KEY WORDS: Maximal test, oxygen consumption, Karate, motivation.

Participants
Five male competitive Karate practitioners participated in this study. All participants were highly competent in Karate kumite (sparring) to at least England National Squad level; Age (M = 31, SD = 9 years), height (M = 1.80, SD = 0.05 m), body mass (M = 78.9, SD = 13.2 kg) and experience (M = 15. 4, SD = 8.6 years). All participants were informed of the procedures of the investigation and completed a pre-test health-screening questionnaire and provided written consent, according to ACSM guidelines (ACSM, 1995) prior to participation. The University Research Ethics Committee approved all experimental procedures.

Sport specific protocol development
Pilot testing took place over a 1-month period. Figure 1 (a-d) illustrates the techniques performed in the test with gas analysis equipment attached. These techniques were selected from video analysis of competition Karate and the most common techniques observed were selected. These included the straight punch and the roundhouse kick.

The test protocol itself involved a sequential set of straight punch and roundhouse kick combinations on a heavy punch/kick bag suspended from a wall mounted bracket. The combination consisted of a leading straight punch (Figure 1a), followed by a rear leg roundhouse kick (Figure 1b), a rear straight punch (Figure 1c), and a leading roundhouse kick (Figure 1d), repeated twice. The time to complete this set of movements accurately and without haste was set at 7 seconds. This allowed sufficient time to execute and prepare each strike in a controlled manner.

The progression in intensity of the exercise during the test was based on a similar sequence of emitted audio beeps as the multistage fitness test (Léger et al., 1988). The test was designed with 2 auditory signals, the first to let the participant know when to begin the bout of exercise and a second sound to indicate when they should rest (7 seconds later).

The time to complete the exercise bout remained the same, 7 seconds, whilst the recovery time between bouts would be progressively decreased. Participants had to perform each strike and kick with as maximum force as possible. The aim here was to maintain maximal exercise intensity whilst progressively making the test more demanding by reducing the recovery between exercise bouts. One additional participant performed 3 experimental trials to obtain sets of recovery periods and subsequent decrements in time allotted for each particular set of recovery periods. This participant was not included in the actual experimental trials due to the imbalance in familiarity with the test. Once the data from these trials was analysed, a final test protocol was developed.

The protocol was again similar to that of Léger et al., 1988, in that stages and levels were adopted to correspond to certain periods of the recovery time. The outcome of pilot testing and the final test protocol is presented in Table 1. The final test protocol was placed onto a laptop PC.

Procedures
30 minutes before the onset of the experiment, participants were allowed to warm- up and stretched for 15 minutes. 10 minutes prior to test commencement participants were allowed a trial, with the portable gas analysis system and heart rate monitor attached, to practice the techniques used on the heavy bag and familiarise themselves with the test procedure. 5 minutes before the test all systems were checked again and the computer programme was put to ready. The test began after a pre-recorded 5 second auditory countdown. Expired air was analysed continuously throughout the test (Cosmed K4b2, Cosmed^TM S.r.l, Italy). Often peak oxygen uptake (VO_2peak) is used to express the highest oxygen uptake obtained when the criteria for achieving a VO₂max are not strictly met (BASES, 1997). Both relative (ml·kg^-1·min^{- 1}) and absolute (L·min^-1) peak oxygen uptake (VO_2peak) were defined as the highest 30 second average VO₂. The K4b2 is an upgraded version of the K2 gas analyser (Vacumetrics Inc., Ventura, CA), which has been shown to be both a valid (Lothian et al., 1993) and reliable (Bigard et al. 1995; Crandall et al., 1994) system for analysing VO₂ uptake. The heart rate (HR) was recorded every 5 seconds during the test via short wave telemetry (Polar Vantage, Polar Electro OY, Kempele, Finland). During the test, one of the researchers held the heavy bag in place to avoid unwanted movement. The test ended when the participant achieved volitional exhaustion at which point time to exhaustion (TE), exercise level and the cycle number obtained on the test were recorded. A second performance of the test was carried out 1 week after the initial test. Following both trials subjective feedback was noted, however this was not formally standardised.

Data handling and statistics
All results are presented as mean ± S.D. Due to the small sample size, a non- parametric differences test (Wilcoxin signed-rank) was carried out on all dependant variables. In addition, relative and absolute VO_2peak values obtained from the two tests were individually plotted against TE to examine possible relationships via linear regression. Where a significant correlation was observed, the coefficient of determination (variance) was determined. A significance level of p < 0.05 was established prior to analysis. Statistical tests were run using SPSS version 12.1 for windows.

There was no significant difference in VO_2peak, HRM and TE between test 1 and test 2 (p > 0.05). However, a significant difference in VE_peak between test 1 and test 2 was observed (p < 0.05). Figure 2 illustrates a significant relationship (R² = 0. 77, p < 0.001) between time to exhaustion and relative VO_2peak. The calculated coefficient of determination was 0.593 (0.77 x 0.77). Therefore, 59.3% (0.593 x 100) of the variance in TE is accounted for by relative VO_2peak. There was no relationship between TE and absolute VO_2peak, (R² = 0.28, p > 0.05). Table 2 presents the observed absolute VO_2peak, relative VO_2peak, VE_peak, HRM, and TE for test 1 and test 2 and the difference in these measures between tests. Absolute VO_2peak, relative VO_2peak, and TE increased (mean ± SD) 0.04 ± 0.1 L·min^{- 1}, 1 ± 1.1 ml·kg^-1·min^-1 and 28 ± 8.1 s respectively. HRM and VE_peak decreased by 3 ± 0.3 beats·min^-1 and 2.8 ± 1.6 L·min^-1 respectively.

