Clinical Science (1984) 67, 229-236
229
The time course during 36 weeks’ endurance training of
changes in Poz max. and anaerobic threshold as determined
with a new computerized method
D. A. SMITH
AND
T . V. O’DONNELL
Department of Medicine, Wellington Clinical School of Medicine, Wellington, New Zealand
(Received 2 Febnrmyf22 November 1983; accepted 25 Janumy 1984)
summary
1. Six healthy male subjects followed a programme of endurance training for 36 weeks. At
12 week intervals each underwent an incremental
exercise test to maximum on a treadmill. Minute
ventilation, cardiac frequency, expired and endtidal concentrations of oxygen and carbon
dioxide, oxygen uptake and carbon dioxide output
were measured continuously during each test.
2. Anaerobic threshold (AT) was determined
non-invasively as the onset of sustained increases
in each of the ventilatory equivalent for oxygen,
expired and end-tidal concentrations of oxygen
followed by an increase in ventilatory equivalent
for carbon dioxide after a brief delay due to
isocapnic buffering. A new computerized cumulative-sum method was employed.
3. Sig@cant increases in maximum oxygen
uptake (Vo, m u . ) of 13.6%, AT (32.3%) and
% AT/Vo2 max. (17.0%) and a significant decrease
(10.2%) in cardiac frequency at an oxygen uptake
of 1.O litres/min df,l.o) were observed at the completion of the training programme.
4. .The time courses of the changes for each of
AT, Vo2 max. and fcl.o were not identical during
the training. Compared with Vo, max. the disproportionate increases in AT occurred eallier and
were usually completed within 12 weeks. Voz max.
increased for the fust 24 weeks of training, but
showed no further increase over the final 12
weeks.
decreased through the whole training
period .
Correspondence: Professor T. V. O’Donnell,
Wellington Clinical School of Medicine, Wellington
Hospital, Wellington, New Zealand.
5 . The improved AT after training is more
likely to be related to improved peripheral utilization of oxygen than to an improved oxygen
delivery to muscles.
Key words: aerobic metabolism, anaerobic
threshold, computer processing, endurance training, exercise, oxygen uptake, respiratory gas
exchange, time course.
Abbreviation: AT, anaerobic threshold.
Introduction
Anaerobic threshold (AT) has been defined as
‘the highest oxygen uptake that can be achieved
without a sustained increase in arterial bloodlactate’ [l]. The onset of lactic acidosis, and thus
the AT, is of considerable relevance to athletes
engaging in prolonged severe exercise such as longdistance running. The close correlation, reported
by Costill [2], between speed during long-distance
running and low lactic acid concentration suggests
that an endurance runner is likely to be at an
advantage when his AT represents a high p ~ o portion of his maximum oxygen uptake (Vo,
max .).
The precise cause of the rise in arterial blood
lactic acid concentration above AT remains uncertain. Wasserman [3] has presented several lines
of evidence to suggest that lactic acid concentration in arterial blood is directly related to oxygen
supply to working muscles. Whipp & Mahler [l]
have suggested that enzymatic rate-limitation in
the muscle fibres or fibre-type recruitment
patterns are other possible mechanisms. Studies
230
D. A. Smith and T. V. O'Donnell
involving endurance training, particularly comparison of the time course of changes in AT with
that of increases in Vo2 max. could yield information concerning the extent to which these
increases occur through identical mechanisms.
There have been several reports of increases in
AT after up to 9 weeks of training on a cycle
ergometer [4,5].
The present project involved the following-of
the time course of changes in both AT and Vo2
max. in athletes taking part in a prolonged programme of running (36 weeks).
The onset of lactic acidosis during incremental
exercise was determined indirectly by following
changes in ventilation and respiratory gas
exchange. In previous studies [5, 61 using this
indirect approach there has been considerable
variation between replicate tests, resulting in a low
test-retest correlation. This variability may reflect
changes in the AT of the subjects between the
replicate tests, or arise as a result of the subjective
nature of the methods used to determine AT. As
part of the present project a computerized method
was developed for determining AT from measurements of ventilation and respiratory gas exchange.
Experimental
Subjects
Six healthy male subjects, mean age 34.8 years
7.2; n = 6), mean height 173.0 cm (SD 7.4;
n = 6), mean weight 71 .O kg (SD 8.2; n = 6), who
were active in recreational sport, undertook an
endurance training programme for 36 weeks.
