Trunsocriotu of rhc Ameriron Fisl~cricsSocirly 133: 1472-1 479, 20W
0 Copyright hy h e American Fifihcrlcs Society 2(KM
An Evaluation of the Use of Critical Swimming Speed for
Determination of Culvert Water Velocity Criteria for
Smallmouth Bass
Department of Biology. \:Jniversity of New A r ~ l s w i c k ,
Frcrlt~r~icton,
New Brunswirk E3B 6E1, Canada
,4bstruct.-Critical swimming speed (U,,,) is a common rneHsure of the relationship between
exercise intensity and duration within the prolonged perforn~nnceenvelope. This relationship is
often used to establish wntcr velocity criterin for fishways and culverts; however, the technique
involves the nssumptions that fish will choose to move at ( I ) n swimming spced cquivulent to U,,,
and (2) a ground speed that results in a passage time equnl to the intervnl used in the criticnl
speed protocol. The objective of this study wns to cvaluute thc validity ol-these assumptions nnd
the usefulness of velocity critcrin based on critical speed dnta in predicting passage success. To
nccomplish this, I measured the critical swimmin~speed of wild smalln~outhbass Miuropter~i~
dolomieu in a respiromcter and Lhe relntionship helwecn the water velocity and swimming speed
of tish in a 50-m rnccway. The lnttcr was then uscd in conjunction with the former to calculate
wntcr velocity criteria hased on ground spccds ~ctuallynttained by fish, thereby eliminating the
second nssump~ion.Criticd swimming speeds rangcd from 65 to 98 c d s and were positively
correlated to fish length (24-44 em). Swimming speeds atrained hy free-swimming fish in the
raceway increased significnntly with wnter velocity. Maximurrl nllowable wnter velocities for
rnceway fish (nd.jusred uppropriately for cxpccted swimming speed) rnngcd from 54 to 63 crnls,
depending on fish length. Dcspite this. thc proportion of individuals that mnde cornpletc ascents
ngainst water velocities ranging from 40 to 120 c d s wns high (82-95%), nnd the probnbility of
a successful nscent did not chnnge signilicnntly with water velocity. fish length, water temperature,
exposure rime, or time in cnptivity. As such, it is concluded that criticnl speed should not be used
to set culvert water velocity criteria for this species.
The vast majority of information relating to the
swimming capacity and exercise physiology of fish
has bcen generated by studying individuals forced
to perform in swim tunnel respirometers (Hummer
1995). As a result of this work, it is gerierally
accepted that the overall performance envelope
can be divided into three main swimming categories: sustained, prolonged, and burst (Beamish
1978). Sustained swimming occurs ut relativcly
low speeds, and can be maintained for long periods
(>200 minj without interruption or failure as the
energy used to suppurt this activity is generated
through aerobic metabolic pathways. The highest
velocity that can be maintained aerobically is
called maximum sustained speed. Prolonged
swimming involves moderate speeds thut require
some anaerobic cnergy and, therefore, ends in futigue after 20 s to 200 min (Bcamish 1978). Finally, high-speed locomotion is classified as burst
swimming, an activity that is exclusively anaerobic and relatively short-lived (<20 s).
Critical swimming speed (I/,,,,)is a special cat-
* Corresponding author: speake@unb.ca
Received Novembcr 9, 2003; acccprcd May 24. 2004
egory of prolonged swimming first introduced by
Brett (1964). The protocol is carried out on fish
enclosed in a swim tunnel respiromcter, during
which subjects ure forced to tnaintain position
against H particular water velocjty for a set tjnle
interval. Water velocity (and therefore swimming
speed) is [hen increnscd by n set increment until
the individual fails to swim for an entire time interval. Critical swimming speed is calculated as
the sum of the penultimate velocity attained and
a fracticln of the vclocity increment proportional
to the time spent swimming at the final velocity
relative to the full time interval (Brelt 1964). Although U,,,has often bcen used to estimate muximum sustained speed, it has also been defined ns
the highest swimming spced that a fish can maintain for a period equal in magnitude to the t h e
interval used in the test (Peake et al. 1997). T i m
intervals as low as 2 ~ n i nand as high as 60 min
have been uscd in these trials, although values between 20 and 30 rnin are most conlmon ( ~ c a m i s h
1978).
