Journal of Gerontology: MEDICAL SCIENCES
2000, Vol. 55A, No. 6, M311–M316
Copyright 2000 by The Gerontological Society of America
Maximal Power Across the Lifespan
J. C. Martin, R. P. Farrar, B. M. Wagner, and W. W. Spirduso
Department of Kinesiology and Health Education, The University of Texas at Austin.
Background. Previous investigators have reported that maximal power increases during growth and decreases with
aging. These age-related differences have been reported to persist even when power is scaled to body mass or muscle
size. We hypothesized that age-related differences in maximal power were primarily related to differences in muscle size
and fiber-type distribution rather than to age per se.
Methods. Maximum cycling power (Pmax) and optimal pedaling rate (Vopt, a surrogate measure for muscle fiber type)
were determined for 195 boys and men, 8–70 years of age, by using inertial load cycle ergometry. Anthropometric dimensions were used to estimate lean thigh volume (LTVest) of all subjects, and magnetic resonance imagery was used to
determine thigh and hip muscle volume (MRIvol) for 24 subjects.
Results. Pmax was highly related to the product of LTVest and Vopt (LTVest ⫻ Vopt; r2 ⫽ .83). Multiple regression revealed that Pmax was significantly related to both LTVest ⫻ Vopt and age (R2 ⫽ .84). Power scaled by LTVest ⫻ Vopt was
stable during growth and exhibited a small but significant decrease with aging. MRIvol was highly correlated with
LTVest, and the ratio of LTVest to MRIvol was independent of age.
Conclusions. These results suggest that muscle volume and optimal pedaling rate are the main determinants of maximal power across the lifespan and that the contractile properties of muscle are developed early in childhood and remain
nearly intact late into the lifespan.
M
AXIMAL muscular power changes across the lifespan:
it increases during childhood, reaches a peak during
adulthood, then declines with old age. Production of maximal
power may be required for many daily activities; climbing
stairs, crossing a busy street, and recovering balance may represent short-term maximal efforts, particularly in young children and older adults. The ability to produce sufficient power
to perform these types of activities may have a major influence on the development of children and the independence of
older adults. At a more basic level, maximal power represents
the integration of neural and muscular function, and serves as
an indicator of the integrity of the neuromuscular system. It is
therefore important to understand the time course of development, and subsequent decline, of maximal power.
Previous investigators have reported that the maximal
power of young adults is two- to threefold higher than the
power of 8- to 10-year-old children (1–5). Such a difference
is not unexpected because adults are more than twice as large
as 8-year-old children (6). However, the power of adults has
been reported to be 35–50% greater than that of children
when scaled to body mass (1–5) and 20–53% greater when
scaled to muscle size (7–9). From young adulthood to older
age, total power and power scaled to body mass have been
reported to decrease by 6–11% per decade (8,10–13) and by
6%–8% per decade when scaled to muscle size (11,14).
Some previous investigators have discussed possible mechanisms for these age-related differences in maximal power
(1,3,4,9,11), but the exact mechanisms remain unknown.
One mechanism known to affect maximal power is total
ATPase activity (15), which is a function of muscle size and
fiber type. Indeed, maximal cycling power has been reported to be highly related to lean thigh volume (16) and to
muscle fiber–type distribution (12,17,18). Additionally, op-
timal pedaling rate has been shown to be highly related to
muscle fiber type (18) and thus may serve as a surrogate
measure for muscle fiber–type distribution.
We hypothesized that age-related differences in maximal
power were primarily related to differences in muscle size
and fiber-type distribution rather than to age per se. To test
that hypothesis, we measured the maximum cycling power,
lean thigh volume, and optimal pedaling rate of males aged 8
to 70 years; we then analyzed those data to determine what
portion of the variability in maximal power was related to
the product of muscle size and optimal pedaling rate and
what portion was related to age. An additional purpose was
to determine if anthropometric dimensions provide a valid
estimate of muscle volume in active males of different ages.