None of the subjects had taken part in regular
physical training for at least 6 months. After the
36 week programme, the mean weight had
decreased to 68.5 kg (SD 7.0; n = 6 ) . The programme involved self-directed jogging through
membership of a local marathon clinic. The
subjects undertook progressively increasing
running averaging 20 km/week (range 6-50 km/
week). This had increased to 73 km/week (range
41-88 km/week) by the completion of the study.
Each subject kept a written record of the distance
covered and associated time in training each week.
After the initial sessions each was advised to train
at close to the maximum pace that he could
sustain for about an hour.
A further 65 healthy subjects, ranging in age
from 16 to 52 years, who did not take part in the
training programme, each performed an incremental exercise test to maximum to allow validation of the computerized method used to
determine AT. A second test was performed
within a week of the first, by 25 of these subjects,
(SD
to assess the repeatability of the determinations
of AT and Vo2max.
A further four of the subjects agreed to have
arterial blood samples withdrawn during the
exercise test. This enabled validation of the indirectly determined AT by comparison with direct
measurements of changes in lactic acid concentration in the arterial blood.
Methods
Each subject carried out an incremental exercise
test to maximum before training commenced and
then at 12 week intervals throughout the training
period. This test was carried out on a treadmill
according to a modified Balke protocol. Initially
treadmill speed remained constant at 5 km/h with
the slope increasing by 21% every 2 min to a
maximum value of 171%. The slope was then
lowered to 10% and the speed increased to 8 km/h.
With the subject now running, the treadmill speed
was increased by 1 km/h or the slope by 24% each
minute until the subject was unable to continue
the test. There was a linear rate of increase of
oxygen uptake (Vo,): mean 1.8 ml min-' kg-'
(SD 0.3; n = 24) for each minute of exercise
(approx. 10 Wlmin).
Minute ventilation (VE), cardiac frequency (f,),
mixed-expired and end-tidal concentrations of
oxygen and carbon dioxide were measured continuously throughout the test by means of a
computerized system (Fig. 1). Subjects breathed
through a low resistance Y breathing valve, with
inspired-gas volume measured with a ParkinsonCowan C D 4 dry-gas meter. An optical shaft
encoder (Ferranti, 2000 pulses per revolution) was
fixed to the indicator needle of the gas meter for
the accurate measurement of tidal -volume. The
non-linear response of the gas meter was corrected
by means of an array, stored on the computer
disk, and relating inspired volume registered by the
gas meter to the number of pulses produced by the
shaft encoder. The onset of expiration was detected
from the shaft encoder signal and respiratory
frequency obtained. Mixed expired concentrations
of oxygen (FEo2)and carbon dioxide (FEco,)
were measured at the distal end of an 8 litres
mixing chamber with Beckman OM-11 and LB-2
gas analysers. End-tidal gas concentrations
(FET02, FETCO2) were measured with a PerkinElmer MGA-1100 mass spectrometer.
In calculating oxygen uptake (Vo2) and carbon
dioxide output (Vco,) a correction was made for
the delay introduced through the use of the
expired-gas mixing chamber. The delay was
variable and obviously shorter at higher levels of
minute ventilation. The correcting algorithm was
Changes in anaerobic threshold during training
23 1
T o subject
FIG. 1. Diagram of equipment used for measurement of ventilation and gas exchange.
Abbreviations: Oz, oxygen analyser with drier; COz, carbon dioxide analyser with
drier; Mixing, 8.0 litres mixing chamber; CPU, PDP 11/40 computer; VDU, visual
display unit; ECG, electrocardiograph.
based on the time taken to inspire a constant
volume (10 litres) of air [7]..To facilitate correction for this delay, average Voz values were also
calculated over the varying time interval corresponding to the same volume of air inspired
(10 litres), rather than averaged over successive
fixed time intervals. Thus data were averaged over
a .20 s interval when inspired minute ventilation
(VI) was 30 l/min but over a 5 s interval when
minute ventilation. was 120 l/min. The equations
used to calculate Voz and Vcozwere the classical
equations described previously [8] modified for
measurement of VI.
The Voz max. was determined as the average
of the last five Voz values before the exercise was
stopped. The mean Voz differences between this
maximum value and those 30 s and 1 min before
the end of exercise were 0.11 l/min (SD 0.13;
n = 24) and 0.16 l/min (SD 0 . ! 3 ; n = 24) respectively. These small changes in Voz,in response to
the increase in work rate over the last minute of
exercise, indicate that the maximum oxygen
uptake was achieved.