One of thc practical upplications of critical
swimming speed data is related to the establishment of water velocity criteria for fishways fin*
culverts (Jones et nl. 1974). For example, if UNII
to 1,800 s (30 min X 60 . , I
criteria for a 50-m c ~ ~ l v eWI'.
rt '
the fish were to swim Tor 30 mil L
,the speed of the individual relnti
'would be 2.8 cmls (5.000 cmll .P
were to swim at 50.0 cnds (i.c.. c.
speed), water velocity in the c ~ r
'cced 47.2 cmls (50.0 cmds - 1
animal would fatigue prior to r x i
Therefore, the muxinluni allow:,b
for thut culvert would be 47.2 u
dure, however, makes two imp( 1.
The first is that fish will c h o w 3
regardless of the water velocir.
with. and the second is that iil
ground speed that results in a F
will not excccd that used in tht ,
tocol. Clearly, these assumptiorls
however, since almost no inforiron how fast fish swim when 1 ' 1 1
containing various watcr velocir
commonly used to cslablish wal.
for lack of a better alternative.
The overall objective of this s
uate the assumptions made in cs
velocity criteria and assess thc . k
teria to uccurately predict fish pi
realistic conditions. This was act
' measuring the criticul swimmin,
mouth bass Micropterus dolo111
time intervals, ( 2 ) estublishinp
ship between watcr velocity ~ I ; L ,
for free-swimming individuals.
lationship to calculate maximur
velocity without the assurnpticl I
speed and passage time, and (
ability of this newly calculatc,l
: fish passage through a 50-m cx1.r
lup
Methods
*
E x p r r i ~ n ~ n ~animals.-Sm
al
I
collected from the Winnipeg 1 .
October 2000 and May to S C
mean\ of angling gear and tr.1
caught quickly with minimal
;transport+
in an aerated live \
( < I k m j to the Canudian Rivl
Manitoba Field Station in Pinav
ada. Individuals were housetl
that the animal could swim at 50.0 cmls for periods
s (30 lnin x 6 0 slmin). If velocity
criteria for a 50-m culvert werc required, und if
the fish wcrc to swim for 30 rtiin during thc ascent.
the speed of the individual rclative to the ground
would be 2.8 cmls (5.000 cm/I,R00 s). If the fish
were to swim at 50.0 ends (LC., criticul swimming
speed), watcr velocity in the culvert could not exceed 47.2 cmls (50.0 cmls - 2.8 cmls), or the
animal would fatigue prior to exiting the structure.
Therefore, the maximum allowable water velocity
for that culvert would be 47.2 c d s . This proccdure, however, makes two important assumptions.
The first is that fish will choose to swim at UCri,
rcgardlcss of the water velocity they are facedwith, and the second is that tish will choose a
ground speed that results in a passage time that
will not exceed that used in the critical speed protocol. Clcurly, thcse assumptions cannot he valid;
however, since almost no information is available
on how fast fish swim when fnccd with culverts
containing various watcr velocities, U,,,, data are
commonly used to establish water velocity criteria
for lack of n better alternative.
The overall objective of this study was to evaluate thc ussumptions made in establishing culvcrt
velocity criteria and assess the ability of these criteria to accurately predict fish passage under more
realistic conditions. This was accnmplishcd by (1)
mcusuring thc critical swimming speed of smallmouth bass Microprerus dolomieu using 30-min
time intervals, (2) establishing the true relatianship between water velocity and swimming speed
for frce-swimming individuals, (3) using this relationship to calculate maximum ullowable water
velocity without the nssumption related to ground
spccd and passage time, and (4) cvaluating the
ability of this newly calculuted metric to predict
fish passage ~hrougha 50-m cxperirnental raceway.
dummnni
m k mll
up to 1,800
'
'
!
'1
Methods
Experimental animals.-Smallmouth bass were
collcctcd from the Winnipeg Rivcr from April to
October 2000 und May to September 2001 by
means of angling gear and trap nets. Fish were
caught quickly with minimal air exposure and
transported in an acrited live well a short distance
(<1 km) to the Canadian Rivers Institute (CRI)
Manitoba Field Station in Pinawa, Manitoba, Can-
SlDE VIEW
\
mmp d
Iiuhl-yna
TOP VIEW
4
,
UOWNSTREAM
I ANK
KACEU'AY
UPSTREAM
TANK
FIGUREI .-Diagrammatic representations (not to
scnlc) of thc ( 8 ) rcspiromctcr find. (b) raceway used in
this study.
0.5 Lls. Smallmouth bass were always allowed at
lcust 24 h of recovery time before being uscd in
an experiment. Based on several literature accounts for this (and rclutcd) spccics (Schrccr et al.
