This article was downloaded by: [University Of Maryland]

On: 15 October 2014, At: 03:28

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office:
Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Sports Sciences

Publication details, including instructions for authors and subscription
information:
http://www.tandfonline.com/loi/rjsp20

The relationship between selected physiological

variables of rowers and rowing performance as
determined by a 2000 m ergometer test
M.J. Cosgrove , J. Wilson , D. Watt & S.F. Grant
Published online: 09 Dec 2010.

To cite this article: M.J. Cosgrove , J. Wilson , D. Watt & S.F. Grant (1999) The relationship between selected
physiological variables of rowers and rowing performance as determined by a 2000 m ergometer test, Journal of
Sports Sciences, 17:11, 845-852, DOI: 10.1080/026404199365407

To link to this article: http://dx.doi.org/10.1080/026404199365407

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”)
contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors
make no representations or warranties whatsoever as to the accuracy, completeness, or suitability
for any purpose of the Content. Any opinions and views expressed in this publication are the opinions
and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of
the Content should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings,
demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising
directly or indirectly in connection with, in relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial
or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or
distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can
be found at http://www.tandfonline.com/page/terms-and-conditions
 Jour nal of Sports Sciences, 1999 , 17, 845± 852

The relationship between selected physiological

variables of rowers and rowing performance as
determined by a 2000 m ergometer test
1 1 2 1
M .J. CO SGROVE, J. WILSON , D. WATT and S.F. GRANT *
1 2
Institute of B iom edical and L ife Sciences and Departm ent of Statistics, U niversity of G lasgow, G lasgow, U K

Accepted 21 M arch 1999

Downloaded by [University Of Maryland] at 03:28 15 October 2014

The aim of this study was to establish the relationship between selected physiological variables of rowers and
rowing performance as determ ined by a 2000 m tim e-trial on a Concept II M odel B rowing ergometer. The
participants were 13 m ale club standard oarsmen. Their m ean (± s) age, body mass and height were 19.9 ± 0.6
years, 73.1 ± 6.6 kg and 180.5 ± 4.6 cm respectively. The participants were tested on the rowing ergom eter to
determine their maxim al oxygen uptake (VÇ O 2m ax ), rowing economy, predicted velocity at VÇ O 2max , velocity and
- 1
VÇ O 2 at the lactate threshold, and their velocity and VÇ O 2 at a blood lactate concentration of 4 mm ol ´ l . Percen t
body fat was estimated using the skinfold method. The velocity for the 2000 m perform ance test and the
predicted velocities at the lactate threshold, at a blood lactate concentration of 4 m mol ´ l- 1 and at VÇ O 2 ma x were
4.7 ± 0.2, 3.9 ± 0.2, 4.2 ± 0.2 and 4.6 ± 0.2 m ´ s - 1 respectively. A repeated-measures analysis of variance showed
that the three predicted velocities were all signi® cantly diþ erent from each other (P < 0.05). The VÇ O 2ma x and
lean body m ass showed the highest correlation with the velocity for the 2000 m tim e-trial (r = 0.85). A stepwise
m ultiple regression showed that VÇ O 2 max was the best single predictor of the velocity for the 2000 m tim e-trial;
a model incorporating VÇ O 2m ax explained 72% of the variability in 2000 m rowing performance. Our results
suggest that rowers should devote time to the improvem ent of VÇ O 2 max and lean body m ass.

K eyw ords : lactate threshold, m aximal oxygen uptake, rowing, rowing economy, velocity at VÇ O 2m ax .

Introduction D uring a 2000 m race, rowers are m ainly dependent

Researchers have exam ined such physiological factors
as the im portance of lactate exchange and rem oval
(M essonnier et al., 1997), the eþ ects of training on
rowing ergom eter perform ance (Wom ack et al., 1996),
the eþ ectiveness of altitude training on sea-level rowing
perform ance (Jensen et al., 1993) and the determ ination
of the anaerobic threshold using a non-invasive m ethod
(Drogh etti, 1986). Although these studies have contri-
buted to the scienti® c understanding of row ing, lim ited
inform ation is available on the relationship between
physiological variables of rowers and rowing perfor-
m ance. Relating physiological variables to perform ance
could be valuable for designing training program m es
and for team selection.
* Address all correspondence to Stan G rant, Institute of Biom edical
and Life Sciences, U niversity of G lasgow, 64 O ak® eld Avenue,
G lasgow G 12 8LT, U K.
on aerobic m etabolism , as the duration of the race is 6± 7
m in, depending on the type of boat and the standard of
the rowers. The relative contribution to the total energy
dem and of a 2000 m race is about 70% aerobic and 30%
anaerobic (M ickelson and H agerm an, 1982; Roth et al.,
1983).
T he perform ance of a crew can be m easured by their
2000 m race tim e. However, it is very diY cult to assess
physiological param eters during `on-water’ row ing. The
im practicality of m easuring physiological perform ance
during row ing has led to the use of rowing ergom eters
for the assessm ent of row ing perform ance. T he m ost
w idely used ergom eter for training purposes is the
C oncept II wind resistance braked rowing ergom eter
(C oncept II, N ottingham , U K). Craven et al. (1993)
analysed the rowing stroke on a C oncept II rowing
ergom eter and concluded that it `provides a close
approxim ation to the m ovem ents of the rowing stroke,
and allows accurate m easurem ents of the physiological