The cardiac frequency at an oxygen uptake of
1.0 l/min (f,l.o) was established from the leastsquare! linear-regression of f, against Voz starting
from V02 values not less than 0.7 l/min, up to f,
values that did not exceed 85% of the maximum
f, for the test.
AT was established using a computerized
cumulative-sum method [9]. Cumulative-sum
techniques are appropriate for use where the
objective is to determine changes from an average
level and to establish where these changes
occurred. A series of cumulative sums (cusums)
can be obtained by subtracting a constant quantity
(e.g. the average value) from each original data
value and then summing the differences algebraically for each successive data element. If there is
only random variation of the data around the
mean value (Fig. 2a), there will be no increase in
the cusum values. After an increase in the average
value of the data, there is a sustained increase in
the cusum values (Fig. 2b).
An analysis of the changes during incremental
exercise was carried out for ventilatory equivalent
for C02( V E / V C O ~ventilatory
),
equivalent for Oz
(ri,/VoZ), F'oz, &To2 and respiratory exchange
ratio. ( R ) . The transient undershoot observed in
VE/VO, and R immediately after the onset of
exercise was associated .with the slower response
dynamics of V, and. Vcoz compared with Vo2
[l]. Values for VE/VOZ, F E O ~ ,
and R
beyond this point were assumed to remain
approximately constant until the AT, and to
increase linearly at higher work rates. The time
interval over which these values remained constant
was obtained by a least-squares curve-fitting pro-
D. A. Smith and T. V. O'Donnell
232
(a1
f 401
35
2
0
0
1
10
20
Time (min)
Time (min)
FIG. 2 . ( a ) Ventilatory equivalent for 0 2 (&/ VOZ)
with increasing time of incremental exercise for
one subject. ( b ) Cumulative sums for VE/VCO,( 0 )
and V,/Vo2 ( 0 ) for the same data shown in (a).
n = l toi
(CusumX)i=
1
(~n-7)
where X = VE/V C 0 2 or v E / TO2 and cusum has no
units. AT identified at 4 . V ~ / V c o was
z constant
until 0.
fo, of the three values indicated by the respective
changes in VE/ Vo,,.FE0 2 and FET02.
Values for VE/Vc02 were analysed in the same
way to establish that no increase was observed at
the AT. This ensured that the disproportionate
increase in VE at AT did in fact represent isocapnic
buffering [ 111. Many subjects exhibited a modest
hyperventilation during the early stages of the
incremental exercise test. This resulted in values
for VE/Vc02 which were higher than the subsequent average constant value, for some minutes
after the start of the test, Thus, as shown in Fig.
2b, cusum values for VE/Vco2show an initial rise
before attaining a constant value. These patterns
of change at low work rates were not included in
the analysis of the cusums for detection of the AT.
This computerized determination of AT was
validated in 65 healthy subjects ranging in age
from 16 to 52 years, by comparison with a visual
determination. The visual method was a modification of that reported by Davis et al. [4] where
cusum values were plotted rather than the original
data values [ 9 ] . In all estimations the visual
determination was carried out before the computerized procedure to preclude observer bias. A
two-way analysis of variance was carried out to
detect whether there were significant differences
between the AT values derived from the three
different physiological variables, and also t o detect
whether the computerized and visual methods gave
rise to significantly different results.
The AT determined indirectly was also compared
with changes in lactic acid concentration in arterial
blood. An enzymatic method was used to estimate
the concentration of lactic acid in whole blood
[12]. The time corresponding to the AT was
identified as described above and was then
designated as time zero. The time at which each
blood sample had been collected was then recalculated in terms of time before or after the AT.
To be able to clearly show any pattern of change,
the pre-AT lactic acid concentrations were
averaged for each subject and all lactic acid values
for that subject were divided by this mean value.