2001; Cooke et al. 2003; Peakc and Farrcll 2004).
this period of time was considered sufficient to
allow individuals to recover from physiological
and cardiovascular disturbances ussociated with
stress, uir exposure, and exhaustive exercise. Fish
were not fed and werc, therefore, in a postabsorptive state during all experiments. Individuals were
always tcstcd and released within 7 d of cupture
to uvoid stress from prolonged captivity.
Critical swimming speed experiments.-The
swim tunncl respiromctcr used in the UCri,experiments wus fashioncd from a 1.5-m-long section
of Plexiglns pipe with an inside diameter of 20 cm
(Figure la). The upstream end of the pipc was
fastened to the side of the downstream tank of the
experimental raceway (Figure lb). Water flowed
by gravity (head diffcrcntial was approximately 6 0
cm) from the downstream tank through a butterfly
valve, past the upstream retention scrcen of the
swim chamber, through the swim chamber (npproximately 90 cm in length), past the downstream
retention screen and a propeller-driven velocity
meter (Swoffcr Model 2100, Forestry Suppliers,
Inc.), through another butterfly valve, and into the
Pinawa Channcl. The upstream valve of the rtspiromctcr controlled the volume of water that entered the device, while the downstream valve was
used to adjust water velocity. Fish were introduced
1474
PEAKE
and removed though an opening on the top, which
could be left open or sealed with u Plexiglas lid
fitted with a rubber O-ring (Figure la). During
experiments, the entire apparatus, except for the
opening through which fish wcre introduced and
removed, was covered with black plnstic to ensure
the fish did not become startled or distracted by
outside activity.
The critical swimming speed of smallmouth
bass was measured by placing fish in the respirometer, setting the water velocity at 30 c d s , and
allowing them to acclimate for at least 8 h. Following ucclimation, individuals were exposed to
stepwisc increases in velocity (V,; 15 c m h ) for
every 30 min ( t , ) of continuous swimming until.
fatigue occurred. Fatigue was asscsscd when fish
became impinged on the downstrcum screen of the
respirometer for 5 s. The highest speed maintuincd
for u complete time interval (V,), and endurance
at the final vclocity (Q),were thcn substituted into
the following equation:
Immediutcly after the critical speed dcterminations, fish wcre mesthetized in a 60 mg/L clove
oilkthanol solution (Peake 1999) and measured
for length and mass. lndividuals were thcn either
( I ) placed in a separate tank (and therefore did not
nced to be marked or tagged), rcstcd for at least
24 h, nnd uscd in ii single raceway experinlent or
(2) released immediately into the Winnipeg River.
R a r ~ w a yexperirnenls.-These cxpcrimcnts were
performed in a wood frame, plywood-covcrcd iipparatus constructcd at the CRI Manitoba Field Stution in Pinawa, Manitoba (Figure lb). Watcr was
siphoned into the structure from the upstream sidc
of a small diversion dam by gravity ( h a d dirferential was approxinlately 4 m ) through six polyvinyl chloride (PVC) pipes (diameter = 20 cm)
fitted with butterf y valves at each end. The volume
of water entering the system through each pipe
(and therefore water velocity) was controlled using
the doynstrcam valves. Water from the siphons
cntcrcd'the upstream tank of the structure (1.2 m
wide, 2.4 m long, 2 , 4 m deep) and movcd through
anothcr tank (I .Z m wide, 1.2 nl long. 1 , 2 m deep)
containing a series of bafflcs dcsigned tu minimize
turbulcnce and straighten the flow. Baffles covered
the entire cross-sectional area of this tank, and
consisted of horizontally stacked PVC pipes (20
cm long, 8 cm diameter). Water then moved into
another tnnk (1.2 rn wide, 2.4 m long, 1.2 m deep),
which served as a rest area for fish that success-
fully uscended the raceway (Figure lb). This tank
empticd into the raceway (50.0 m long, 0.6 m widc3
0.8 m deep), which was equipped with plywood
covers to provide overhead shade. The raceway
terminated at the downstream tank, which was approximately 3.6 m long, 3.6 m wide, und 1.2
deep. Wuter moving through the downstream tank
exited the systtm through a sluice gate built inlo
the back wall. The height of the sluice gate could
be adjusted via a hand-cranked winch to fine-tune
water depth and velocity.
To Facilitate entry of fish into the raccway, the
downstream tank was outfitted with guide walls
that extended from the raceway cntrance to the
sidcs of the tank at an angle of approximately 3(j0
(Figure lb). In addition, a ramp was placed in the
downstream tnnk thut extended from the floor to
the raceway entrance, creating an incline of approximately 15'. Thc floor of the raceway was approximately 30 cm higher than that of the tank so
that the system could be emptied without inadvertcntly stranding the test subjects. Water velocity
was set using a propeller-driven flowmctcr identical to thnt used in the critical speed tests. Preliminary trials indicated that smallmouth bass gcnerally swam near the bottom of the raceway. Consequently, the flowmctcr was plnced approximately 8 cm from the bottom to reflect the water speed
directly experienced by the fish.
