown how to use the photocells to control speed and the monitor output was explained in detail. On that day each subject was asked to run on the treadmill and get accustomed to changing speeds with the photocells and monitor output. During the second acclimation trial on the following week, the subjects were required to complete a 5-km time trial on the treadmill.

The subjects were matched for age, training volume, and aerobic capacity(Table 1, Fig. 1) and were placed into either a running only (RT) or a “cross training” group (cycling and running, CT). While most subjects accepted random assignment into groups, two were willing to participate only if placed into RT. Since one of these subjects was assigned to the RT group by chance, only one matched pair had to be assigned to specific groups.

Testing sessions. All subjects were tested at the same time of day between 0500 and 0700 h, over a 2-d period during week 0 (PRE), and at the conclusion of weeks 3 (MID) and 6 (POST). Heart rate, blood pressure, and blood samples were obtained after 15 min rest from RT, CT, and CON. RT and CT also completed a 10-min submaximal run (75% of initial ˙VO2max) during which heart rate (UNIQ Heart Watch, CIC, Hempsted, NY) and ratings of perceived exertion were obtained. Expired gases were collected into Douglas bags for determination of respiratory exchange ration and oxygen consumption. At least 10 min after the submaximal run, each runner completed a computer simulated 5-km time trial on the treadmill.

Intensified training. The subjects were asked to maintain normal running distance and intensity for 6 wk (Fig. 2) and reported to the lab (0500-0700 h) on Monday, Wednesday, and Friday for additional training sessions. During Monday sessions all subjects completed 5× 5 min at a workload calculated to elicit 95% of˙VO2max/peak with a 5-min recovery between work periods. Wednesday sessions consisted of workouts of 50-60 min at a workload calculated to elicit 70% ˙VO2max/peak. Friday sessions incorporated 3 × 2.5 min at 105% ˙VO2max/peak (2.5 min recovery) and 6 × 1.5 min at 115%˙VO2max/peak (1.5 min recovery).

Group RT completed Monday and Friday sessions on a 400-m outdoor track and Wednesday sessions on a measured road course, whereas CT completed all their additional sessions outdoors on a bicycle ergometer. A notebook was completed by the subjects each week containing training logs, forms for profile of mood states (POMS), and subjective perception of training effort, leg fatigue, and muscle soreness. The instruction set for the POMS was changed to read“how you are feeling today.” Global mood state was calculated by summing the five negative affect scores from the POMS (tension/anxiety, depression/dejection, fatigue/inertia, confusion/bewilderment, anger/hostility), adding a factor of 100 and then subtracting the vigor/activity score (16).

Blood analyses. Serum was frozen at -20 C until analyzed by solid phase single antibody RIA using commercially available kits for testosterone(TT), free testosterone (FT), and cortisol (C) (Diagnostic Products Corp, Los Angeles, CA). Serum creatine kinase was determined using an enzymatic method(Sigma Chemical Co., St. Louis, MO).

Statistical analysis. Data were analyzed using a split-plot two-way ANOVA. When a significant difference was detected, a Newman-Keulspost-hoc test was used to determine where significant differences existed between means. An alpha level of 0.05 was set before the investigation.

Although we believe that the experimental design and statistical analyses were sound, the experiment may not have provided adequate statistical power to test the null hypothesis. Low statistical power is potentially associated with Type II statistical error.

Back to Top | Article Outline


There were no significant differences between RT and CT for treadmill or cycling ˙VO2max/peak(Table 1), initial training volume (CT 46.9 ± 14.9; RT 52.5 ± 12.97 km·wk-1, Fig. 1), or initial 5-km performance times (Fig. 3). Training volumes were significantly lower for week 6 compared with volumes reported during the 2 wk of preliminary training, and weeks 1, 2, 4, and 5 (Fig. 1). This decline was expected, since the runners were instructed to taper their training the last few days of the final week to give optimal effort for the last 5-km time trial.

Resting measures. There were no significant differences for resting heart rate or blood pressure (Table 2) between groups or across time. Global mood state (16) was not significantly changed over time nor was it different between treatments(Table 3). ANOVA revealed a significant F value for the main effect of time for subjective perceptions of training effort, muscle soreness, and leg fatigue (Table 3); however, Newman-Keuls post-hoc analysis did not reveal the location of the significant differences. Serum creatine kinase activity was not different between groups or across time.

Endocrine measures. Testosterone (Fig. 4) and free testosterone (Fig. 4) were not significantly different among groups or across time. In addition, there were no significant differences for the testosterone-to-cortisol ratio. Cortisol levels were not significantly changed across time; however, there was a significant group effect, such that cortisol was lower in the control group compared with both RT and CT (Fig. 5).