Jour nal of Sports Sciences ISSN 0264-041 4 print/ISSN 1466-447 X online Ó Taylor & Francis Ltd
 846
changes produced by the work’ . The Concept II allows
standardization of testing, which is not possible during
`on-water’ row ing.
Several studies have com pared the physiological
characteristics of elite rowers to less skilled rowers,
suggesting that elite rowers have a higher m axim al
oxygen uptake (VÇ O 2max ; Secher et al., 1982), a larger
percentage of slow -twitch m uscle ® bres (Roth et al.,
1983), a greater body m ass (Secher, 1983) and a higher
VÇ O 2m ax at a blood lactate concentration of 4 m m ol ´ l-
1
than less skilled rowers (Roth et al., 1983; M arx, 1988).
Kram er et al. (1994) studied various descriptive, ® eld
and laboratory variables of 20 C anadian intercollegiate
oarswom en and exam ined the relationship of these
variables to perform ance. O f the variables tested, only
abso lute VÇ O 2max had a correlation coeY cient greater
Downloaded by [University Of Maryland] at 03:28 15 October 2014
than 0.71 with perform ance.
Economy is de® ned as the volum e of oxygen con-
sum ed by the working m usculature at a given steady-
state workload. This has received little attention in
row ing research. O nly one study (M acLennan et al.,
1994) has investigated the economy of rowing; un-
fortunately, the participants in the study had no row-
ing experience and were taught how to use a row ing
ergom eter. In contrast to research on row ing, several
studies have exam ined running economy. H owever,
the evidence is con¯ icting as to w hether running
economy is a good predictor of perform ance. T he m ore
heterogeneous the group, the better the prediction of
endurance perform ance is likely to be (W ilcox and
Bulbulian, 1983).
T he velocity at VÇ O 2m ax has been de® ned in a variety
of ways and several protocols have been used to deter-
m ine it (Hill and Rowell, 1996). D aniels (1985) used
the linear extrapolation of the subm axim al steady-state
oxygen uptake± velocity relationship to estim ate the
velocity at VÇ O 2 max in runners. D aniels (1985) and H ill
and Rowell (1996) both reported that the velocity at
VÇ O 2m ax is a good predictor of endurance running per-
form ance; no study has exam ined this velocity in rowers.
In endurance running and cycling events, blood
lactate variables have been shown to be the best pre-
dictors of perform ance over a wide range of distances
(Weltm an, 1995). Wolf and Roth (1987) found that
the power which elicits a blood lactate concentration
of 4 m m ol ´ l- 1 is the best predictor of com petition
perform ance in trained rowers, especially those w ho
com pete in sm all boats. O thers have demonstrated that
rowers w ith high absolute VÇ O 2 at a blood lactate con-
centration of 4 m m ol ´ l - 1 perform better in 6 or 7 m in
m axim al tests than rowers with low VÇ O 2 at the sam e
blood lactate concentration (Roth, 1979; M arx, 1988).
The higher VÇ O 2 of less skilled rowers at a blood lactate
concentration of 4 m m ol ´ l- 1 suggests that they are likely
to fatigue earlier than elite rowers. A possible m echan-
Cosgrove et al.
ism for this fatigue is a decrease in intracellular pH
resulting from glycolysis and H + accum ulation, which
has been shown to interfere w ith the m yosin and actin
interaction and to result in a decrease in force pro-
duction (Jones and Round, 1990).
T he aim of this study was to exam ine the relationship
between selected physiological variables of m ale rowers
and rowing perform ance as determ ined by a 2000 m
tim e-trial.
M ethods
Participants
Thirteen m ale rowers from G lasgow U niversity Boat
Club volunteered to participate. T hey all had experi-
ence of com petitive rowing over 2000 m , ranging from 1
to 9 years, and were club standard oarsm en of var ying
ability. All participants trained regularly on a Concept II
row ing ergom eter. Inform ed consent was obtained from
all participants and ethical approval was received from
the U niversity of G lasgow E thical Com m ittee. T he
characteristics of the participants are given in Table 1.
Protoco l
The participants perform ed three tests on separate
days with at least 2 days between tests; however, each
individual com pleted all three tests within 14 days. O n
day 1, height, body m ass, percent body fat (Durnin
and Wom ersley, 1974) and VÇ O 2max were m easured. O n
day 2, a lactate pro® le and rowing economy test were
perform ed; on day 3, a perform ance test was com pleted.
Equipm ent
A Concept II M odel B rowing ergom eter (C oncept II,
N ottingham , U K) was used for all tests. D uring testing,
the larger of the two drive cogs was used and the vanes
were kept fully closed.
H eart rate was m onitored using a Sport Tester
TM
PE3000 heart rate m onitor (Polar Electro, K empele,
Finland). As a precautionary m easure during the
m axim al test, a H ewlett Packard 43200A ECG m onitor
was used.
Table 1 C haracteristics of the participants (mean ± s)
Age (years) 19.9 ± 0.6
Height (cm) 181 ±5
Body mass (kg) 73.1 ± 6.6
Body fat (%) 16.1 ± 3.3
Lean body mass (kg) 60.9 ± 5.2
- 1
VÇ O 2 max (l ´ m in ) 4.5 ± 0.4
 Rowing performance and physiological variables of rowers
For gas analysis, expired air was collected in D ouglas
bags. T he sam ples were passed through a dr ying tube
and then analysed using a PK M organ 801D (M organ,
Rainham , U K) and Ser vom ex 570A (Ser vom ex,
C row borough, U K) system to determ ine carbon
dioxide and oxygen content respectively. Volum e was
calculated using a Parkinson Cowan volum e m eter
(C ranlea, Birm ingham , U K) calibrated against a Tissot
spirom eter (Collins, M assachusetts, U SA). F rom these
data, VÇ O 2 and respiratory values were calculated.
W hole blood was analysed for blood lactate using an
Analox G M 7 (Analox Instrum ents, London, U K).
Procedures
Downloaded by [University Of Maryland] at 03:28 15 October 2014
T he tim e to com plete 500 m at the current intensity and
stroke rate (500 m split tim e) was calculated and dis-