Activity in the raceway was monitored by means
of an array of 15 custom-designcd and custambuilt light gate sensors placed on the outside wulls
of the raceway at 3.6-m intervals along its length
(Figure lb). Thc system was conceptually similar
to that described by Nelson et al. (2002). with
several modifications to make it weatherproof and
rcduce power consumption (power was supplied
by 12-V marine butteries). Technical details on the
system can be obtained by contacting the electronics shop at Simon Fraser University, Burnahy.
British Columbia, Canada. Briefly, a single light
gntc station consisted of a Plexiglas-encased circuit board equipped with a red light emitting diodc.
This emitter was plnced on thc outside wnll of the
raccway, against a Plexiglas window reccsscd inlo
the plywood wall. Focuscd red light was pulscd 31
a rate of 5 Hz and traveled 60 cm through the
raccwiiy to another Plexiglas window laced on
the opposite wall. A Plexiglas-encased detector
was mounted against this window (on the outside
wall of the raceway). Fish moving past a light gate
station would block light from rcaching the sensor.
which changed the pulse rate. When this occurre*,
a custom-designed processor/data logger would re-
M u n tihh Icnylh.
4,
Fish length rang?.
M c n n fish muss, i
Fish mnw nlnyt.. I
8
Meun tcrnpcnltur.
T~.mpcrulurcr n n p
Mcuo cuplivity ti11
Ceplivily t i m e ran:,
f
cord the identity o f the sensol
blocked and at what time ( t ( 1
The time required for a f i s h
sensor to the next was used
speed for that intcrvul, Thu.<
, yielded 14 individual ground
; averagcd and added to thc n
!, tcrmine mean swimming spec
- rtsolution of the light gate s)
'
and the swimming speed dar:
parable to thnt collected by a
era system (see Peake and Fa
During raccway experimc
WGS set at 40, 60, 80, 100, or 1
speed attainable). Tcst subjec~:
'the 2,000-L holding tank and
strcam tank. Fish were then all
throughout the system for a ra
, ranging from 0.5 to 39.5 h. A
to determine the outcome of
test was discontinued and in
data loggcr was downloadcd,
: exposure period was arbitrar?
\
I8
21
24
27
30
33
Fork lenpl) (
ICUKE
2.-The
relationship
1)s
CRITICAL SPEED FOR VELOCITY CRITERIA
TABLEI.-Details pertaining to fish Iength, muss, water temperature, and time in cnptivity fur criticnl swi~nming
(Unit), swim speed, nnd ascent succesv experiments.
cxpcnments
Mcnn fish length, (cm)
Fish length range, (cm)
Mcnn Rsh muss, (8)
Fish m m mnge. (0)
Meun tcmpcrnturc. VC)
Tcmpcrarurc m g c , ("C)
Mtan cnplivlty time, (d)
Captivity tirnc mnge. (d)
I
I
536
114-1188
19.8
17.5-21.0
3
1-7
Raceway swim
y x e d cxpenrncnt~
793
272-1498
15.6
11.0-21.0
3
1-7
Rncowny ~ u c c c s s
sxpcrimonts
-
778
272-3448
15.8
1 1.G21.0
3
1-7
cord the identity of the sensor (1-15) that had heen
blocked and at what timc (to the nearest 0.01 s).
Thc timc required for a fish to move from onc
sensor to the next was used to calculate ground
speed for that interval. Thus, a complete ascent
yielded 14 individual ground speeds, which wcre
averaged and added to the water velocity to determinc mean swimming speed. The accuracy and
resolution of the light gate system was cvaluated.
and rhc swimming s p e d data yielded was comparable to that collecrcd by a separate video camera system (see Peake and Farrell 2004).
During ruceway experiments, water vclocity
by the occurrence (or number) of attempts o r ascents. After this, water was drained from the raceway and fish were removed and released, or segregated and kept for use (after a 24-h rest period)
in one additional raceway experiment conducted
at u different water velocity. Data from all complete ascents made during thc exposure periods
were used in swimming speed determination cxperimcnts. Fish that wcre able to fully ascend thc
raceway at least once during the exposure period
were considered to have been successful.