The rise in this lactic acid concentration ratio was
then compared graphically with the time identified
as the gas exchange AT.
cedure using the program MODFIT [lo]. The
average value for each variable was calculated over
this time interval determined by the curve-fitting
procedure using a second computer program. This
program was then used to determine the AT
Statistical methods
(Vo,) as indicated by a significant rise (> 2 , s ~ )
Statistical analyses were carried out using the
above the cusum average for each of V E / V O ~ ,
SAS general linear models procedure for analysis
Fb;02and FRTOZ. The cumulative sum was calcuof variance [ 1 3 ] . Duncan's multiple-range test
lated for each data value and the elapsed time and
was used for all comparisons of means. The
Vo2 were determined at which the sustained rise
in the cusum was first observed. The VO, Pearson product moment correlation coefficient
was used for correlation analyses [14].
corresponding t o the AT was taken as the mean
Changes in anaerobic threshold during training
233
Results
Validation of computerized method for AT
A
With either the visual or the computerized
method there were no significant differences
between the AT as indicated by changesjn each of
the three physiological variables VE/VO,, FEO,
and &TO, (P= 0.91). When the AT was taken as
the average of those indicated by these three
variables, the mean AT for the 65 subjects determined by the visual method was 2.66l/min
(SEM 0.08; n = 65) and the mean AT from the
computerized method was 2.67 l/min ( S E M 0.08;
n =65). This difference, 0.014 l/min (SD 0.13;
n = 65), was not significant (P= 0.40).
In four subjects, the AT as determined from the
gas exchange measurements coincided with the
sustained rise in lactic acid concentration in
arterial blood. This concentration remained
approximately constant at work rates below that
corresponding to. the AT but rose significantly at
higher levels of Vo, (Fig. 3) further validating the
computerized cusum method for determining AT.
For AT the mean difference between replicate
tests was 0.001 l/min (SD 0.22; n = 25). The testretest correlation coefficient was 0.95. c e mean
difference between replicate values of Vo, max.
was 0.02 l/min (SD 0.19; n = 25), resulting in a
test-retest correlation coefficient of 0.97.
Results with endurance training
At the commencement of the training the mean
voz max. was 3.61 l/min ( S E M 0.18;. n = 6)
corresponding to 51.3 ml min-' kg-' ( S E M 2.5;
n = 6). The mean AT was 62% of this value, at
2.27 l/min ( S E M 0.17; n = 6). The cardiac frequency (fcl.o) at a standard work rate (Vo, = 1.0
l/min) was 97.8 min-' ( S E M 4.5; n = 6) at the
commencement but decreased throughout the
training period (Table 1).
A
-
-13 -'4
o
4 S li
Time from AT (min)
FIG. 3. Lactic acid concentration, in arterial blood
during incremental exercise, related to the
anaerobic threshold (AT) as indicated by gas
exchange (four subjects: 0, 0, A, D). The lactic acid
concentration values for each subject have been
expressed as a ratio of the mean pre-AT lactic acid
concentration for that subject. Time 0 represents
the AT (computerized non-invasive method). The
onset of the sustained rise in lactic acid concentration ratio coincides with the AT as indicated by
gas exchange.
-is - i z
AT, voz max. and fcl.o showed different
patterns of change over the course of the training.
After the first 12 weeks the mean AT had
increased significantly from 32.2 to 40.3 ml min-'
kg-' (25.2%>P < O . O l ) . After 24 weeks' training
the mean AT was 41.3 ml min-' kg-', which was
also significantly higher than the pre-training value
(28.3%, P < O . O l ) . This AT value, however, was
not significantly different from that after 12
weeks' training. The mean AT of 42.3 ml min-'
kg-' at the completion of the programme was not
significantly different from that at 24 weeks (I' >
0.50) or at 12 weeks (P> 0.40). These short-term
changes in AT occurred rapidly and were largely
completed after 12 weeks of training such that
TABLE1. Mean values (+ SEM) after each 12 weeks interval o f training f o r AT,
VO,
m a . and fclo0 ( n = 6 )
AT (ml min" kg-l)
PO,
max. (ml min-' kg-I)
fc 1.O (min-')
After
After
After
24 weeks
training
training
36 weeks
32.2 f 2.2
A*
51.3 k2.5
40.3 t2.4
41.3 22.5
42.3 2 2.9
B
B
B
56.2 f3.0
58.2 22.1
R
-
r!
58.2 t2.9
97.824.5
94.9 f 3.6
A
91.0; 3.5
81.6i3.6
B
B
C
Before
training
_A A
A
12 weeks
* Mean values having the same letter are not sipsrificantly different (P > 0.05).
training
C
D. A. Smith and T. V. O'Donnell
234
TABLE2. he-training, peak and final values of % AT/VOz max. and stage o f training
when peak observed (six subjects)
Subject
he-qhing
%AT/VO,max.
Peak
%AT/VO,max.
1
57
18
2
3
4
5
6
60
69
64
63
61
19
81
64
I1
88
increases after the first 12 weeks were small and
did not reach statistjcal significance.
