Swimming performance of two Iberian cyprinids: the Tagus nase Pseudochondrostoma polylepis (Steindachner, 1864) and the bordallo Squalius carolitertii (Doadrio, 1988)

The main purpose of this study was to gather swimming performance information for two endemic cyprinids of the Iberian Peninsula to contribute to the optimization of fish ways. Critical swimming speed (U_crit) was determined for the Tagus nase Pseudochondrostoma polylepis (Steindachner, 1864) and for the bordallo Squalius carolitertii (Doadrio, 1988) in a swimming tunnel. From a total of 80 P. polylepis tested, the mean (± SD) U_crit observed was 0.78 ± 0.15 ms^?1 (c. 3.74 ± 0.93 BL s^?1); the 68 S. carolitertii tested presented an U_crit of 0.54 ± 0.1 ms^?1 (c. 4.43 ± 0.74 BL s^?1). Significant interspecific differences were found between the U_crit of the tested cyprinids. Intraspecific comparisons between the U_crit and the variables of size, sex, condition factor and gonado‐somatic index were also made. No sex‐or gonad maturation‐related differences between the U_crit were identified, but the robust P. polylepis were found to be stronger swimmers. Water velocities in fish ways for P. polylepis and S. carolitertii should aim, on average, for lower than 0.7 and 0.5 ms^?1, respectively.     

Sexually dimorphic morphology and swimming performance relationships in wild‐type zebrafish Danio rerio

This study compared prolonged swimming performance (U_crit) between male and female Danio rerio, and characterized how body shape was associated with this performance measure in each sex. When swimming in small (n = 6) mixed‐sex groups at 28° C, males swam, on average, over 10 cm s^?1 faster than females despite being significantly smaller. Body shape was sexually dimorphic, with males and females exhibiting small, but statistically significant differences in most aspects of body shape. Body shape explained 18 and 43% of the variation in U_crit among males and females. In general, effects of body shape on swimming performance appeared to be sex limited, whereby different aspects of body shape affected performance in each sex, although the contribution of the distance between pelvic and anal fins to swimming performance was weakly sexually antagonistic.     

The critical swimming speed of Iberian barbel Barbus bocagei in relation to size and sex

The swimming capacity of Barbus bocagei was measured with the critical swimming speed (U_crit) standard test in a modified Bla?ka‐type swim tunnel. Sixty B. bocagei were tested and they exhibited a mean ±s .d . U_crit of 0·81 ± 0·11 m s^?1 or 3·1 ± 0·86 total lengths per second (L_T s^?1). Sex had no effect on U_crit but significant differences were found between the swimming performance of fish with distinct sizes.     

Critical swimming speeds of wild‐caught sand‐smelt Atherina presbyter larvae

Swimming abilities of wild‐caught sand‐smelt Atherina presbyter larvae were assessed as critical swimming speed (U_crit) throughout ontogeny. The mean U_crit increased with size, ranging from 3·6 to 18·7 cm s^?1, over the size range of 6·6–21·0 mm L_T. This indicates that at hatching A. presbyter larvae, far from being passive floaters, are already capable of active behaviours, which may influence their dispersal patterns.     

Effects of age and size on critical swimming speed of juvenile Chinese sturgeon Acipenser sinensis at seasonal temperatures

Changes in the critical swimming speed (U_crit, cm s^?1) with ontogeny of 2·5–12·5 month‐old juvenile anadromous Chinese sturgeon Acipenser sinesis were measured in a modified Blazka‐type swimming tunnel. The absolute U_crit increased with length, mass and age; the relative U^′_crit (body lengths, s^?1), however, decreased. Juvenile A. sinesis did not display a parr–smolt transformation at the length or age threshold to tolerate full‐strength seawater.     