Stutistical una1ysis.-The relationships between
( I ) critical swimming specd, fish length, and water
the 2,000-L holding tank and placed in the downstream tank. Fish were then allowcd to move freely
throughout the systcm for a random period of time
ranging from 0.5 to 39.5 h. As there was no way
to determine the outcome of the trid before the
test was discontinued and infornia~ionfrom the
data logger was downlouded, the duration of the
exposure pcriod was arbitrary and not influenced
analysis. The models ultimately selected were the
simplest forms that resulted in thc highest coefficients of determination. The relationship between
mean swimming speed, water temperature, fish
length, and time in captivity at each water velocity
tested was dctcrmined using backward, stepwisc
regression analysis. The relationship bctween the
probability of a successful raceway ascent and fish
length, water tempcrature. water velocity, time in
captivity, m d exposure duration was ussesscd using logistic regression analysis. Statistical significance was indicated by P-values of 0.05 or less.
All analyses wcre performed using NCSS statistical software (Hintze 2001).
Results
Critical Swirnrnirt~Speed Experiments
The length and mass of smallmouth bass used
in these expcriments are given in Table 1. Critical
speed increased significantly with fork length (F
= 27.5, df = 51, P < 0.001, r2 = 0.53; Figure 2)
according to the following three-parameter Hill
Fork length (an)
FIGURE2.-The relationship between criticnl swimming speed (Uc,,)
and fork length for smnllmouth bess.
I
I
relationship between water vclocity and swimming
speed for free-swimming smallnlouth bnss, and
this study is among the first to provide quantitative
data on voluntary swimming speeds attained by
fish moving through a culvcrt-like structure
against water velocities of varying intensity. Interestingly, mean ground speeds (swimming speed
minus water velocity) maintuined by smallmouth
bass during nsccnts in the raceway increased with
' water vclocity (Figure 3), which resulted in a concomitant decrease in passage time. Peakc and Farre11 (2004) also reported that smallmouth bass exhibited this behavior in a 25-m raceway. and used
postexercise oxygen consumption and physiological datn to argue thnt individuals moving at high
speeds may have been trading exercise intensity
for duration, a strategy that allowed fish to conserve anaerobic energy stores and reduce oxygen
dcbt.
The third objective of this study was to use the
relationship hetwcen swimming speed and water
velocity (equation 3 in conjunction with the critical specd model [equation 21) to establish fieldapplicable water velocity criteria for culverts. As
mentioned previously, critical swimming speed
provides an indication of the highest swimming
speed that a fish can maintain for periods up to
(and including) the magnitude of the time interval
used in the prolocol (Bunt et nl. 1999). However,
it must be stressed that these data provide no inJ formation pertaining to endurance at higher
speeds. This means thnt it is impossible to predict
whether successful pussage will occur if a fish
chooses to exceed its critical speed during an asj cent. Consequently, adherence to this artificial
, "spccd limit" is a central assumption in culvert
data. With this conpassage modcls based on U,,,,
straint in mind, field-applicable water velocity criteria can now be established for smallmouth bass.
For example, critical specds for the smallest (27
cm) and largest (47 cm) fish used in the raccway
experiments are 78 and 91 c d s , respectively
(equation 2). However, in a raceway (or culvcrt)
thesc fish would only attain thesc swimming
'
speeds whcn faced with water velocities of 54 and
63 cm/s, respectively (Figure 3). Thercfore, water
Speeds between 54 and 63 cmls should represent
reasonable velocity criteria for free-swimming
The final objective of this study was to evaluate
e ability of culvert watcr velocity criteria based
on Ucr,, data, corrected appropriately for ground
peed, to predict the capacity of smallmouth bass
to move through an actuul 50-rn raceway. Results
of this study clearly indicate that these criteria
(54-63 c d s ) did not accurately predict threshold
water velocities for this species under the conditions tested (Figure 4). Indeed, smallmouth bass,
regardless of size and woter temperature, were able
to success full^ ascend the raceway against watcr
currents moving at more than twice those of the
predicted thresholds. Whilc it could be argued that
U,,,,
tests involving intervals closer to the passage
times dernonstratcd by fish in the present study
( I ,4-4.2 min) might yield highcr critical specds
and more appropriate velocity crirerin, Cooke and
Bunt (2001) used 5-min time intervals and reported that the mean critical speed of smallmouth
bass tested at 17°C was 11 1 c d s . This swimming
speed would have bcen attained in the raccway in
response to a wuter velocity of about 78 c d s (Figure 3). This value, although better, still docs not
predict a reasonable maximum allowable watcr velocity blised on the rcsults of this study (Figure
4). If this finding is general and not specific to
smallmouth bnss, it may not be possible to adjust
critical speed data, or the protocol, to address the
issue. Clearly, additional work on other species
and efforts to develop and evaluate new swimming
performancc protocols are warranted.