played on the ergom eter. T he stroke rate was lim ited to
24± 28 strokes ´ m in - 1 at 500 m split tim es of 2 m in or
m ore. At 500 m split tim es less than 2 m in, the stroke
rate was increased to 28± 32 strokes ´ m in - 1 . T his allowed
the intensity to be increased so that the 500 m split tim e
could be reduced to the desired value.
T he restriction on stroke rate resulted in the exercise
intensity being kept constant for each stage and for
all participants (M ickelson and H agerm an, 1982).
T here were slight deviations in intensity, owing to a
com bination of individual variation in stroke rate and
the inability to keep 500 m split tim e constant. Such
problem s are inherent in the design of the rowing
ergom eter and the m echanics of the row ing stroke
(M ickelson and Hagerm an, 1982).
A warm -up of 6 m in at a 500 m split tim e of 2 m in
30 s was perform ed by all participants before all tests.
T he participants then rested for 6 m in, during w hich
tim e they perform ed stretching exercises.
Percent body fat. The percent body fat of each par-
ticipant was estim ated using skinfold m easurem ents.
Skinfolds were m easured using a H oltain caliper
(C rosswell, C rymych, UK ) following the m ethod of
D urnin and Wom ersley (1974).
VÇ O 2max test. The participants perform ed a continuous
increm ental test to volitional exhaustion. The target
duration of the test was 8± 14 m in. The test began
w ith each participant exercising at a 500 m split tim e
of 2 m in 30 s. T hereafter, the split tim e was decreased
by 5 s each m inute until the participant reached
volitional exhaustion. H eart rate was m easured after
50 s at each exercise intensity. O nce heart rate reached
170 beats ´ m in - 1, sam pling of expired gas began.
Exp ired gas sam ples were collected for 1 m in at each
exercise intensity after this tim e.
847
Blood sam ples were taken from the tip of the thum b
for lactate analysis at 1, 3 and 5 m in after the VÇ O 2max
test. Two analyses were perform ed on each sam ple and
the m ean of the two was taken as the blood lactate
concentration.
D eterm ination of the velocity at VÇ O 2max. T he velocity at
VÇ O 2max was calculated following the m ethod recom -
m ended by Hill and Rowell (1996) to be the m ost
accurate and appropriate for evaluating perform ance;
that is, the linear extrapolation of the subm axim al
velocity± VÇ O 2 relationship.
B lood lactate pro® le and rowing economy test. Blood
lactate concentration and rowing economy were
m easured using a discontinuous increm ental sub-
m axim al test. T he test involves a target of ® ve stages of
5 m in exercise each. The 500 m split tim e for the ® rst
stage was determ ined for each individual from the
results of his VÇ O 2max test. The starting split tim es ranged
from 2 m in 30 s for the rowers w ith the lowest VÇ O 2max
to 2 m in 15 s for those with the highest VÇ O 2 max . Each
subsequent stage was 5 s faster than the previous one.
A 30 s rest was allowed between each stage, during
w hich a blood sam ple was taken from the tip of the
thum b for lactate analysis. T he test continued until
the average blood lactate for a given stage reached
4 m m ol ´ l- or m ore.
1
O ne-m inute expired gas sam ples were collected
between the third and fourth and the fourth and ® fth
m inutes of each stage to assess row ing econom y. The
m ean of these two values was recorded for each stage.
H eart rate was m easured 50 s into each m inute of the test.
D eterm ination of the physiological parameters. Velocity
and VÇ O 2 were determ ined at the lactate threshold and
at a blood lactate concentration of 4 m m ol ´ l- 1. Blood
lactate concentrations were plotted against the 500 m
split tim es at w hich they were produced. To achieve
objectivity, a m athem atical regression analysis was used
to m odel the relationship between lactate and rowing
velocity (O rr et al., 1982). T he lactate threshold was
de® ned as the in¯ ection point of the lactate pro® le w hich
indicated a sharp increase in blood lactate concentration
(K inderm ann et al., 1979). A blood lactate concentra-
tion of 4 m m ol ´ l- 1 was determ ined from the velocity and
VÇ O 2 that corresponded to a blood lactate of 4 m m ol ´ l- .
1
Two horizontal lines were drawn across from the pre-
determ ined lactate threshold and 4 m m ol ´ l - 1 concen-
tration to intercept with the blood lactate values on the
vertical axis. From these points, lines perpendicular to
the intercepts were drawn down to the horizontal axis to
obtain the velocity and VÇ O 2 for individual lactate
thresholds and 4 m m ol ´ l - 1 blood lactate concentrations.
 848
Rowing performance. Rowing perform ance was assessed
by a 2000 m tim e-trial on the Concept II row ing
ergom eter. H eart rate was m easured 50 s into each
m inute and on com pletion of the test. Thum b prick
blood sam ples for lactate analysis were taken 1, 3 and
5 m in after the 2000 m for com parison w ith values
obtained for the VÇ O 2max test.
Statistical analyses
Initially, a repeated-m easures analysis of variance
with an appropriate Bonferroni m ultiple com parisons
follow-up was perform ed to identify any signi® cant
diþ erences at the P < 0.05 level between the velocity
for the 2000 m tim e-trial and the predicted velocities at
the lactate threshold, at a blood lactate concentration of
Downloaded by [University Of Maryland] at 03:28 15 October 2014
4 m m ol ´ l- and at VÇ O 2 max .
1 Predicted velocity at VÇ O 2max
Pearson correlation coeY cients were used to exam ine
the interrelationships between variables. F rom these
sim ple linear regressions, a subset of the initial physio-
logical variables (the nine variables in Table 3 show ing
a statistically signi® cant relationship with velocity in
the 2000 m tim e-trial, P < 0.05) were entered into a
forward stepwise m ultiple linear regression using a
variable selection procedure w ith tim e-trial velocity as
the response variable. This showed which of these