Laboratory protocol to calibrate sea lamprey (<Emphasis Type="Italic">Petromyzon marinus</Emphasis> L.) EMG signal output with swimming

A correct application of electromyogram (EMG) telemetry in the field can be a powerful tool to evaluate activity patterns and swimming strategies of fishes. We evaluated the swim performance of seven untagged sea lampreys (Petromyzon marinus L.) with critical swim speed (U _crit) tests. The average U _crit observed was c. 1.03 ms⁻¹ (i.e., 1.14 BL s⁻¹). The strongest reotaxic response was observed during tests using water velocities between 0.4 ms⁻¹ and 0.8 ms⁻¹. During two consecutive years (i.e., 2004 and 2005), in order to model upstream migration of sea lampreys with CEMG transmitters (Lotek Wireless), we calibrated EMG signal with swim speed. A high correlation between EMG records and swim speed was observed in both years (r ² = 0.74–0.93). However, in spite of methodology improvements and standardization in the second year of study, differences in intercepts and slopes were observed between individuals, making the determination of a unique calibration equation for all tagged animals unfeasible. Therefore, it appears to be necessary to obtain the relationship between EMG signals and swimming speed for each lamprey using laboratory procedures, prior to release in the wild. It is unknown whether this variability results from individual locomotor behaviour, physiological state and/or variation in placement and functioning of the EMG transmitters. The results of five laboratory calibrated lampreys, released in the River Mondego, revealed considerable differences between swim speeds calculated with EMG signal (calibration equation) and ground speed therefore it was not possible to successfully calibrate the EMG signal output with swimming speed. In order to accomplish this, longer continuous swimming tests in laboratory are necessary. Nevertheless, the calibrated swimming effort gives reliable information about the swimming behaviour and permits comparison of the results between animals.     

Thermal tolerance and metabolic physiology among redband trout populations in south‐eastern Oregon

Streamside measurements of critical thermal maxima (T_crit), swimming performance (U_crit), and routine (R_r) and maximum (R_max) metabolic rates were performed on three populations of genetically distinct redband trout Oncorhynchus mykiss in the high‐desert region of south‐eastern Oregon. The T_crit values (29·4 ± 0·1° C) for small (40–140 g) redband trout from the three streams, and large (400–1400 g) redband trout at Bridge Creek were not different, and were comparable to published values for other salmonids. At high water temperatures (24–28° C), large fish incurred higher metabolic costs and were more thermally sensitive than small fish. U_crit(3·6 ± 0·1 L_F s^?1), R_r(200 ± 13 mg O₂ kg^?0·830 h^?1) and metabolic power (533 ± 22 mg O₂ kg^?0·882 h^?1) were not significantly different between populations of small redband trout at 24° C. R_max and metabolic power, however, were higher than previous measurements for rainbow trout at these temperatures. Fish from Bridge Creek had a 30% lower minimum total cost of transport (C_min), exhibited a lower refusal rate, and had smaller hearts than fish at 12‐mile or Rock Creeks. In contrast, no differences in U_crit or metabolism were observed between the two size classes of redband trout, although C_min was significantly lower for large fish at all swimming speeds. Biochemical analyses revealed that fish from 12‐mile Creek, which had the highest refusal rate (36%), were moderately hyperkalemic and had substantially lower circulating levels of free fatty acids, triglycerides and albumin. Aerobic and anaerobic enzyme activities in axial white muscle, however, were not different between populations, and morphological features were similar. Results of this study: 1) suggest that the physiological mechanisms that determine T_crit in salmonids are highly conserved; 2) show that adult (large) redband trout are more susceptible to the negative affects of elevated temperatures than small redband trout; 3) demonstrate that swimming efficiency can vary considerably between redband trout populations; 4) suggest that metabolic energy stores correlate positively with swimming behaviour of redband trout at high water temperatures; 5) question the use of T_crit for assessing physiological function and defining thermal habitat requirements of stream‐dwelling salmonids like the redband trout.     

Effects of surgically implanted acoustic transmitters >2% of body mass on the swimming performance, survival and growth of juvenile sockeye and Chinook salmon