In contrast the Vo2max. continued to increase
throughout the first 24 weeks .of the training.
After the first 12 weeks mean Vo2 max. had increased by 9.6% (P<O.Ol) from 51.3 to 56.2
mlmin-' kg-'. A further increase of 3.6% (P<
0.02) was observed after the second 12 weeks, to a
value of 58.2 ml min-' kg-'. This represented a
13.5%(P< 0.001) increase above the pre-train-ing
mean Po2 max. value. No further increase in Voz
max. was observed after 36 weeks from the mean
value at 24 weeks (P> 0.90).
Thus in three subjects %AT/vo2max. reached
peak values by 12 weeks and decreased subsequently. In the remainder peak values were
reached later (Table 2).
Values of fcls0 showed a steady decrease
throughout the whole training period, but in
contrast to the pattern of change for AT, the
greatest decrease of fcla0 was observed over the
final 12 weeks. The mean heart rate value of
94.9 min-' after the first 12 weeks was not
significantly different from the pre-training value
of 97.8 min-' (P>O.50). After 24 weeks the
mean fcl,o value had decreased to 91.0 min-'. This
was a significantly lower value than that observed
before training (7.0%, P<O.O5), but was not
significantly different from that after 12 weeks.
In the final 12 weeks of training, values of fcl.o
continued to decrease significantly (3.7%, P =
0.05), compared with those observed after 24
weeks.
At the end of the training period there was a
significant increase .in AT (32.3%), Vo2 max.
(13.6%) and %AT/Vo2 max. (17.0%), and fCl.o
showed a significant decrease (10.2%) compared
with pre-training values.
Discussion
There are many approaches to the assessment of
the capacity for undertaking physical exercise.
Week
observed
12
12
12
0-36
24
36
Fbal
% AT/ VO, max.
61
61
I7
64
15
88
Poz max. is such a yardstick for maximal performance by highly motivated healthy subjects.
AT provides a basis for a submaximal assessment
at a level of oxygen uptake which is of metabolic
significance in energy production and above which
changes in breathing or respiratory symptoms
occur. A lower AT has been demonstrated in some
patients with symptoms of cardiac disease [ l l ] .
Healthy subjects have limitations in the sustaining
of exercise at work rates above the AT [15].
Indices such as fcl.o reflect particularly circulatory performance at a moderate level of physical
exercise.
AT can be determined from the direct measurement of lactic acid concentration in multiple
samples of blood during progressively increasing
exercise [ 161. Noninvasive methods, obviously
preferred, are based on the detection of a change
in VE/VO, followed closely by a change of VE/
Pco,. The changes may be followed breath-bybreath but for reasons of accuracy and stability
[7, 91, we have calculated average values over the
time for successive 10 litres of inspired gas. The
identification of AT requires the detection.of the
earliest significant sustained increase of V,/Vo,
and VE/Vc02 above base line. Some difficulty may
arise from a scatter of values with the subject at
rest or during the initial stages of incremental
exercise. Our method utilized the computer in the
identification of significant changes from base-line
values after its use in least-squares curve-fitting
procedures and the cumulative-sum statistical
technique. There was no difference between the
AT values obtained by this computerized cusum
method and those determined by visual methods
on independent occasions (P= 0.40). If AT is to
be compared for different modes of exercise [6]
or before and after training [4], a reliable and
objective method is desirable. Validation was
achieved through comparison with lactic acid
changes in arterial blood. Repeatability of the AT
measurements was similar to that which has been
achieved for Vo2 m a . [4, 171, and compared
Changes in anaerobic threshold during training
favourably with that for AT obtained by visual
methods [4,6].
The subjects in the present study were active
although none was previously involved in regular
training. Values for AT and Vo2 max. observed at
the commencement of the training were considerably higher than those reported by Davis
et al. [4] for sedentary middle-aged subjects. We
have found previously that AT values were not
significantly different for active subjects within
the three age decades, 25-34, 35-44 and 45-54
years [9]. We do not then attribute the higher AT
of our subjects to an age difference. Their pretraining values were similar to those of Davis et al.
[6] for active students and probably reflect a
greater level of regular physical activity.