This is not the first study to show thnt forced
swimming performance datn can underestimate
voluntary locomotory capacity and behavior of
fish in the field. Mallen-Cooper (1992) used a
small experimental fishway to study the passage
performancc of juvenile Australian bass Macquariu novernaculcara and barramundi (ulso
known as b~irramundiperch) Lutes calcurifer. und
found that fish were able to swim faster in the
fishway than would have been predicted by flume
studies. Peake (2004) also found that watcr speeds
that resulted in swimming failure in respirometerconfined juvenile northern pike Esox lucius were
lower than those that resultcd in impingement on
an experimental irrigation intake screen.
The simplest explanation for the lack of agreement bctween Ucr,,-based velocity criteria and the
raceway data in this study is that critical swimming
speed does not accurattly describe the relationship
between swimming specd and endurance for smallmouth bass. Although this finding, if general in
nature, would invalidate many assumptions currently made by researchers studying fish locomotion, it is worth noting that it has never bcen
proven that the relationship hetwcen swimming
specd and timc.to fatigue nctually holds for fish
outside of a swim tunnel respirometer. Elucidation
1478
PEAKE
of why this relationship does not appear to hold
for smallmouth bass awaits further investigation;
however, it is likely that causation is associatcd
with ( I ) the behavioral refusal of confined fish to
swim to complete physiological exhaustion, as has
been dernonstratcd by various authors,for several
species (Tarby 1981; Reidy et al. 1995; Swanson
c t al. 1998), andlor (2) differenccs in the energetic
costs associated with confined and unconfined
swimming. Fish exercising in rcspirometers arc
gcnerally cxposed to straightened flows describcd
as "uniformly microturbulent" (Plaut and Gordon
1994), conditions that probably d o not occur oftcn
in nature or in fishwgys and corrugated or baffled
culvcrts. Although turbulence has recently been
shown to increasc the energetic costs of Iocomotion in fish (Endcrs et al. 2003), such a difference
would tend to lead to an overestimation of freeswimming capacity, rather than the undercstirnation observcd here. Fish in turbulent water huvc
also been shown to reduce their cncrgetic expcnditure in turbulent water, both in the laboratory
(Liao et al. 2003) and in tbc field (Hinch and Rand
2000; Standcn et al. 2002). Howcver, apart from
normal open channcl houndary layer issuc.;, the
racewny used in this study wns relatively free of
turbulence, and s~nallmouthbass videotaped by
Penkc and Farrell (2004) in the satne raceway
swam actively (albeit near the bottom) and did not
appear to bc gaining any hydrodynamic advantugc
from the flow. Similar advantages can also be associated with vertical and horizontal walls, whcrc
boundary layer water moves slowcr and ground
effect increases lift and reduces drag (Webb 1993).
Howcver, wall effects were prohably not an important determinant of thc pcrformancc differenccs
observed in this study as (1) the cffect becorncs
lcss pronounced at high water velocities (Wcbb
1993), (2) both confined and unconfined individuals swam near thc bottom of their respective
flumes, and (3) similar fish filmed in the raceway
by Pcakc und Farrell (2004) did not favor the walls
over the middle of the channel. It may be that
diffcrcnces in wall spacing bctween the rcspirometcr and the raceway was a,contributing factor, as
Webb (1993) demonstruted that the critical speed
of fish that swam in a narrow area bctween two
walls was lower than that of individuuls that had
more space. The author attributed the difference
to constraints on tail bcat amplitude in the former.
It is possiblc that confined smallmouth bass in the
present study were not able to swim as fast as those
in the raceway becausc the walls of the respirometer limited tail beat amplitude. Howcver, the re-
1
duction in U,,,,
(about 20%) reported by wFbb
(1993) was not sufficicnt to fully explain the dif.
-ferenccs shown here.