METHODS
A total of 195 males, 8 to 70 years of age, were recruited
from schools and from a list of licensed competitive cyclists. Subjects under 18 years of age were physically active
and enrolled in physical education classes. Subjects between the ages of 18 and 30 years were either physically active or were competitive cyclists. All subjects over 30 years
of age were competitive cyclists. Prior to testing, the subjects had the procedures explained verbally and the test
demonstrated. Each minor provided written assent, and his
parent or guardian provided written consent. Subjects over
18 years of age provided written informed consent. This
study was approved by the Institutional Review Board of
The University of Texas at Austin.
Cycling power was measured using the inertial load
method (16), which determines maximal power across a
range of pedaling rates in one short exercise test. The ergometer was fitted with bicycle-racing handlebars, cranks, pedM311
M312
MARTIN ET AL.
als, and seat and was fixed to the floor. Each subject wore
cycling shoes fitted with a cleat that locked into a springloaded binding on the pedal. The ergometer frame was modified to accommodate the lower seat position required by
the younger subjects, and three handlebar stems were available to provide a comfortable position for each subject.
This method has been described previously (16). Briefly,
the trajectory of an ergometer flywheel was determined by
means of a slotted disc mounted on the flywheel and an infra-red photodiode on the ergometer frame. The slots were
spaced at an angle (⌬θ) of ␲/8 radians and alternately
passed or interrupted the infra-red beam. Time between
consecutive interrupts (⌬t) was recorded by a microprocessor with a clock accuracy of ⫾ 0.5 ␮s. Flywheel angular velocity was calculated as ⌬θ/⌬t. The time–angular velocity
data were low-pass filtered at 8 Hz by using a 5th order
spline (19). Power was calculated as the rate of change of
kinetic energy for each revolution of the cranks, and no frictional resistance was applied to the flywheel. Maximum
power (Pmax) was identified as the apex of the power–pedaling rate relationship. Optimal pedaling rate (Vopt) was the
pedaling rate at which Pmax occurred. The inertial load was
varied to accommodate the various sizes of the subjects.
Loads were 3.2 kg ⫻ m2 for body masses of 24–39 kg, 5.24
kg ⫻ m2 for masses of 40–49 kg, 8.6 kg ⫻ m2 for masses of
50–69 kg, and 9.5 kg ⫻ m2 for masses of 70 kg and above.
Leg circumference (C) and length were measured at gluteal furrow, mid-thigh, and proximal patella. Skin-fold
thickness (SF) was measured using a Harpenden skin-fold
caliper. Lean limb circumference (LC) was estimated as LC ⫽
C ⫺ SF ⫻ ␲ (20). Lean limb diameter (LD) at gluteal furrow, mid-thigh, and proximal patella was estimated from
the circumference (LD ⫽ LC/␲). Lean cross-sectional area
was estimated as ␲LD2/4. Lean thigh volume (LTVest) was
estimated as the sum of two truncated cones (21).
Magnetic resonance imagery was used to obtain precise
measures of hip and thigh muscle volume (MRIvol) in a subset of 24 subjects. T1 images were recorded by a General
Electric Signa (Milwaukee, WI) system at 16 kHz with 10mm slice thickness from the knee to the iliac crest. Images
were cropped to remove the background and most non-muscular tissue by using image processing software (Adobe
Photoshop 5.0, San Jose, CA). Muscle area within each image was determined using National Institutes of Health image processing software (Scion Image for Windows), which
determined the number of pixels in each image within a
range of gray-scale density that was characteristic of muscle
tissue. Incremental volumes were calculated as the product
of cross-sectional area and slice thickness. Total MRIvol was
calculated as the sum of incremental volumes from the origin of gluteus maximus to the knee joint.
Subjects who were not cycling-trained performed three
days of practice prior to experimental data collection,
whereas the trained cyclists performed only the experimental trials. This procedure has been previously shown to produce reliable results in active men and cyclists, respectively
(22). During all practice and experimental trials, subjects
performed a 5-minute warm-up of cycling at 100–120 rpm
and then rested for 2 minutes prior to performing four
power tests with 2 minutes resting recovery between tests.