Our evidence of co-operation with training programmes depended on the written records kept by
each subject. The mean level per week covered at
fnst 20 km, occupying about 2 h, but increased to
a mean of 73 kmlweek, occupying 6-7 h, by the
end of the programme. When each subject had
completed his fust marathon the mean Vo2
during the event. was estimated using the relationship between Vo2 and the speed of outdoor
running described by Pugh [181. There was a clpse
correlation (r = 0.98) between this estimated Vo2
and the AT value measured at the end of trai@g
[9]. We have assumed that a similar level of Vo2
during the training runs of more than an hour was
likely. T h i s training intensity was lower than for
some previous studies although the duration per
week was considerably longer [4,5].
Our subjects showed a smaller mean percentage
increase in Vo2max. than those of Davis et al. [4],
who trained at a regular intensity of 75-85%
Vo2 max. for 3-4h weekly. After training, the
mean percentage rise of AT was similar to values
in those. who trained at a higher load. The smaller
rise of Vo2max. is likely to reflect the higher level
of pre-training fitness among our men [19]. The
increases of AT are not related so clearly to this.
The greater increase of AT and its more rapid
t+ecourse of change compared with that for
Vo, max. raise the probability that the increase of
each is dependent principally on a different
mechanism.
Po, max. is limited particularly by the blood
flow to the exercising skeletal muscles [20]. At
maximum exercise after training an increased
supply of oxygen to these accompanies the
increase of. cardiac output associated with an
increased Vo2 max. It has been tempting to
attribute the improvement of AT to an altered
distribution of the cardiac output, favouring the
exercising muscles at a particular work rate. However, Clausen et al. [21] reported a decrease of
235
cardiac output at a given submaximal work-rate
after training but, at the same time, an increased
splanchnic-hepatic blood flow. Such results
would not favour significant increases of blood
flow to exercising skeletal muscles occurring as a
result of training. Support for this conclusion is
provided by the decreases of blood flow in vastus
lateralis muscles measured in men by '=Xe
clearance techniques at a f s e d submaximal work
rate before and after 6 weeks' training on a bicycle
ergometer [22].
An alternative explanation for the improvement of AT with training involves an improved
metabolic efficiency in the exercising muscles
associated with a greater oxygen extraction from
arterial blood [19]. Significant increases in the
activities of oxidative enzymes in muscle have
been observed after training [22, 231. Andersen &
Henriksson [23] reported a 40%increase in the
activities of succinate dehydrogenase and cytochrome oxidase after a programme of training at
about 80% Vo2 max. over 8 weeks. They showed
an increased capillary density in the quadriceps
femoris muscle with a time course of increase
similar to that for Vo2 max. The activities of the
qxidative enzymes increased more rapidly than
Vo2 max. This more rapid time course may be
relevant to that for the AT cornpared with the
Vo2 max. in the present study. The development
of an increased capillary density with training
might have assisted the extraction of oxygen by
providing a shorter pathway for diffusion. Should
this have been the explanation for the increase in
AT with training then we would have expcted the
time courses of the changes of AT and Voz max.
to have been similar in each subject. This was not
so in our subjects, in three of whom there was a
clear dissociation between increases in AT and
those in Voz max.
In these subjects in particular training at a level
close to AT over the 36 weeks period, no sigtllfic p t increase occurred in AT after 12 weeks or in
Vo, max. beyond 24 weeks in spite of the
sustained training effort. The percentage increase
of AT was similar to that observed by Davis et al.
[4] but was achieved at a training intensity lower
than the 7 5 4 5 % of Vo2 max. undertaken by his
subjects.
The fcl.o is a complex index subject to many
humoral and neurological influences. In spite of
the widespread use of decreasing cardiac frequency
at a standard work rate in indicating the effects of
training, its changes are less and have a slower
time course than AT. A change in AT with training
is not only more substantial but provides an advantage to the distance runner wishing to undertake exercise based on prolonged aerobic activity.
236
D. A . Smith and T. V
O O’DonneN
Acknowledgments
This study was supported by grants from the New
Zealand Lottery Board and the War Pensions
Medical Research Trust Board of New Zealand. We
acknowledge also with thanks the statistical
contribution from Mr N. Pearce and the secretarial
assistance of Ms Jacqui Williams.
References
1. Whipp, B.J. & Mahler, M. (1980) Dynamics of pulmonary gas exchange during exercise. In: Pulmonary
Gas Exchange, vol. 11. Academic Press Inc., New York.
2. Costill, D.L. (1970) Metabolic responses during
distance running. Journal of Applied Physiology, 28,
251-255.