variables were collectively im portant in the prediction
of tim e-trial velocity.
Results
VÇ O 2max test
D espite the relatively low body m asses of the partici-
pants, som e high VÇ O 2max values were recorded, the
sam ple m ean being 4.5 ± 0.4 l ´ m in - . Absolute VÇ O 2m ax
1
- 1
(l ´ m in ) correlated well w ith velocity in the 2000 m
tim e-trial (r = 0.85), the participants with higher values
generally perform ing quicker (Fig. 1a). However, the
relative values (m l ´ kg - ´ m in - ) did not correlate well
1 1
with tim e-trial velocity (r = 0.13). T he correlation
between VÇ O 2m ax divided by lean body m ass and the
tim e-trial velocity was r = 0.01; that between VÇ O 2m ax
expressed in m l ´ kg - ´ m in - and tim e-trial velocity was
2/3 1
r = - 0.05.
Rowing economy
As a result of the variation in starting velocity and
individual diþ erences in ability, only three stages at
the sam e velocity were com pleted by all participants.
Econom y values were spread over a narrow range at all
three velocities (Table 2). The VÇ O 2 at a 500 m split tim e
of 2 m in 5 s resulted in the m ost signi® cant negative
correlation w ith velocity in the 2000 m tim e-trial
Cosgrove et al.
(r = 0.62). T he other two intensities com pleted by all
participants showed poorer correlations with tim e-trial
velocity: r = - 0.20 at a 500 m split tim e of 2 m in 15 s and
r = - 0.51 at a 500 m split tim e of 2 m in 10 s.
To assess the accuracy of the m ethod used to deter-
m ine row ing economy, a related t-test was perform ed on
the VÇ O 2 values recorded in the fourth and ® fth m inute
of each stage for each participant to identify any sig-
ni® cant diþ erence between the two values. This would
con® rm if a steady-state oxygen consum ption had
been reached. The diþ erences in oxygen consum ption
between the fourth and ® fth m inute of each stage did
not reach statistical signi® cance for any participant,
indicating that a steady state had been attained.
The predicted velocity at VÇ O 2max correlated well with
tim e-trial velocity (r = 0.77; Fig. 1b). In general, the
participants who achieved a high predicted velocity at
VÇ O 2m ax produced faster 2000 m tim es.
B lood lactate pro® le
The VÇ O 2 at the lactate threshold had a non-signi® cant
relationship with tim e-trial velocity (r = 0.39), whereas
VÇ O 2 at a blood lactate concentration of 4 m m ol ´ l-
1
correlated well tim e-trial velocity (r = 0.68). T he parti-
cipants with low values of these two VÇ O 2 param eters
tended to attain these blood lactate concentrations at
slow speeds. T he predicted velocity at a blood lactate
concentration of 4 m m ol ´ l- 1 had a fairly strong relation-
ship with tim e-trial velocity (r = 0.73).
Perfor mance test
The 2000 m tim es varied widely, ranging from 6 m in 30
s to 7 m in 45 s, w ith speeds between 4.3 and 5.1 m ´ s - 1 .
To m aintain a low 500 m split tim e, the individual m ust
be able to produce a large force during the stroke. Lean
body m ass was a signi® cant factor (r = 0.85; Fig. 1c),
as it identi® ed individuals with greater m uscle m ass.
Statistical analyses
A repeated-m easures analysis of variance showed that
the velocities for the 2000 m tim e-trial, at the lactate
threshold, at a blood lactate concentration of 4
m m ol ´ l- 1 and at VÇ O 2 max were diþ erent (P < 0.001). An
appropriate follow-up analysis showed that the velocity
for the tim e-trial was signi® cantly diþ erent from the
velocities at the lactate threshold and at a blood lactate
concentration of 4 m m ol ´ l- 1 , but not signi® cantly dif-
ferent from the velocity at VÇ O 2m ax . Follow -up analysis
also showed that the velocities at the lactate threshold, at
 Rowing performance and physiological variables of rowers
Downloaded by [University Of Maryland] at 03:28 15 October 2014
a blood lactate of 4 m m ol ´ l- 1 and at VÇ O 2m ax were all
signi® cantly diþ erent from one another.
Table 3 presents correlation coeY cients and P-valu es
for a test of w hether the variables had a statistically
signi® cant relationship with velocity in the 2000 m
tim e-trials. T he m ost signi® cant relationships were
between VÇ O 2max and tim e-trial perform ance (r = 0.848,
P < 0.001) and between lean body m ass and tim e-trial
perform ance (r = 0.848, P < 0.001). T here were also sig-
ni® cant relationships between tim e-trial perform ance
and endurance in the VÇ O 2 max test (r = 0.791, P = 0.001),
the highest value obtained in the velocity at VÇ O 2 max test
(r = 0.773, P = 0.002), and velocity at a blood lactate
concentration of 4 m m ol ´ l - 1 (r = 0.734, P = 0.004). L ess
signi® cant relationships were noted between tim e-trial
perform ance and body m ass, the lactate m axim um 5
m in after the VÇ O 2m ax test, and the lactate m axim um
849
F igu re 1 Fitted line plots showing the relationships between
(a) VÇ O 2 max , (b) velocity at VÇ O 2ma x and (c) lean body m ass with
velocity in the 2000 m tim e-trial. Solid line = regression line;
dotted lines = 95% con® dence intervals.
5 m in after the perform ance test (r = 0.698, P = 0.008;
r = 0.579, P = 0.038; r = 0.584, P = 0.036, respectively).
T he three rowing economy variables all showed a
negative relationship with perform ance, although only
rowing economy at a 500 m split tim e of 2 m in 5 s
displayed statistical signi® cance (r = - 0.62, P = 0.024).
A stepwise m ultiple regression was carried out w ith
velocity in the 2000 m tim e-trial as the response
variable, and the nine variables that had shown som e
statistically signi® cant correlation with perform ance as
the explanator y variables. T he results of this analysis are
presented in Table 4. T he analysis showed that VÇ O 2max
was the best single predictor of tim e-trial velocity;
a m odel incorporating VÇ O 2 max explained 72% of the
variance in 2000 m rowing perform ance. The inclusion
of blood lactate concentration 5 m in after the perform -
ance test im proved the prediction of tim e-trial velocity;
 850 Cosgrove et al.