Subjects began each test from rest and accelerated maximally for approximately 3-4 seconds on a verbal command
with standardized encouragement. For the cyclists, seat
height (measured from the top of the saddle to the pedal
axle when the pedal was at its lowest point) was set to
match their accustomed riding position, which was found to
average 108.5% of leg length (i.e., height minus seated
height). Seat height for the active subjects was likewise set
at 108.5% of leg length. Subjects remained seated throughout each bout, and feedback regarding Pmax was provided after each power test.
Subjects were grouped by age: 8–11, 12–19, and each decade to age 70. The effects of training on Pmax, scaled by
LTVest ⫻ Vopt, were determined by comparing the active (n ⫽
25) and cycling-trained subjects (n ⫽ 25) ages 18–30 with a
Student’s t test. If scaled Pmax of those subjects did not differ, the data for active and cycling-trained subjects were
combined for subsequent analyses. Values for body mass,
LTVest, Vopt, LTVest ⫻ Vopt, and Pmax (absolute and scaled to
body mass, LTVest, and LTVest ⫻ Vopt) were analyzed using
analysis of variance, and Bonferroni post hoc procedures
were used to determine which age groups differed ( p ⬍ .05)
from the young adults (20–29 years of age). Linear regression was performed to determine the relationship of Pmax
with LTVest and LTVest ⫻ Vopt. Multiple linear regression of
Pmax with LTVest ⫻ Vopt and age was performed to determine what additional portion of the variability in Pmax was
related to age per se. Linear regression of LTVest with
MRIvol was performed to determine how well the anthropometric estimate predicted the actual muscle volume of the
leg. Linear regression was also performed on the ratio of
LTVest to MRIvol versus age to determine if an age-related
increase in intramuscular fat might bias the accuracy of the
anthropometric estimate. Linear regression of Pmax versus
LTVest ⫻ Vopt, and Pmax versus MRIvol ⫻ Vopt were performed to determine if MRIvol was more predictive of Pmax
than was LTVest.
RESULTS
Age, body mass, LTVest, and Vopt for each age group are
presented in Table 1. Members of the youngest group
were significantly less massive than were the young
adults (20–29 years). LTVest of the two youngest groups
were significantly less than that of the young adults. Vopt
of the youngest and oldest groups were significantly less
that of the young adults. LTVest ⫻ Vopt of the two youngest
and the oldest groups were significantly less that of the
young adults.
Pmax, scaled by LTVest ⫻ Vopt, of young active (2.22 ⫾
0.06) and cycling-trained (2.25 ⫾ 0.05) men did not differ.
Therefore, active and trained subjects were combined for
analysis.
Absolute and scaled values for Pmax are presented in Table 2. Absolute Pmax (Figure 1) of the young adults was approximately 220% greater than that of the youngest boys,
whereas Pmax of men in the oldest age group was 30% less
than that of the young adults; age groups 8–11, 12–19, 50–
59, and 60–70 had significantly lower Pmax than the young
adults. Pmax scaled to body mass (Figure 2) of the young
adults was approximately 40% greater than that of the
MAXIMAL POWER ACROSS THE LIFESPAN
M313
Table 1. Subject Characteristics
Age Group (years)
n
Age (years)
Body Mass (kg)
LTVest (L)
Vopt (rpm)
LTVest ⫻ Vopt
8–11.99
12–19.99
20–29.99
30–39.99
40–49.99
50–59.99
60–70
21
21
38
27
21
40
27
10.4 ⫾ 0.9
16.0 ⫾ 2.1
24.8 ⫾ 2.8
34.5 ⫾ 3.2
44.4 ⫾ 2.4
54.3 ⫾ 2.7
64.4 ⫾ 3.0
33 ⫾ 6*
66 ⫾ 21
76 ⫾ 12
78 ⫾ 10
77 ⫾ 8
82 ⫾ 12
76 ⫾ 10
1.75 ⫾ 0.46*
3.89 ⫾ 1.33*
5.02 ⫾ 0.91
5.15 ⫾ 1.01
4.96 ⫾ 0.79
4.66 ⫾ 0.82
4.37 ⫾ 0.82
115 ⫾ 12*
124 ⫾ 8
124 ⫾ 8
124 ⫾ 8
123 ⫾ 10
112 ⫾ 8
114 ⫾ 8*
202 ⫾ 60*
483 ⫾ 168*
624 ⫾ 120
640 ⫾ 150
612 ⫾ 131
558 ⫾ 108
498 ⫾ 115*
Notes: Age, mass, LTVest, Vopt, and LTVest ⫻ Vopt are presented as mean ⫾ SD for each age group.