Submaximal exercise. Running economy was not significantly different either between groups or across time (RT PRE 47.9 ± 1.6 mL·kg-1·min-1, RT MID 47.6 ± 1.6, RT POST 47.4 ± 1.6; CT PRE 49.1 ± 1.5, CT MID 49.1 ± 1.6, CT POST 49.9 ± 2.0). However, ratings of perceived exertion (RPE, 0-10 scale) were significantly lower (time effect) for both MID and POST trials compared with those for PRE (Fig. 6). RPE was not significantly different between MID and POST, nor were there any significant group effects.

5 km time trial. Performance times in the 5-km time trial were significantly faster (time effect) at MID (1076.1 s) and POST (1068.6 s) compared with PRE (1096.6 s; Fig. 3); however, there were no significant differences between RT and CT.

Back to Top | Article Outline


This project was undertaken to confirm and extend the previous findings regarding cross training(5,7,15,18,20). Specifically, we examined whether a cross training regimen would enhance performance in a group of well-trained runners. Additionally, we wanted to examine whether a program of intensified training would elicit similar physiologic and psychologic stress when different modes of exercise were used to intensify training. Six weeks of intensified training via cross training or mode specific training resulted in a significant performance improvement (5-km time trial) in already well-trained runners. Furthermore, intensified training with either mode specific or cross training did not significantly alter subjective perceptions of training stress, global mood state, or endocrine parameters, suggesting that these parameters may only be disrupted when the change in training volume/intensity are substantially larger than that employed in this investigation.

Resting measures. Resting heart rate and blood presure were not different after 3 or 6 wk of intensified training. This was not surprising since these parameters are often reported to be unchanged when more rapid and severe increases in training volume have been employed(5,6,13,14). Hooper et al.(10), for example, reported no change in morning heart rate even in swimmers classified as stale. It could be argued that the relatively long time span between testing (3 and 6 wk) failed to detect changes in heart rate or blood pressure. However, subjective perceptions of training effort, leg fatigue, and muscle soreness tended to peak between weeks 2 and 4, suggesting that our week 3 trial fell during the period of greatest training stress. We believe that it is more likely that the intensified training was not of sufficient magnitude to disrupt typical indices of training stress. For example, Kirwan et al. (13) reported that doubling training volume of collegiate swimmers for 10 d did not change resting heart rate.

Endocrine measures. A number of investigators have reported unchanged testosterone (14,26), cortisol(4-6,12,26), and testosterone-to-cortisol ratio (6) after hard training. Since our training regimen was designed to be “stressful,” yet allow for enough recovery and adaptation to provide a stimulus for performance enhancement, it was not surprising that the training was of insufficient magnitude to alter resting endocrine parameters. Lehmann et al.(14) reported that increasing training volume by 100% for 3 wk did not significantly alter resting testosterone or cortisol. We previously reported (6) that increasing volume (88%) of swim training for 2 wk significantly decreased free testosterone, but these changes were not evident with modest increases in run training volume (35%). It is possible that the endocrine parameters typically used to monitor training may not be sensitive to the subtle changes in training volume that a runner may undergo. That is to say, it appears substantial increases in training volume or intensity (14) may be required to alter the serum hormone levels measured in the present study, and even at high levels of training the hormones generally fall within the normal range.

Submaximal exercise. Using a random crossover design, we previously found (20) that 10 d of intensified training using cycle ergometry resulted in a significant increase in oxygen consumption during a fixed velocity treadmill run. However, a similar increase in running volume did not significantly alter running economy. In the present study we found no significant changes in running economy for RT or CT. In our earlier study (20), the volume of training was increased by 100% and the intensity was maintained at about 70% ˙VO2max. In the present study the volume increase was about 35-45%, but the intensity of work was increased substantially for most runners. Lehmann et al.(14) have reported that 3 wk of increased volume (100%) led to changes in performance parameters indicative of overtraining, while 3 wk of increased intensity improved performance parameters. Reduced running economy was not associated with decrements in endurance performance in our earlier study (20), but there were significant changes in free testosterone, dehydroepiandrosterone-sulfate, and subjective perceptions of training effort, muscle soreness, and leg fatigue(5). Therefore, it is difficult to determine whether the different findings relative to running economy were a function of differences in levels of training stress or a differential response to varying training volume or intensity. Since performance improved with increased intensity in the present study and did not improve with increased volume in our earlier study, this might explain the discrepancy. It is also possible that reduced running economy in our earlier study was a transient response to the introduction of a novel activity (cycling) and that the subjects in the present study had become acclimated to cycling after 3 wk.

Using resistance training as a mode of cross training has been reported by others to improve running economy (11). For example, Johnston et al. (11) found that an upper and lower body resistance training program improved running economy in women runners without affecting ˙VO2max. These authors did not report performance changes; however, Hickson et al. (9) reported that resistance training improved cycling time to exhaustion (80%˙VO2max) for cyclists but not runners, while “short term endurance” (4-8 min performance) improved in both runners and cyclists.