The influence of surgical implantation of an acoustic transmitter on the swimming performance, growth and survival of juvenile sockeye salmon Oncorhynchus nerka and Chinook salmon Oncorhynchus tshawytscha was examined. The transmitter had a mass of 0·7 g in air while sockeye salmon had a mass of 7·0–16·0 g and Chinook salmon had a mass of 6·7–23·1 g (a transmitter burden of 4·5–10·3% for sockeye salmon and 3·1–10·7% for Chinook salmon). Mean critical swimming speeds (U_crit) for Chinook salmon ranged from 47·5 to 51·2 cm s^?1 [4·34–4·69 body lengths (fork length, L_F) s^?1] and did not differ among tagged, untagged and sham‐tagged groups. Tagged sockeye salmon, however, did have lower U_crit than control or sham fish. The mean U_crit for tagged sockeye salmon was 46·1 cm s^?1 (4·1 L_F s^?1), which was c. 5% less than the mean U_crit for control and sham fish (both groups were 48·6 cm s^?1 or 4·3 L_F s^?1). A laboratory evaluation determined that there was no difference in L_F or mass among treatments (control, sham or tag) either at the start or at the end of the test period, suggesting that implantation did not negatively influence the growth of either species. None of the sockeye salmon held under laboratory conditions died from the influence of surgical implantation of transmitters. In contrast, this study found that the 21 day survival differed between tagged and control groups of Chinook salmon, although this result may have been confounded by the poor health of Chinook salmon treatment groups.     

Swimming performance of 12 Schizothoracinae species from five rivers

A series of stepped velocity tests were carried out in a Brett‐type swimming respirometer and the overall range in swimming performance for 12 Schizothoracinae species was measured. The relative critical swimming speed U_crit and burst speed U_burst decreased with body length, while absolute U_crit and U_burst increased with body length. U_crit increased with temperature up to approximately 15^° C and then decreased. Species with a high U_crit also displayed a higher U_burst.     

Morphological predictors of swimming speed performance in river and reservoir populations of Australian smelt Retropinna semoni

Dam construction is a major driver of ecological change in freshwater ecosystems. Fish populations have been shown to diverge in response to different flow velocity habitats, yet adaptations of fish populations to river and reservoir habitats created by dams remains poorly understood. We aimed to evaluate divergence of morphological traits and prolonged swimming speed performance between lotic and lentic populations of Australian smelt Retropinna semoni and quantify the relationship between prolonged swimming speed performance and morphology. Prolonged swimming speed performance was assessed for 15 individuals from each of three river and two reservoir populations of R. semoni using the critical swimming speed test (U_crit). Body shape was characterized using geometric morphometrics, which was combined with fin aspect ratios and standard length to assess morphological divergence among the five populations. Best subsets model-selection was used to identify the morphological traits that best explain U_crit variation among individuals. Our results indicate R. semoni from river populations had significantly higher prolonged swimming speed performance (U_crit = 46.61 ± 0.98 cm s⁻¹) than reservoir conspecifics (U_crit = 35.57 ± 0.83 cm s⁻¹; F_1,74 = 58.624, Z = 35.938, P < 0.001). Similarly, R. semoni sampled from river populations had significantly higher fin aspect ratios (AR_caudal = 1.71 ± 0.04 and 1.29 ± 0.02 respectively; F_(1,74) = 56.247, Z = 40.107, P < 0.001; AR_pectoral = 1.85 ± 0.03 and 1.33 ± 0.02 respectively; F_(1,74) = 7.156, Z = 4.055, P < 0.01). Best-subset analyses revealed U_crit was most strongly correlated with pectoral and caudal fin aspect ratios (R²_adj = 0.973, AIC_c = 41.54). Body shape, however, was subject to a three-way interaction among population, habitat and sex effects (F_3,74 = 1.038. Z = 1.982; P < 0.05). Thus sexual dimorphism formed a significant component of unique and complex variation in body shape among populations from different habitat types. This study revealed profound effects of human-altered flow environments on locomotor morphology and its functional link to changes in swimming performance of a common freshwater fish. While past studies have indicated body shape may be an important axis for divergence between lotic and lentic populations of several freshwater fishes, fin aspect ratios were the most important predictor of swimming speed in our study. Differences in body morphology here were inconsistent between river and reservoir populations, suggesting this aspect of phenotype may be more strongly influenced by other factors such as predation and sexual dimorphism.     

The effects of temperature on swimming performance of juvenile shortnose sturgeon (Acipenser brevirostrum)

The swimming performance of juvenile shortnose sturgeon (～16 cm TL, ～20 g), Acipenser brevirostrum, was quantified with regards to temperature (5 to 25°C) using both increased (U_crit) and fixed velocity (endurance) tests in a laboratory setting. Sturgeons were found to show reduced U_crit values at 5 and 10°C (25.99 and 28.86 cm s^?1 respectively), with performance beginning to plateau at 15°C through 25°C (33.99 cm s^?1). For the endurance protocol, fish were tested at speeds of 35, 40 and 45 cm s^?1 at 5, 15 and 25°C. Performance within a single speed was similar at all temperatures, indicating the usage of anaerobic metabolism to fuel locomotion at these higher velocities. Overall, shortnose sturgeon demonstrated high tolerance towards a wide range of temperatures but showed few differences between performance levels at colder or warmer water conditions.     