3. Wasserman, K. (1981) Physiology of gas exchange and
exertional dyspnoea. Clinical Science, 61, 7-13.
4. Davis, J.A., Frank, M.H., Whipp, B.J. & Wasserman,
K. (1979) Anaerobic threshold alterations caused by
endurance training in middle-aged men. Journal of
Applied Physiology: Respiratory Environmental and
Exercise Physiology, 46,1039-1046.
5. Ready, A.E. & Quinney, H.A. (1982) Alterations in
anaerobic threshold as the result of endurance training
and detraining. Medicine and Science in Sports and
Exercise, 14, 292-296.
6 . Davis, J.A., Vodak, P., Wilmore, J.H., Vodak, J. &
Kurtz, P. (1976) Anaerobic threshold and maximal
aerobic power for three modes of exercise. Journal
of Applied Physiology, 41,544-550.
7.Hughson, R.L., Kowalchuk, J.M., Prime, W.M. &
Green, H. J. (1980) Opencircuit gas exchange analysis
in the non-steady-state. Chadian Journal of Applied
Sports Science, 5 , 15-18.
8. Jones, N.L., Campbell, E.J.M., Edwards, R.H.T. &
Robertson, D.G. (1975) Clinical Exercise Testing.
W. B. Saunders Co., Toronto.
9. Smith, D.A. (1981) Non-invasive methods for estimation of anaerobic threshold. In: The Anaerobic
Threshold, its Objective Assessment - Effects of age,
Training and Athletic Background. PhD Thesis, University of Otago, Dunedin.
10. McIntosh, J.E.A. & McIntosh, R.P. (1980) Mathematical Modelling and Computers in Endocrinology.
Springer-Verlag, Berlin.
11. Wasserman, K. & Whipp, B.J. (1975) Exercise physiology in health and disease. American Review of
Respiratory Disease, 112,219-249.
12. Hohorst, H.J.L. (1965) L(+)-Lactate. In: Merhods of
Enzymatic Analysis. Ed. Bergmeyer, H.U. Academic
Press, New York.
13. SAS Institute Inc. (1979) SAS Users Guide: 1979
edition. Publications Division, SAS Institute Inc.,
Raleigh, North Carolina.
14. Snedecor, G.W. & Cochran, W.G. (1967) Statistical
Methods, 6th edn. Iowa State University Press, Ames,
Iowa.
15. Wasserman, K., Van Kessel, A.L. & Burton, G. (1967)
Interaction of physiological mechanisms during
exercise. Journal of Applied Physiology, 22, 71-85.
16. Ivy, J.L., Withers, R.T., Van Handel, P.J., Elger, D.H.
& Costill, D.L. (1980) Muscle respiratory capacity and
fibre type as determinants of the lactate threshold.
Journal of Applied Physiology: Respiratory Environmental Exercise Physiology, 48, 523-527.
17. Froelicher, V.F. Jr., Brammell, H., Davis, G., Noguera,
I., Stewart, A. & Lancaster, M.C. (1974) A comparison of three maxitllal treadmill exercise protocols.
Journal of Applied Physiology, 36, 720-725.
18. Pugh, L.G.C.E. (1970) Oxygen intake in track and
treadmill running with observations o n the effect of
air resistance. Journal of Physiology (London), 207,
823-835.
19. Saltin, B., Blomqvist, G., Mitchell, J.H., Johnson,
R.L. Jr., Wildenthal, K. & Chapman, C.B. (1968)
Response to exercise after bed rest and after training.
Circulation, 38 (Suppl. 7), 1-78.
20. Reybrouck, T., Heigenhauser, G.F. & Faulkner, J.A.
(1975) Limitations to maximum oxygen uptake in
arm, leg and combined arm-leg ergometry. Journal
of Applied Physiology, 38,774-779.
21. Clausen, J.P., Klausen, K., Rasmussen, B. & TrapJensen, J. (1973) Central and peripheral circulatory
changes after training of the arms or legs. American
Journal of Physiology, 225,675-682.
22. Bergman, H., Bjorntorp, P., Conradson, T.-B., Fahlen,
M., Stenberg, J. & Varnauskas, E. (1973) Enzymatic
and circulatory adjustments t o physical training in
middle-aged men. European Journal of Clinical
Investigation, 3,414-418.
23. Andersen, P. & Henriksson, J. (1977) Capillary
supply of the quadriceps femoris muscle of man:
adaptive response to exercise. Journal of Physiology
(London), 270,677-690.