5-km time trial. The most interesting finding in the present study was that 5-km performance times were improved to a similar extent for CT and RT. Foster et al. (7) found that increasing the training volume of already well-trained men and women improved 3.2-km performance times in both swim/run (-13.2 s) and run only groups (-26.4 s). These improvements (7) were accomplished after increasing training by 10% per week for 8 wk and were not associated with a concomitant increase in ˙VO2max or running economy. Mutton et al.(18) also reported a substantial improvement in performance and ˙VO2max after both cycle/run training or run training 4 d·wk-1 (2 run/2 cycle vs 4 run) for 5 wk. The present study confirms and extends the findings of previous investigators(7,18) since our subjects, while not elite, were of higher caliber. Since there were some lower caliber runners in each group, we also statistically analyzed the changes in performance times for the six fastest runners in the RT and CT groups (mean 5 km time 17.17 min). The improvements in 5-km performance for the fastest runners were similar (PRE-MID-21.2 s; PRE-POST -24.6 s; P < 0.05) to the mean improvement observed when all subjects were included, and there were no differences between training groups.

The mechanism behind the performance improvement is not readily apparent from the measurements obtained. For example, running economy was not improved during the intensified training period; however, the subjects perception of effort was significantly lower at MID and POST compared with that at PRE. While the ˙VO2max is somewhat stable in well-trained athletes(22), it is possible that the high intensity training improved cardiovascular function or muscle oxidative capacity. Additionally, the high intensity cycle and run training may have improved the lactate threshold, allowing the runners to sustain a higher velocity during the time trial. Inactive muscle has been suggested as a primary site of lactate removal during exercise (23), and it is possible that training muscle not normally recruited during running would furnish an additional site for lactate disposal and improve performance (7). However, Foster et al. (7) reported that a higher velocity was required to elicit a 4 mM lactate level in well-trained individuals who improved running performance by adding running to their baseline training, but not in a group who improved running performance by adding swimming to their baseline training level. That is, improving the aerobic potential of upper extremity muscle did not influence lactate threshold. We found (20) that submaximal lactate values were lower after 10 d of hard training (200% of normal training) when the training consisted of either all running or equal amounts of cycling and running; however, performance was not altered in these runners and it is possible that the lactate reduction was secondary to glycogen depletion(8). Finally, since it was not possible to“blind” subjects to treatment, it is also difficult to factor out psychological benefits (e.g., motivation, confidence) that may have resulted from the subjects' knowledge that they were training harder. Our subjects were involved in minimal interval training before the study period and although it is likely that the increased training intensity contributed to their performance improvement, the specific mechanisms remain to be determined.

In conclusion, 6 wk of increased training, via cross training or mode specific training, resulted in similar improvements in performance and no significant changes in indices of training stress. While these findings require replication using an elite subject population, the data clearly indicate that addition of high intensity cycling to the normal running regimen may substantially improve performance in well-trained male distance runners.

Figure 1-Running distance(km·wk

Figure 1-Running distance(km·wk

Figure 2-Schematic of testing and training. Intensified training, three additional training sessions per week on either the bicycle ergometer(cross training) or outdoor track/measured road course (running only)

Figure 2-Schematic of testing and training. Intensified training, three additional training sessions per week on either the bicycle ergometer(cross training) or outdoor track/measured road course (running only)

Figure 3-Running time (s) for the 5-km time trial. RT, running only; CT, cross training.

Figure 3-Running time (s) for the 5-km time trial. RT, running only; CT, cross training.

Figure 4-Total testosterone(ηmol·L

Figure 4-Total testosterone(ηmol·L

Figure 5-Cortisol (ηmol·L

Figure 5-Cortisol (ηmol·L

Figure 6-Rating of perceived exertion (0-10 scale) during submaximal running at the same absolute workload week 0, 3, and 6. RT, running only; CT, cross training.

Figure 6-Rating of perceived exertion (0-10 scale) during submaximal running at the same absolute workload week 0, 3, and 6. RT, running only; CT, cross training.