Measuring Ucrit and endurance: equipment choice influences estimates of fish swimming performance

This study compared the critical swimming speed (U_crit) and endurance performance of three Australian freshwater fish species in different swim‐test apparatus. Estimates of U_crit measured in a large recirculating flume were greater for all species compared with estimates from a smaller model of the same recirculating flume. Large differences were also observed for estimates of endurance swimming performance between these recirculating flumes and a free‐surface swim tunnel. Differences in estimates of performance may be attributable to variation in flow conditions within different types of swim chambers. Variation in estimates of swimming performance between different types of flumes complicates the application of laboratory‐based measures to the design of fish passage infrastructure.     

Relationships between metabolic rate, muscle electromyograms and swim performance of adult chinook salmon

Oxygen consumption rates of adult spring chinook salmon Oncorhynchus tshawytscha increased with swim speed and, depending on temperature and fish mass, ranged from 609 mg O₂ h^?1 at 30 cm s^?1 (c. 0·5 BL s^?1) to 3347 mg O₂ h^?1 at 170 cm s^?1 (c. 2·3 BL s^?1). Corrected for fish mass, these values ranged from 122 to 670 mg O₂ kg^?1 h^?1, and were similar to other Oncorhynchus species. At all temperatures (8, 12·5 and 17° C), maximum oxygen consumption values levelled off and slightly declined with increasing swim speed >170 cm s^?1, and a third‐order polynomial regression model fitted the data best. The upper critical swim speed (U_crit) of fish tested at two laboratories averaged 155 cm s^?1 (2·1 BL s^?1), but U_crit of fish tested at the Pacific Northwest National Laboratory were significantly higher (mean 165 cm s^?1) than those from fish tested at the Columbia River Research Laboratory (mean 140 cm s^?1). Swim trials using fish that had electromyogram (EMG) transmitters implanted in them suggested that at a swim speed of c. 135 cm s^?1, red muscle EMG pulse rates slowed and white muscle EMG pulse rates increased. Although there was significant variation between individual fish, this swim speed was c. 80% of the U_crit for the fish used in the EMG trials (mean U_crit 168·2 cm s^?1). Bioenergetic modelling of the upstream migration of adult chinook salmon should consider incorporating an anaerobic fraction of the energy budget when swim speeds are ≥80% of the U_crit.     

Pop up satellite tags impair swimming performance and energetics of the European eel (Anguilla anguilla)

Pop-up satellite archival tags (PSATs) have recently been applied in attempts to follow the oceanic spawning migration of the European eel. PSATs are quite large, and in all likelihood their hydraulic drag constitutes an additional cost during swimming, which remains to be quantified, as does the potential implication for successful migration. Silver eels (L_T = 598.6±29 mm SD, N = 9) were subjected to swimming trials in a Steffensen-type swim tunnel at increasing speeds of 0.3–0.9 body lengths s⁻¹, first without and subsequently with, a scaled down PSAT dummy attached. The tag significantly increased oxygen consumption (MO₂) during swimming and elevated minimum cost of transport (COT_min) by 26%. Standard (SMR) and active metabolic rate (AMR) as well as metabolic scope remained unaffected, suggesting that the observed effects were caused by increased drag. Optimal swimming speed (U _opt) was unchanged, whereas critical swimming speed (U _crit) decreased significantly. Swimming with a PSAT altered swimming kinematics as verified by significant changes to tail beat frequency (f), body wave speed (v) and Strouhal number (St). The results demonstrate that energy expenditure, swimming performance and efficiency all are significantly affected in migrating eels with external tags.     

Thermal sensitivity of metabolic rates and swimming performance in two latitudinally separated populations of cod, <Emphasis Type="Italic">Gadus morhua</Emphasis> L.