*Groups whose results differ ( p ⬍ .05) from those of the young adults (20–29.99 years).
youngest boys, whereas scaled Pmax of men in the oldest age
group was 30% less than that of the young adults; age
groups 8–11, 40–49, 50–59, and 60–70 had significantly
lower scaled Pmax than the young adults. Pmax scaled to
LTVest (Figure 3) of the young adults was not different than
that of the youngest boys, whereas scaled Pmax of men in the
oldest age group was 19% less than that of the young adults;
age groups 50–59 and 60–70 had significantly lower scaled
Pmax than the young adults. Finally, Pmax scaled to LTVest ⫻
Vopt (Figure 4) of the young adults was not different than
that of the youngest boys, whereas scaled Pmax of men in the
oldest age group was 12% less than that of the young adults;
age groups 50–59 and 60–70 had significantly lower scaled
Pmax than the young adults.
Regression analyses indicated that Pmax was highly related to LTVest and to LTVest ⫻ Vopt, and that age accounted
for a small but significant portion of the variability.
P max = 232.7LTV est + 53.7 ,
2
r = .755 , F = 595 , p < .001
P max = 1.835LTV est × V opt + 97.5 ,
2
r = .826 , F = 913 , p < .001
P max = 1.906LTV est × V opt – 2.68age + 160 ,
Vopt was more predictive of Pmax (r2 ⫽ .86, F ⫽ 134) than
was LTVest ⫻ Vopt (r2 ⫽ .79, F ⫽ 87).
DISCUSSION
The main finding of this study is that the product of muscle size and optimal pedaling rate, an indicator of muscle fiber type, accounted for 82% of the variability in Pmax across
six decades of age, whereas age accounted for less than 2%
of the variability. When Pmax was scaled by LTVest ⫻ Vopt,
the previously observed age-related increase from childhood to young adulthood was eliminated and the decrease in
the older groups was reduced to 12% over four decades (approximately half as large as the previously reported agerelated decrease). The stability of Pmax, scaled by LTVest ⫻
Vopt, suggests that the contractile properties of muscle of
any specific fiber type are developed early in childhood and
remain nearly intact late into the lifespan.
Obtaining valid measures of maximal muscular power
from the diverse subjects in this investigation required that
the protocol be safe and that the subjects be highly motivated. We chose cycling as our power measurement technique because the chance of injury was minimal and the test
was of such short duration (3–4 seconds) that it was not influenced by cardiovascular fitness. During the testing of our
younger subjects, other children from their physical educa-
2
R = .843 , F = 516 , p < .001
For the 24 subjects in the magnetic resonance imaging
portion of this study (Figure 5), LTVest was highly correlated to MRIvol (r2 ⫽ .914). The ratio of MRIvol to LTVest
was independent of age (r2 ⫽ .03, F ⫽ .71, p ⫽ .41). MRIvol ⫻
Table 2. Absolute and Scaled Maximum Power
Age Group
(years)
Power (watts)
Scaled to Body
Mass (W/kg)
Scaled to
LTVest (W/L)
Scaled to
LTVest ⫻ Vopt
8–11.99
12–19.99
20–29.99
30–39.99
40–49.99
50–59.99
60–70
417 ⫾ 24*
1,060 ⫾ 79*
1,322 ⫾ 38
1,335 ⫾ 52
1,203 ⫾ 58
1,057 ⫾ 31*
927 ⫾ 37*
12.4 ⫾ 0.3*
15.9 ⫾ 0.5
17.4 ⫾ 0.4
16.9 ⫾ 0.4
15.5 ⫾ 0.6*
13.0 ⫾ 0.3*
12.1 ⫾ 0.3*
241 ⫾ 7
274 ⫾ 7
266 ⫾ 6
261 ⫾ 6
243 ⫾ 8
229 ⫾ 5*
214 ⫾ 7*
2.11 ⫾ 0.07
2.21 ⫾ 0.06
2.14 ⫾ 0.04
2.11 ⫾ 0.05
1.97 ⫾ 0.05
1.92 ⫾ 0.04*
1.89 ⫾ 0.06*
Notes: Power was scaled by body mass, LTVest, and LTVest ⫻ Vopt (mean ⫾ SD).