Back to Top | Article Outline


1. Aldercreutz, H., M. Harkone, K. Kuoppasalmi, et al. Effect of training on plasma anabolic and catabolic steroid hormones and their response during physical exercise. Int. J. Sports Med. 7:27-28, 1986.
2. Barron, G. L., T. D. Noakes, W. Levy, C. Smith and R. P. Millar. Hypothalamic dysfunction in overtrained athletes. J. Clin. Endocrinol. Metab. 60:803-806, 1985.
3. Bloom, Marc. Special Section: Cross training.Runner's World 25:46-58, 1990.
4. Dressendorfer, R. H. and C. E. Wade. Effects of a 15-d race on plasma steroid levels and leg muscle fitness in runners. Med. Sci. Sports Exerc. 23:954-958, 1991.
5. Flynn, M. G., F. X. Pizza, and P. G. Brolinson. Hormonal responses to excessive training: influence of cross training. Int. J. Sports Med. 18:191-196, 1997.
6. Flynn, M. G., F. X. Pizza, J. B. Boone Jr., F. F. Andres, T. A. Michaud. and J. R. Rodriguez-Zayas. Indices of training stress during competitive swimming and running seasons. Int. J. Sports Med. 15:21-26, 1994.
7. Foster, C., L. L. Hector, R. Welsh, M. Schraeger, M. A. Green, and A. C. Snyder, Effects of specific versus cross training on running performance. Eur. J. Appl. Physiol. 70:367-372, 1995.
8. Foster, C., A. C. Snyder, N. N. Thompson, and K. Kuettel. Normalization of the blood lactate profile in athletes. Int. J. Sports Med. 9:198-200, 1988.
9. Hickson, R. C. B. A. Dvorak, E. M. Gorostiaga, T. T. Kurowski, and C. Foster. Potential for strength and endurance training to amplify performance. J. Appl. Physiol. 65:2285-2290, 1988.
10. Hooper, S. L, L. Traeger-Mackinnon, A. Howard, R. D. Gordon, and A. W. Bachmann. Markers for monitoring monitoring overtraining and recovery. Med. Sci. Sports Exerc. 27:106-112, 1995.
11. Johnston, R. E., T. J. Quinn, R. Kertzer, and N. B. Vroman. Strength training in female distance runners: impact on running economy (Abstract). Med. Sci. Sports Exerc. 27(Suppl.):S8, 1995.
12. Keizer, H., G. M. E. Jannsen, P. Menheere, and G. Kranenburg. Changes in basal plasma testosterone, cortisol, and dehydroepiandrosterone sulfate in previously untrained males and females preparing for a marathon. Int. J. Sports Med. 10(Suppl 3):S139-S145, 1989.
13. Kirwan, J. P., D. L. Costill, M. G. Flynn, et al. Physiological responses to successive days of intense training in competitive swimmers. Med. Sci. Sports Exerc. 20:255-259, 1988.
14. Lehmann, M., U. Gastman, K. G. Petersen, et al. Training-overtraining: performance and hormone levels after a defined increase in training volume versus intensity in experienced middle- and long-distance runners. Br. J. Sports Med. 26:233-242, 1992.
15. Loy, S. F., J. J. Hoffmann, and G. J. Holland, Benefits and practical use of cross training in sports. Sports Med. 19:1-8, 1995.
16. Morgan, W. P., D. R. Brown, J. S. Raglin, P. J. O'Connor and K. A. Ellickson. Physiological monitoring of overtraining and staleness. Br. J. Sports Med. 21:107-114, 1987.
17. Murphy, P. Cross training may not reduce injury rate.Physician Sportsmed. 15:49-52, 1987.
18. Mutton, D. L., S. F. Loy, D. M. Rogers, G. J. Holland, W. J. Vincent, and M. Heng. Effect of run vs combined cycle/run training on˙VO2max and running performance. Med. Sci. Sports Exerc. 25:1393-1397, 1993.
19. O'Shea, P. The science of cross training: theory and application for peak performance. J. Strength Condit. Res. 12:40-44, 1991.
20. Pizza, F. X., M. G. Flynn, R. D. Starling, et al. Run training vs cross training: influence of increased training on running economy, foot impact shock and run performance. Int. J. Sports Med. 16:180-184, 1995.
21. Pizza, F. X., M. G. Flynn, T. Sawyer, P. G. Brolinson, R. D. Starling, and F. F. Andres. Run training versus cross training: effect of increased training on circulating leukocyte subsets. Med. Sci. Sports Exerc. 27:355-362, 1995.
22. Sjodin, B. and J. Svedenhag. Applied physiology of marathon running. Sports Med. 2:83-99, 1985.
23. Stainsby, W. N. and G. A. Brooks. Control of lactic acid metabolism in contracting muscles and during exercise. In:Exercise and Sports Science Reviews, Vol. 18, K. B. Pandolf and J. O. Holloszy (Eds.) Baltimore: Williams and Wilkins, 1990, pp. 29-63.
24. Stamford, B. Task specific training vs cross training.Physician Sportsmed. 19:113-114, 1991.
25. Urhausen, A., T. Kullmer, and W. Kindermann. A 7-week follow up study of the behavior of testosterone and cortisol during the competition period in rowers. Eur. J. Appl. Physiol. 56:528-533, 1987.
26. Urhausen, A., H. Gabriel, and W. Kindermann. Blood hormones as markers of training stress and overtraining. Sports Med. 20:251-256, 1995.


©1998The American College of Sports Medicine