In conclusion, it is clear that culvert water ve.
locity critcria based on critical swimming speed
duta for rcspirometer-confined fish arc very conbass, and there is little
scrvative for
to indicate that this is a species-specific phenomenon. While the problem is likely caused by a com.
bination of the behavioral, physiological, and hy.
draulic issues discussed above, it is clear that new
techniques need to be developed to establish appropriate culvert und fishway water velocity criteria.
mallm mouth
Acknowledgments
This project was made initially mude possible
through a grant to S.J.P. from the Manitoba Hydro
Reseurch and Development Committee, with help
from Roy Bukowsky and Dennis Windsor. I thank
Phillip Wang, Mamo Klein, Cory Ruhlen, Dennis
Smith, and Jeff Long for assistunce in designing
and consrruoting thc experimentul raceway. Initial*
funding was matched by an NSERC collaborntive
rcsearch and development grant to R . A. McKcown, with sdditionnl funding and in-kind contributions coming from Manitoba Conservation,
Manitoba Department of Highways (Bridges and
Structures), the Deep River Science Academy, and
the Town of Pinawa. The following scholarships
were provided (to S.J.P.): an NSERC postgraduate
scholarship, two Simon Fraser University graduate
fellowships, an Abhotmrettwcll scholarship, and
an Anne Vallee scholarship. The contributions of
research assistants Cathy Sumi. Jaymie MacDonald, Cristy Smith, und Jeff Long are gratefully acknowledged.
References
Beamish, F. W. H. 1978. Swimming capncity. P u p
101-187 i n W. S. Hoar nnd D. J. Rnndall. rdilorsFish physiology, volume 7. Academic Press, New
Yurk.
Urctt, I. R. 1964. The respiriltury metabolijm nnd swimming perrrirnlnnce of young sockeye salmon. Journnl of the Fisherics Kcsearch B o d of C~nadn21:
1183-1226.
Bunt. C. M..C. K~topodis,and R. S. McKinley. 1999.
Alltaction nnd passage efficiency of white suckers
and smallmouth bass by two Dtnil fishwnys. Noflh
American Journal of Fisheries Management 19:
793-803.
Cooke. S . J.. and C. M. Bunt. 2001. Assesslncnl of
internal and externnl antenna configurations of t d o
transmitters irnpltlnted i n smallmouth bass. Nonh
,
,
ookc, S. J.. K . G. Ostr;~ntl.( ' '.I
D.H. Wnhl, rmd D.P, Ph~lii,:..
responses of lurpemouth bu, -. I '
and brief air exposure over a
peratures. Transactions of 111 :
'Enders. E. C . , D. Boihcluir, ant1
'
effect of turbulence on the c(
juvenile Atli~nLicsnlmon i . S 1
Journnl of Fisherics and Aqu:,
. \
Farlinger, S., and E W. H. BC,II:
time and velocity incrcmcn~~
ming speed of largemouth I :
moides). Transaclions of the :'>.I
ciety 106~436--439.
Hnrnmcr, C. H. 1995. Fntigur : I > I I
fish. Comparative Biochenli!
112A:l-20.
Hinch, S. G., and P. S.Rnnd. 2000
and forward-assikted propul<:r
behaviours o l uprivcr rnigrniil
Journal of Fishcries nnd Aquat
2478.
Hinrze, J. 1001. NCSS and P A \
stntisticnl systems. Knysville.
Jones. D. R.. J. W. Kiceniuk, and (
Evnluatiun of thc swimming PI
fish species from the M~cken
the Fisheries Research BoarJ
1647.
Liao, J. C., D. N. Beal, G. V. L a u
tafyllou. 2003. The Karma11
nematic9 of rainbow trout : V
street. Journal of Experimenkr
1073.
Mallen-Cooper. M. 1992. swim^:
Australian bass, Mucqrrarin
dnchner), and juvenile harm
(Bloch) in an experimenl:~
Journal of Marine nnd Fresh.
834.
Nelson, J. A,, F! S. Cotwalt, S.
Webber. 2002. Beyond U,,,
performnnce lests to the p h
the nnin~al,including a ncw
pnrntivc Biochemistry nnd I'
289-302.
Peake, S. J. 2004. Swimming cnp
American Journal of Fisheries Mraagament 2 1 :
236-24 1.
Cooke, S. J.. K. G. Ostmnd. C. M. Bunt, J. E Schreer.
D. H. Wnhl, nnd D. P. Philipp. 2003. Cnrdiovnsculnr
responses of largemouth hnss to exhaustive exercise
nnd brief nir cxposurc over n rnnge of wntcr tcmperntures. Transnctions of the Amcricnn Fishcrics
Society 132: 1 154-1 165.
Enderg, E. C., D. Boisclnir, and A. G. RopW03. The
effect of turbulence on the cost of swimming for
juvenile Atlantic salmon (Sulmo snlar). Cmadiap
Journal of Fishcries and Aquntic Sciences 60: 11491 160.
Furlinger, S.. and F. W. H. Beamish. 1977. Effects of
time and velocily increments on the criticnl swirnming speed of largemouth bnss (Mlcroprerus salmoides). Transactions of the American Fisheries Society 106:436-439.