*Groups whose results differ ( p ⬍ .05) from those of the young adults (20–
29.99).
Figure 1. Absolute power versus age. Maximum cycling power increased 220% across the young age groups, reached a peak during
adulthood, then decreased by approximately 7.5% per decade with
aging.
M314
MARTIN ET AL.
Figure 2. Power scaled to body mass versus age. Scaled power increased 40% across the young age groups and decreased approximately 7.6% per decade with aging.
Figure 4. Power scaled to LTVest ⫻ Vopt versus age. Scaled power
did not change across the young age groups and decreased approximately 3% per decade with aging.
tion class cheered for the boy performing the test, providing
high motivation. The cyclists volunteered primarily because
they hoped to learn something that would help them with
their racing. Indeed they were so highly motivated that most
of them traveled several hours just to perform our testing.
Another advantage of recruiting cyclists was that no practice trials were necessary (22), so our older subjects were
only required to visit the laboratory one time. This one-time
visit allowed us to recruit subjects who might not have volunteered to make several visits. Finally, because the older
subjects were trained at cycling, they were confident and
able to exert a truly maximal effort.
An important difference between this investigation and
previous investigations of age and power was the use of optimal pedaling rate as a surrogate measure for muscle fiber–
type distribution. It is well known that muscle fiber type affects maximal power (23,24); however, the relationship of
optimal pedaling rate with muscle fiber type has only re-
cently been reported by Hautier and colleagues (18). They
reported that Vopt is strongly related to the proportion of
type II muscle fiber area (r2 ⫽ .77). Prior to that investigation, McCartney and colleagues (25) reported that the maximum pedaling rate of a subject with 72% type II fibers was
380 rpm compared with 225 rpm for a subject with 53%
type II fibers, suggesting a relationship between fiber type
and maximum pedaling rate. Conversely, Sergeant and colleagues (26) reported that optimal pedaling rate was similar
in four subjects with similar muscle fiber–type distribution.
Figure 3. Power scaled to LTVest versus age. Scaled power increased 10% across the young age groups and decreased approximately 5% per decade with aging.
Figure 5. Magnetic resonance images of the mid-thigh of 9- (upper
left), 20- (upper right), 40- (lower left), and 61-year-old (lower right)
subjects.
MAXIMAL POWER ACROSS THE LIFESPAN
Our use of optimal pedaling rate as a surrogate measure for
muscle fiber type is, therefore, well supported.
The composite variable LTVest ⫻ Vopt was used to account for the combined effects of muscle size and fiber-type
distribution. Another method to account for two independent variables is multiple linear regression, which would be
appropriate if muscle size and fiber-type distribution affected muscular power via additive mechanisms. However,
Swoap and colleagues (23) reported that rat soleus, a muscle
that is 95% slow twitch, produced 26 W/kg of muscle mass
during cyclic contraction, whereas the plantaris, which is
95% fast twitch, produced 144 W/kg. Because fiber type affects the power a muscle can generate per unit mass, the effects of muscle size and fiber-type characteristics are multiplicative. Thus, use of the product of LTVest and Vopt as a
single independent variable is appropriate and based on
physiological principles.
To investigate the effects of age, without confound associated with sedentary lifestyle, we sought out men who maintained vigorous training late into the lifespan. We were unable, however, to obtain a sufficient sample of cyclingtrained boys. Therefore, we determined the effects of training
in 50 men 18–30 years of age. In those men, Pmax was nearly
identical when scaled by LTVest ⫻ Vopt. This equivalence in
scaled power suggests that at least for our power test, the effects of training are manifested in muscle size and optimal
pedaling rate. Therefore, we believe it was appropriate to
pool data for the young men and to compare the active boys
with the trained men. Even if the data were not pooled, our
conclusions would persist because the same results could be
shown during growth for the active subjects, and the same results for aging could be shown for the trained men.