- 1
Table 2 Economy scores (l ´ min ) at the three submaxim al Table 3 Pearson correlation coe Y cients ( r ) for each variable
intensities (500 m split time) com pleted by all participants versus 2000 m tim e-trial performance P -values

VÇ O 2 at VÇ O 2 at VÇ O 2 at Variable r P
Participant 2 : 15 min : s 2 : 10 m in : s 2 : 05 m in : s
Height 0.207 0.497
A 2.60 2.65 3.22 Body mass 0.698 0.008
B 2.63 2.94 3.19 Percent body fat - 0.254 0.402
C 2.48 2.85 3.17 Lean body mass 0.848 < 0.001
D 2.68 2.94 3.19 VÇ O 2 ma x 0.848 < 0.001
E 2.75 3.10 3.29 Velocity at VÇ O 2m ax 0.773 0.002
F 2.56 2.90 3.19 Endurance tim e 0.791 0.001
G 2.80 3.06 3.45 Lactate maximum 5 m in after VÇ O 2 max
H 2.63 3.14 3.21 test 0.579 0.038
I 2.64 3.20 3.45 Rowing economy
J 2.74 3.23 3.57 VÇ O 2 max at 2 : 15 min : s - 0.200 0.513
K 2.66 2.99 3.23 VÇ O 2 max at 2 : 10 min : s - 0.514 0.072
Downloaded by [University Of Maryland] at 03:28 15 October 2014