Hammer, C. H. 1995. Fatigue and exercise tests with
fish. Comparntive Biochemistry and Physiology
1 12A: 1-20.
Hinch, S. 0..and P. S. Rand. 2000. Optimal swim speeds
nnd forward-tlssisted propulsion: energy conserving
behaviours of upriver migrating salmon. Canndinn
Journal of Fisheries and Aquatic Sciences 57:24702478.
Hintze, J. 2001. NCSS nnd PASS. Number cruncher
statistical systems, Kaysville, Utah.
Jones. D. R., J. W. Kiceniuk, and 0 . S. Ramford. 1974.
Evaluation of the swimming performance of severnl
fish species from the Mnckenzie River. Journnl of
the Fisheries Research Board of Canada 3 1 : 164 11647.
Liau. J. C.. D. N. Beal. G. V Lauder, and M. S. Trinntafyllou. 2003. The Karmun gait: novel hody kinematics of rainbow trout swimming in n vortex
street. Journnl of Experimcntnl Biology 206:10591073.
, Mallen-Cooper, M. 1992. Swimming nbility of juvcnile
Australinn bass, Macquaria nov~maculeara(Stein1
dachncr), and juvenile barramundi. Lates ca1carifPr
(Bloch) in an experimental fishwny. Austrnlinn
i, Journal
of Mtlrine and Freshwater Research 43:823!
834.
, Nelson, J. A,. P. S. Gotwalt, S. P. Reidy, and D. M.
b
Webber. 2002. Beyond U,,,: matching swimming
performance tests to the physiological ecology of
the animal, including a new fish "drag mip." Compnriltive Biochemistry nnd Physiology Pnrt A 133:
289 -302.
Peake, S. J. 2004. Swimming capacity of juvenile north-
1
ern pike in relation to impingcrnent on irrigation
intukc screens. North Americun Journal of Fisheries
Mnnagemcnt 24:390-396.
Penke, S. J . 1999. Sodlum blcarbonute and clove oil as
potentinl nnesthetics for non-salmonid fishes. North
Americnn Journnl of Fisheries Management 18:
9 19-924.
Penke, S . J., nnd A. P. Farrell. 2004. Locomotory behaviour and bostexercise physiology in relation to
swimming speed, gait transition, and metabolism in
frce-swimming smallmouth hnss Microptfrus dolornleu. Journal of Experirnentnl Biology 207: 15631575.
Peake, S. J., R. S. McKinley, and D. A. Scmton. 1997.
Swimming performance o f various freshwarer Newfoundland salmonids rclativc to hnhitnt selection
and fishwtly design. Journnl of Fish Biology 51:
710-723.
Plaut. I.. and M. S. Gordon. 1904. Swimming metaholism of wild-type and cloned zchrnfish B I - u L I L . ~ ~ .
danio rerio. Journnl of Experimentnl Biology 194:
209-223.
Reidy, S. P.. J. A. Nelson. Y. Tang, and S. R. Kcrr. 1995.
Postexercise metabolic rate in Atlantic cod and its
dependence upon the method of exhaustion. Journnl
of Fish Biology 471377-386.
Schrccr. J. E. S. J. Cookc, nnd R. S. McKinley. 2001.
Cardinc response to variable forced exercisc at difIerent temperatures: an angling simulation for
smallmouth bass. Transactions of the American
Fisheries Society 130:783-795.
S m d e n . E. M., S. G. Hinch. M. C. Healey, and A. P.
Farrell. 2002. Energetic costs of migration through
the Frnser River Canyon. British Columbia, in itdult
pink (Oncorhynchus gorhuscha) and sockeye (Oncorhynchus nerku) snlmon ns nssessed by EMG teIcmetry.. Cnnndinn Journal of Fishcries nnd Aquntic
Sciences 59: 1809-1818.
Swnnson, C.. P. S. Young, nnd J. J. Cech. 199R. Swimming performance of delta smelt: maximum pcrformance and behavioral and kinematic limitations
on swimming at submaximal velocities. Journal of
Experimental Biology 2 0 1 3 3 - 3 4 5 .
Tarby, M. J. 1981. Metabolic expenditure of walleye
(Stizostedion vitrpum) as determined by rate of oxygen consumption. Canadian Journal of Zoology
59:882-HH9.
Webh, P. W. 1993. The effect of solid and porous channel
wnlls on steady swimming of stcelhend trout, 011corhynchus mykiss. Journtll of Experirnenttll Biology 17B:Y7-108.