The observation that Vopt differed across age groups suggests an age-related difference in the percentage of crosssectional area occupied by type II muscle fibers. Previous
research of growth and aging supports such age-related differences. Lexel and colleagues (27) reported a decrease in
the percentage of type I fibers from 68% at age 5 to 50% at
age 20. Additionally, atrophy of type II fibers in the elderly
has been reported (28,29), which would result in a decrease
in the proportion of muscle cross-sectional area occupied by
type II fibers. Thus, our data for Vopt are supported by studies that used rigorous methods to determine muscle fiber
characteristics.
Magnetic resonance imagery was performed because anthropometric estimates of thigh volume do not account for
intramuscular fat, which is thought to be higher in older
adults (11,30). If intramuscular fat were higher in the older
subjects, the anthropometric technique would overestimate
the lean thigh volume of the older men and bias the scaled
power in favor of the younger subjects. However, in data
from 24 subjects ranging in age from 9 to 69 years, the ratio
of MRIvol to LTVest was independent of age. Thus, intramuscular fat did not increase with age, and consequently the
anthropometric estimate of muscle size seems valid for all
of the subjects in this study. We believe that this finding has
major implications for research because it suggests that
field studies in which only simple anthropometric techniques are logistically possible can provide valid results in
exercise-trained males of any age.
M315
An important difference between LTVest and MRIvol is
that MRIvol included muscle volume at the hip (from gluteal
furrow to iliac crest) that was not included in LTVest. This
volume includes gluteus maximus, which is known to be a
powerful hip extensor. The inclusion of this additional muscle volume may be the reason that MRIvol more accurately
predicted Pmax in these subjects.
Previous investigators have reported that the power of
males increases 27–50% during growth when scaled to
body mass (1–5) and 20–53% (7–9) when scaled to muscle
size. The present data indicate a similar increase in power
scaled to body mass (40%) but less of an increase in power
scaled to muscle volume (10%). Previous investigators of
power and aging have reported that age-related decreases in
power range in magnitude from 6% to 11% per decade in
absolute power and power scaled to body mass or muscle
size (8,10,11,13). The present data indicate a similar decrease of 30% over four decades (7.5% per decade) for total
power and power scaled to body mass, and 20% over four
decades (5% per decade) for power scaled to estimated lean
thigh volume. Indeed, when expressed as power scaled to
body mass (Figure 2), data from the present study are strikingly similar to data reported by Margaria and colleagues
(13) in Figure 5 of that study. When Pmax was scaled by
LTVest ⫻ Vopt, the growth-related increase was eliminated
and the decrease with aging was reduced to 12% over four
decades (3% per decade; Figure 4). Thus, our results suggest that previously reported differences in power during
growth are primarily related to muscle size and fiber type.
Additionally, when those variables are taken into account,
the age-related decrease was less than half of what has been
previously reported.
The youngest subjects in this study (ages 8–12) produced
417 ⫾ 24 watts or 12.4 ⫾ 0.3 W/kg. That value was higher
than previously reported values of 5.9⫺10.2 W/kg (1–5,31)
for boys of similar age. These higher values may be due to
differences in measurement techniques (16) or to the 3 days
of practice performed by our subjects. During the 3 practice
days, the power of 13 of the youngest subjects was monitored to evaluate ergometer settings. Their power increased
44% from 275 ⫾ 18 watts (8.3 ⫾ 0.4 W/kg) on the first
practice test to 394 ⫾ 27 watts (8.3 ⫾ 0.4 W/kg to 11.9 ⫾
0.5 W/kg) during experimental data collection. That 44%
learning-related increase is much larger than the increases
of 11%–13% that have been previously reported for adults
(22,32). Thus, the high power of our young subjects is at
least partially due to the practice bouts that allowed them to
produce truly maximal power. This finding suggests that investigators who do not provide adequate practice may obtain power data that are biased by learning effects.