Subjective feedback from participants suggested that the test accurately simulated the actions used in competitive sparring situations. In addition, participants reported that they "felt more motivated" performing the new test than previous aerobic assessment tests.

At present, this is the first sports specific test to assess the aerobic fitness of competition Karate practitioners. The first thing to detail is the fact that the new test simulates the actions used in competition Karate. All participants are familiar with other means of aerobic fitness testing, the most common being the multistage fitness test or 'bleep' test (Léger et al., 1988). Participants reported that the new test accurately simulated the actions they used in competition sparring, and this in turn "increased motivation" to perform well on the new test compared to previous aerobic capacity testing.

The difference between test 1 and test 2 absolute VO_2peak, relative VO_2peak, HRM, and TE was 0.04 L·min^-1, 1 ml·kg^-1·min^-1, -3 beats·min^-1, and 28 s, respectively. These relatively small, non-significant differences indicate that the newly developed test demonstrates potential test re-test reliability in these variables. VE_peak displayed a significant difference between the two tests, decreasing 2.8 L·min^-1 in the second test. This finding indicates the new test is potentially less reliable at assessing VE_peak.

There was a strong relationship between TE and relative VO_2peak (Figure 2) and potentially a large percentage (59.3%) of the variance in TE is attributed to relative VO_2peak. Performance on the new test involves weight-bearing exercise utilising both upper and lower limbs. It is well established that relative VO_2peak relates strongly to performance for weight-bearing whole body exercise as witnessed in the present study (Haug et al., 1999). The present finding indicates that the new test potentially assesses the appropriate capacity of the aerobic system.

In developing the new test, a format similar to that of Léger et al., 1988 was used. The 20m shuttle run test was designed to increase to such intensity that it was almost impossible, if not impossible, to complete. It was our aim that the new test should also be of an intensity that prevented completion. However, it is noticeable that one participant completed the new test on both occasions, indicated by a TE of 1182 seconds. Another participant completed the test on the second testing occasion, increasing time to exhaustion by 76 s. Both participants had a VO_2peak equal to or greater than 55 ml·kg^-1·min^-1. Moreover, these participants reported that they could have continued exercising longer had the test not ended.

Current guidelines for maximal aerobic capacity testing indicate VO_2peak should be achieved within 8 - 15 mins (BASES, 1997). In the present test, only one participant achieved VO_2peak within this time (Participant 4, Table 2). For those participants that did not complete the test (n = 3), average time to achieve VO_2peak was 16.9 mins and 17.2 mins for tests 1 and 2 respectively. This, coupled with the fact that time to completion of the test is already 19 mins and 42 s (1182 s/60), indicates that the new test is perhaps already too long. Further time additions (e.g. an increase in number of exercises/stages) would be impractical. It would be more practical to consider the new test to be of insufficient intensity to achieve VO_2peak within guideline duration. Likewise, the new test appears of insufficient intensity for individuals with a VO_2peak 55 ml·kg^-1·min^-1 or greater.

Implications for test development
Due to the small n, consideration to the possible presence of both type I and type II error needs to be given. A larger n would reduce the possibility of these errors occurring and provide sufficiently statistical significant results. This should be addressed accordingly in continuation study of the new test. Additional further developments to the test include a need to increase intensity at an earlier stage in an effort to prevent completion of the test. This could be achieved by eliminating early stages used as a warm up period, and increasing the number of stages within each exercise. However, when performing a maximal oxygen uptake test on a treadmill or cycle ergometer, work rate or intensity progressively increases by either increasing speed (treadmill) or workload (cycle) until volitional exhaustion. One of the restrictions of the new Karate test is the fact that work rate cannot be assessed nor altered in the same manner as treadmill or cycle ergometry workload. In the new test participants are asked to perform each technique with as much force as possible. Towards the end of the test, it was noticeable that participants were not able to maintain the same level of force as that of earlier stages. The problem lies in distinguishing if participants are giving a maximal effort each time they perform the activity bout. Test developments will look to address this issue via methods of assessing power output during the test. The use of force transducers within a boxing punch bag (Karpitowski et al., 1994) and boxing manikin (Smith et al., 2000) to monitor strike forces has been shown to provide accurate and reliable force and power outputs in boxers. The adoption of such methods for assessment of force/power output in the new test warrants further investigation.

Finally, subjective feedback provided information regarding the effectiveness of the new test in simulating actions specific to competitive karate. As a result, participants reported feeling "more motivated" performing the new test compared to previous non-specific aerobic assessment tests. The use of appropriate psychological inventories (Reeve and Deci, 1996) may have allowed for quantification of this improved motivation. Further developments will look to employ such methods, providing additional ecological validity to the work.

In summary, the new test accurately simulates the actions used in competition Karate and anecdotal reports are that the test is more motivating than other familiar methods of assessing aerobic capacity in Karate practitioners. Time to exhaustion during the test was strongly related to relative VO_2peak. At present, the new test does not allow us to quantitatively measure performance effort, making it difficult to distinguish a 'true' VO_2peak. The use of force transducers could be assessed to address this issue.

• This is the first attempt at an aerobic fitness test specific to competitive Karate practitioners
• Anecdotal reports are that the new test accurately simulates the actions used in competition Karate
• Relative VO_2peak was significantly related to time to exhaustion, with 63.5% of the variance in time to exhaustion attributed to relative VO_2peak.
• Test developments include the use of force plates and transducers to assess force/power output during the test