L 2.73 3.05 3.34 VÇ O 2 max at 2 : 05 min : s - 0.619 0.024

M 2.68 2.88 3.28 VÇ O 2 at lactate threshold 0.39 0.190
Velocity at lactate threshold 0.39 0.190
Mean ± s 2.66 ± 0.09 2.99 ± 0.16 3.29 ± 0.13 - 1
VÇ O 2 at 4 mm ol ´ l 0.68 0.011
Velocity at 4 m mol ´ l- 1 0.734 0.004
Lactate m axim um 5 m in after the
the m odel with these two predictors explained 87% of performance test 0.584 0.036
the variance in 2000 m row ing perform ance.

Table 4 Stepwise multiple regression with 2000 m

D iscussion perform ance as the response variable

The m ain ® nding of this study is that VÇ O 2max and lean
body m ass are the best predictors of rowing per-
form ance, indicating their im portance for success in
row ing. T he strong correlation between absolute
VÇ O 2m ax and velocity in the 2000 m tim e-trial (r = 0.85,
P < 0.001) is in line w ith the results of Secher et al.
(1982), Secher (1983) and Kram er et al. (1994).
Steinacker (1993) em phasized that a high oxidative
capacity is needed to m aintain a high speed throughout
a race. Although a high VÇ O 2m ax is a good predictor of
row ing perform ance, there was a poor relationship with
relative VÇ O 2 max (m l ´ kg ´ m in - 1 ). The large body m ass of
elite rowers results in a low relative VÇ O 2max com pared
with elite endurance runners. A large m uscle m ass does
not penalize rowers, whose body weight is supported in
the boat.
N evill et al. (1992) showed that rowing perform ance
is related to both VÇ O 2max and body m ass and that
these two variables are directly related. Secher (1983)
reported that rowers w ith a large body m ass are at an
advantage, quoting an average value of 93 kg for inter-
national oarsm en. T he m ean body m ass of the present
rowers was m uch lower (73 kg), but did show a fairly
strong relationship w ith velocity in the 2000 m tim e-trial
(r = 0.70). The correlation between tim e-trial velocity
and lean body m ass, and between tim e-trial velocity at
VÇ O 2m ax , was 0.85. H owever, som e doubt exists regard-
ing the validity of calculating lean body m ass using a
Step Variable entered R
2
Beta weights
Constant 1.96
1 VÇ O 2 ma x 0.720 0.497
2 Lactate m axim um 5 m in
after perform ance test 0.870 0.0499
skinfold technique. C onsequently, one should have
som e reservations about the strength of the relationship
between lean body m ass and row ing perform ance.
There is som e error in the D urnin and Wom ersley
(1974) m ethod of estim ating body fat, indicating that
caution should be exercised w hen the relationship
between lean body m ass and row ing perform ance is
exam ined. T he equations of D urnin and Wom ersley
were based on a West of Scotland population. As
we recruited participants from the sam e region, we
considered it appropriate to use their equations. As
VÇ O 2m ax and lean body m ass were highly correlated and
had the sam e correlation coeY cient w ith tim e-trial
velocity, we decided to force VÇ O 2 max into the regression
m odel.
D uring the perform ance tests, we noted that the
rowers who were able to obtain lower 500 m split tim es
tended to produce faster perform ance tim es. O btaining
a low 500 m split tim e involves an elem ent of dynam ic
strength and the ability to produce a large force during
 Rowing performance and physiological variables of rowers
the stroke (Secher, 1983). T his could explain the high
correlation between lean body m ass and tim e-trial
velocity (r = 0.85). Individuals w ith a high lean body
m ass possess a larger m uscle m ass than individuals w ith
a low lean body m ass and, therefore, are potentially able
to produce a greater force during the stroke.
T he high correlation between lean body m ass and
velocity in the 2000 m tim e-trial shows that m uscle m ass
is an im portant variable in row ing perform ance. The
im plications of this are that less heavy individuals in the
lightweight rowing category (<72.5 kg) should increase
their lean body m ass and, where appropriate, reduce
their body fat to allow them to m axim ize their m uscle
m ass w hile rem aining within their weight category. As
no `true’ heavyweight rowers took par t in this study, the
relationship between lean body m ass and velocity in the
Downloaded by [University Of Maryland] at 03:28 15 October 2014
2000 m tim e-trial cannot be determ ined for individuals
in this category. Future research needs to investigate
this relationship in a group of heavyweight rowers.
T he VÇ O 2 at the highest subm axim al workload com -
pleted by all participants showed a signi® cant negative
relationship with tim e-trial velocity (r = - 0.62). The
VÇ O 2 at the next highest workload showed a negative
relationship with tim e-trial velocity that approached