Our data indicated that 1.7% of the variation in Pmax was
accounted for by age. That 1.7% might be ascribed to age
per se; however, an alternative explanation is an increase in
the proportion of connective tissue within the muscle as a
person ages. Specifically, the volume of intramuscular connective tissue of the older subjects in this investigation
might have been maintained even though the muscle fibers
experienced some atrophy, thus decreasing the proportion
of muscle volume occupied by contractile elements. This
mechanism was not measured in the present investigation,
M316
MARTIN ET AL.
but it may be responsible for some portion of the decrease in
scaled power.
In summary, these data demonstrate that the primary determinants of muscular power are the muscle characteristics:
volume and optimal pedaling rate (as a surrogate measure for
muscle fiber type). Across six decades of age, these muscle
characteristics accounted for over 82% of the variation in
Pmax. Indeed, even the effects of training are accounted for by
those variables. Conversely, age per se only modestly affected Pmax. Magnetic resonance imagery indicated that the
exercise-trained older men in this study did not exhibit higher
levels of intramuscular fat in their upper legs. Consequently,
anthropometric dimensions provide a valid estimate of muscle volume in trained older men and active boys. Practice dramatically affected the maximal power of the boys in this investigation, suggesting that previously reported age-related
changes in maximal power may have been influenced by a
lack of familiarity with the testing protocol. These results
should not be extended to the generally sedentary population.
Rather, our findings represent a best case scenario for active
growth and aging and serve to separate what is possible from
what is assumed to be inevitable.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Acknowledgments
20.
This study was possible only because of the cooperation of several individuals and groups. The authors wish to extend their sincere appreciation to
the following: the participants in this study, for their enthusiastic participation; Dr. Mary Bull, V.A. Miller, Melanie Milner, and Latanya Leply of the
Round Rock Independent School District for allowing us to conduct testing;
Dr. Jean Nichols of San Diego State University for providing us with laboratory space; Suzanne Martin and Nick Brown for help with data collection;
Dr. Bill Walker and Jerry Pate of the Central Texas Imaging Center for providing magnetic resonance imaging; and Dr. Mark Bracker of the University of California at San Diego School of Medicine for providing medical
oversight for the testing of our older participants.
21.
Address correspondence to James C. Martin, Department of Exercise
Science, University of South Carolina, 1300 Wheat St., Columbia, SC
29208. E-mail: [email protected]
25.
22.
23.
24.
26.
References
1. Falk B, Bar-Or O. Longitudinal changes in peak aerobic and anaerobic
mechanical power of circumpubertal boys. Ped Exerc Sci. 1993;5:
318–331.
2. Bedu M, Fellmann N, Spielvogel H, Falgairette G, Van Praagh E,
Coudert J. Force–velocity and 30-s Wingate tests in boys at high and
low altitudes. J Appl Physiol. 1991;70:1031–1037.
3. Falgairette G, Bedu M, Fellmann N, Van-Praagh E, Coudert J. Bio-energetic profile in 144 boys aged from 6 to 15 years with special reference
to sexual maturation. Eur J Appl Physiol. 1991;62:151–156.
4. Inbar O, Bar-Or O. Anaerobic characteristics in male children and adolescents. Med Sci Sports Exerc. 1986;18:264–269.
5. Naughton G, Carlson J, Fairweather I. Determining the variability of
performance on Wingate anaerobic tests in children aged 6–12 years.
Int J Sports Med. 1992;13:512–517.
6. McCammon RW. Human Growth and Development. Springfield, IL:
Charles C. Thomas; 1970.
7. Blimkie CJ, Roache P, Hay JT, Bar-Or O. Anaerobic power of arms in
teenage boys and girls: relationship to lean tissue. Eur J Appl Physiol.
1988;57:677–683.
8. Ferretti G, Narici MV, Binzoni T, et al. Determinants of peak muscle
power: effects of age and physical conditioning. Eur J Appl Physiol.
1994;68:111–115.
9. Sargeant AJ, Dolan P. Optimal velocity of muscle contraction for
short-term (anaerobic) power output in children and adults. In: Rutenfranz J, Mocellin R, Klimt F, eds. Children and Exercise. Champaign,
IL: Human Kinetics; 1986.
10. Chamari K, Ahmaidi S, Fabre C, Masse-Biron J, Prefaut C. Anaerobic
27.
28.
29.
30.
31.
32.
and aerobic peak power output and the force–velocity relationship in
endurance-trained athletes: effects of aging. Eur J Appl Physiol. 1995;
71:230–234.
Grassi B, Cerretelli P, Narici MV, Marconi C. Peak anaerobic power
in master athletes. Eur J Appl Physiol. 1991;62:394–399.
Makrides L, Heigenhauser GJ, McCartney N, Jones NL. Maximal
short term exercise capacity in healthy subjects aged 15–70 years. Clin
Sci. 1985;69:197–205.
Margaria R, Aghemo P, Rovelli E. Measurement of muscular power
(anaerobic) in man. J Appl Physiol. 1966;21:1662–1664.
Bonnefoy M, Kostka T, Arsac LM, Berthouze SE, Lacour JR. Peak
anaerobic power in elderly men. Eur J Appl Physiol. 1998;77:182–
188.
Barany M. ATPase activity of myosin correlated with speed of muscle
shortening. J Gen Physiol. 1967;50(suppl):197–218.
Martin JC, Wagner BM, Coyle EF. Inertial-load method determines
maximal cycling power in a single exercise bout. Med Sci Sports Exerc. 1997;29:1505–1512.
Froese EA, Houston ME. Performance during the Wingate anaerobic
test and muscle morphology in males and females. Int J Sports Med.
1987;8:35–39.
Hautier CA, Linossier MT, Belli A, Lacour JR, Arsac LM. Optimal
velocity for maximal power production in non-isokinetic cycling is related to muscle fibre type composition. Eur J Appl Physiol. 1996;74:
114–118.
Woltring HJ. A FORTRAN package for generalized, cross-validatory
spline smoothing and differentiation. Adv. Engineering Software 1986;
8:104–113.
Forbes GB. Human Body Composition: Growth, Aging, Nutrition, and
Activity. New York: Springer-Verlag; 1999.
Jones PR, Pearson J. Anthropometric determination of leg fat and
muscle plus bone volumes in young male and female adults. J Physiol
(Lond.). 1969;204:63P–66P.
Martin JC, Diedrich D, Coyle EF. Learning effects during maximum
power testing. Int J Sports Med. In press.
Swoap SJ, Caiozzo VJ, Baldwin KM. Optimal shortening velocities
for in situ power production of rat soleus and plantaris muscles. Am J
Physiol. 1997;273:C1057–C1063.
Inbar O, Kaiser P, Tesch P. Relationships between leg muscle fiber–
type distribution and leg exercise performance. Int J Sports Med.
1981;2:154–159.
McCartney N, Heigenhauser GJ, Jones NL. Power output and fatigue
of human muscle in maximal cycling exercise. J Appl Physiol. 1983;
55:218–224.
Sargeant AJ, Hoinville E, Young A. Maximum leg force and power
output during short-term dynamic exercise. J Appl Physiol. 1981;51:
1175–1182.
Lexell J, Sjostrom M, Nordlund AS, Taylor CC. Growth and development of human muscle: a quantitative morphological study of whole
vastus lateralis from childhood to adult age. Muscle Nerve. 1992;15:
404–409.
Coggan AR, Spina RJ, King DS, et al. Histochemical and enzymatic
comparison of the gastrocnemius muscle of young and elderly men
and women. J Gerontol Biol Sci. 1992;47:B71–B76
Lexell J, Taylor CC, Sjostrom M. What is the cause of the ageing atrophy? Total number, size and proportion of different fiber types studied
in whole vastus lateralis muscle from 15- to 83-year-old men. J Neurol
Sci. 1988;84:275–294.
Overend TJ, Cunningham DA, Paterson DH, Lefcoe MS. Thigh composition in young and elderly men determined by computed
tomography. Clin Physiol. 1992;12:629–640.
Docherty D, Gaul CA. Relationship of body size, physique, and composition to physical performance in young boys and girls. Int J Sports
Med. 1991;12:525–532.
Capriotti PV, Sherman WM, Lamb DR. Reliability of power output
during intermittent high-intensity cycling. Med Sci Sports Exerc.
1999;31:913–915.
Received January 3, 2000
Accepted January 5, 2000
Decision Editor: John E. Morley, MB, BCh