signi® cance (r = - 0.51), but VÇ O 2 at the lowest workload
dem onstrated a very poor negative relationship w ith
tim e-trial velocity (r = - 0.20).
T he economy scores at the highest workload showed a
stronger relationship to 2000 m tim e than those at the
lower workloads; this m ight have been anticipated. The
highest workload was closest to that exp erienced during
a 2000 m race. Had econom y scores been obtained for
the rowers at true race workloads (a 500 m split tim e
of approxim ately 1 m in 40 s for this group), an even
stronger relationship w ith tim e-trial perform ance m ight
have been dem onstrated. T hat the rowers in this study
had ver y sim ilar economy values suggests that this
variable m ight not be im portant for rowing. Running
studies have shown that econom y can be a distinguish-
ing feature between athletes with sim ilar VÇ O 2m ax values.
If rowers could reduce the oxygen cost of row ing at
a given velocity, they could probably enhance their
perform ance.
Studies using heterogeneous groups in term s of
running economy have found a good relationship
between running economy and perform ance; those
using hom ogeneous groups have not (W ilcox and
Bulbulian, 1983). The economy values were relatively
hom ogeneous in the present study, and this could
explain w hy the correlation w ith 2000 m tim e-trial
perform ance was poor.
Evans et al. (1995) and H ill and Rowell (1996)
dem onstrated a positive relationship between the
velocity at VÇ O 2 max and perform ance in runners. We also
found a signi® cant positive relationship between these
851
two variables (r = 0.77). N o com parison can be m ade
w ith other groups of oarsm en, as we could ® nd no other
reports of the velocity at VÇ O 2max for oarsm en in the
literature. It m ight have been anticipated that velocity
at VÇ O 2 max would correlate well with tim e-trial perfor-
m ance, as the velocity was close to that in the 2000 m
perform ance test, whereas the blood lactate variables
dem onstrated velocities m uch lower than that at
VÇ O 2max . O ur results suggest that the velocity at VÇ O 2max
for rowers is worth investigating, as it correlates well
w ith perform ance.
T he VÇ O 2 at a blood lactate concentration of 4
m m ol ´ l- 1 in the present study showed a positive rela-
tionship with tim e-trial velocity (r = 0.68). T his is in line
w ith the results of Wolf and Roth (1987) and Steinacker
(1993), w ho found a strong relationship between these
two variables. Individuals with high VÇ O 2 scores at a
blood lactate concentration of 4 m m ol ´ l- 1 generally
produce faster tim es. As all of the par ticipants in the
present study had sim ilar rowing economy, those w ith
high VÇ O 2 at a blood lactate concentration of 4 m m ol ´ l - 1
would have reached this concentration at a higher work-
load. As our participants are capable of row ing at higher
workloads before accum ulating blood lactate in signi-
® cant quantities, it is to be exp ected that their rowing
perform ance w ill be better than that of individuals who
accum ulate blood lactate at lower workloads.
Rowing economy has received scant attention in
previous studies. M ore research is required to deter-
m ine the extent to which this variable in¯ uences
perform ance. We have noted a positive relationship
between rowing economy and perform ance, w hich
becam e signi® cant at the highest workload com pleted
by all participants. T he velocity at VÇ O 2 max of endurance
runners has previously been demonstrated to be a good
predictor of perform ance. In the present study, vel-
ocity at VÇ O 2max was a fairly strong predictor of 2000 m
tim e in rowers. However, further studies are needed to
investigate this variable further.
T he best predictors of row ing perform ance were
VÇ O 2max and lean body m ass, followed by endurance tim e
in the VÇ O 2max test, velocity in the VÇ O 2 max test and vel-
ocity at a blood lactate concentration of 4 m m ol ´ l- 1 .
T he best single predictor of velocity in the 2000 m
tim e-trial was VÇ O 2max ; the m odel incorporating VÇ O 2max
only explained about 72% of the variance in 2000 m
rowing perform ance. T he addition of blood lac-
tate concentration 5 m in after the perform ance test
im proved the prediction of tim e-trial perform ance. The
m odel incorporating these two predictors explained
87% of the variability in 2000 m rowing perform ance.
Rowers and coaches could use these ® ndings when
designing training program m es. M ore tim e should
be devoted to the im provem ent of VÇ O 2 max and lean
body m ass.
 852
References
Craven, R.P., Kinch, R.F.T., Parker, D.F. and Walter, S.J.
(1993). Stroke analysis using a m odi® ed Concept II rowing
ergom eter. Jour nal of Physiology , 459 , 133P.
Daniels, G.T. (1985). A physiologist’ s view of running
economy. M edicine and Science in Sports and E xercise , 17 ,
332± 338.
Droghetti, P. (1986). Determ ination of the anaerobic thresh-
old on a rowing ergometer by the relationship between
work output and heart rate. Scandina vian Jour nal of Sports
Sciences , 8 , 59± 62.
Durnin, J.V.G.A. and Womersley, J. (1974). Body fat assessed
from total body density and its estim ation from skin-
fold thickness: M easurem ents on 481 men and wom en
aged from 16 to 72 years. B r itish Jour nal of N utr ition , 32 ,
77± 97.
Downloaded by [University Of Maryland] at 03:28 15 October 2014
Evans, S.L., Davey, K.P., Stevenson, E.T. and Seals, D.R.
(1995). Physiological determ inants of 10 km performance
in highly trained fem ale runners of diþ erent ages. Jour na l of
Applied Physiology , 78 , 1931± 1941.
Hill, D.W. and Rowell, A.L. (1996). Running velocity at
VÇ O 2ma x. M edicine and Science in Sports and E xercise , 28 ,
114± 119.
Jensen, K., Nielsen, T.S., Fiskestrand, A., Lund, J.O.,
Christensen, M .J. and Secher, N.H. (1993). High altitude
training does not increase maximal oxygen uptake or work
capacity at sea level in rowers. Scandinavian Jour na l of
M edicine and Science in Sports , 3 , 256± 262.
Jones, D.A. and Round, J.M . (1990). Skeletal M uscle in H ealth
and Disease . M anchester: M anchester University Press.
Kindermann, W., Sim on, G. and Keul, J. (1979). The
signi® cance of the aerobic/anaerobic transition for the
determination of workload intensities during endurance
training. E uropean Jour nal of A pplied Physiology , 42 , 25± 34.
Kramer, J.F., Leger, A., Paterson, D.H. and M orrow, A.
(1994). Rowing perform ance and selected descriptive,
® eld, and laboratory variables. C anadian Jour nal of Applied
Physiolog y , 19 , 174± 184.
MacLennan, S.E., Silvestri, G.A., Ward, J. and M ahler, D.A.
(1994). Does entrained breathing improve the economy
of rowing? M edicine and Science in Sports and Exercise , 26 ,
610± 614.
Marx, U. (1988). Untersuchungen zur Trainingssteuering
im Rudern mit einem M ehrstufentest und einem Zwei-
streckentest. U npublished doctoral dissertation, Faculty of
Theoretical M edicine, University of Ulm.
Messonnier, L., Freund, H., Bourdin, M ., Belli, A. and
Cosgrove et al.
Lacour, J.R. (1997). Lactate exchange and rem oval abilities
in rowing perform ance. M edicine and Science in Sports and
Exercise , 29 , 396± 401.
Mickelson, T.C. and Hagerman, F.C. (1982). Anaerobic
threshold measurements of elite oarsmen. M edicine and
Science in Sports and E xercise , 14 , 440± 444.
Nevill, A.M ., Lakomy, H.K.A. and Lakomy, J. (1992). Rowing
ergom eter perform ance and m axim um oxygen uptake of
the 1992 Cam bridge University boat crews. Jour nal of
Sports Science s , 10 , 574± 575.
Orr, G.W., Green, H.J., Hughson, R.L. and Bennett, G.W.
(1982). Computer linear regression m odel to determine
ventilatory anaerobic threshold. Jour na l of Applied
Physiolog y , 52 , 1349± 1352.
Roth, W. (1979). Ergebnisse sportphysiologischer Studien zur
Leistungsentwicklung ausgewahlter Sportarten in den
Jahren 1964± 1978 und den Pro® l leistungsbestinmender
Merkm ale sowie der muskelzellularen Grundlagen der
spezi® schen Leistungsfahigkeit in der Sportart Rudern.
Unpublished doctoral dissertation, University of Greifs-
wald.
Roth, W., Hasart, E., Wolf, W. and Pansold, B. (1983). U nter-
suchungen zur Dynamik der Energiebereitstellung
wahrend maxim aler M ittelzeitausdauerbelastung. M ediz in-
und-S port , 23 , 107± 114.
Secher, N.H. (1983). The physiology of rowing. Jour na l of
Sports Science s , 1 , 23± 53.
Secher, N.H., Vaage, O. and Jackson, R.C . (1982). Rowing
perform ance and m aximal aerobic power of oarsmen.
Scandinavian Jour nal of Sports Scienc e , 4 , 9± 11.
Steinacker, J.M . (1993). Physiological aspects of training in
rowing. Inter nationa l Jour na l of Sports M edicin e , 14 , s1 and
s3± s10.
Weltm an, A. (1995). The B lood Lactate Response to Exercise .
Current Issues in Exercise Science M onograph No. 4.
Champaign, IL: Human Kinetics.
Wilcox, A.R. and Bulbulian, R. (1983). Running economy
and race perform ance of male and female cross-country
runners. M edicine and Science in Sports and E xercise , 15 ,
108± 113.
Wolf , W.V. and Roth, W. (1987). Validitat Spoergometrischer
Param eter f à r die Wettkamp¯ eistung im Rudern. M ediz in-
und-S port , 27 , 162± 166.
Wom ack, C .J., Davis, S.E., Wood, C.M ., Sauer, K ., Alvarez, J.,
Weltm an, A. and Gaesser, G.A. (1996). Eþ ects of training
on physiological correlates of rowing ergometry perfor-
mance. Jour na l of Strength and C onditioning R esearch , 10 ,
234± 238.