Atlantic cod populations live in a wide thermal range and can differ genetically and physiologically. Thermal sensitivity of metabolic capacity and swimming performance may vary along a latitudinal gradient, to facilitate performance in distinct thermal environments. To evaluate this hypothesis, we compared the thermal sensitivity of performance in two cod stocks from the Northwest Atlantic that differ in their thermal experience: Gulf of St Lawrence (GSL) and Bay of Fundy (BF). We first compared the metabolic, physiological and swimming performance after short-term thermal change to that at the acclimation temperature (7°C) for one stock (GSL), before comparing the performance of the two stocks after short-term thermal change. For cod from GSL, standard metabolism (SMR) increased with temperature, while active metabolism (AMR, measured in the critical swimming tests), EMR (metabolic rate after an exhaustive chase protocol), aerobic scope (AS) and critical swimming speeds (U _crit and U _b–c) were lower at 3°C than 7 or 11°C. In contrast, anaerobic swimming (sprint and burst-coasts in U _crit test) was lower at 11 than 7 or 3°C. Factorial AS (AMR SMR⁻¹) decreased as temperature rose. Time to exhaustion (chase protocol) was not influenced by temperature. The two stocks differed little in the thermal sensitivities of metabolism or swimming. GSL cod had a higher SMR than BF cod despite similar AMR and AS. This led factorial AS to be significantly higher for the southern stock. Despite these metabolic differences, cod from the two stocks did not differ in their U _crit speeds. BF cod were better sprinters at both temperatures. Cod from GSL had a lower aerobic cost of swimming at intermediate speeds than those from BF, particularly at low temperature. Only the activity of cytochrome C oxidase (CCO) in white muscle differed between stocks. No enzymatic correlates were found for swimming capacities, but oxygen consumption was best correlated with CCO activity in the ventricle for both stocks. Overall, the stocks differed in their cost of maintenance, cost of transport and sprint capacity, while maintaining comparable thermal sensitivities.     

Effect of repeated exercise fatigue on the swimming performance of juvenile rock carp (Procypris rabaudi,Tchang)

The swimming performance of juvenile rock carp (Procypris rabaudi, Tchang) subjected to repeated fatigue exercise was studied using a flume-type respirometer at 20°C. The critical swimming speed (U_crit) and oxygen consumption rate (MO₂) of juvenile rock carp were measured during two successive stepped velocity tests, following a 60 min rest interval. U_crit of rock carp was giving a recovery ratio (R_r) of 92.64%, and exertion exercise decreases U_crit. When MO₂ was plotted as a linear function of U, the slope for trial 1 was 1.06 and 1.50 for trial 2, indicating a decreasing in swimming efficiency. The maximum metabolic rate (MMR) increased from 17.06 ± 1.14 mmol O₂/(kg·hr) to 19.14 ± 1.23 mmol O₂/(kg·hr), and the exercise post oxygen consumption rate (EPOC) increased from 9.00 to 9.65 mmol O₂/kg. Repeated fatiguing exercise increased both the aerobic and anaerobic cost of reaching U_crit, but anaerobic metabolism accounted for a larger proportion in the trial 2. The data investigation on the swimming performance and the physiological response to fatigue provide important design criteria for fishways.     

Swimming Performances of Four California Stream Fishes: Temperature Effects

The critical swimming velocity (U_crit) of four California stream fishes, hardhead, Mylopharodon conocephalus, hitch, Lavinia exilicauda, Sacramento pikeminnow, Ptychocheilus grandis, and Sacramento sucker, Catostomus occidentalis was measured at 10, 15, and 20°C. Hardhead, Sacramento sucker, and Sacramento pikeminnow swimming performances tended to be lowest at 10°C, higher at 15°C, and then decreased or remained constant at 20°C. Hitch swimming performance was lower at 10°C than at 20°C. There were no significant differences among species at 10 or 15°C, although pikeminnow and hitch were ca. 20% slower than hardhead or sucker. At 20°C hardhead, Sacramento sucker, and Sacramento pikeminnow had remarkably similar U_crit but hitch were significantly (by 11%) faster. We recommend that water diversion approach velocities should not exceed 0.3ms^–1 for hitch (20–30cm total length) and 0.4ms^–1 for hardhead, Sacramento pikeminnow, and Sacramento sucker (20–30cm TL).     

Thermal acclimation to 4 or 10°C imparts minimal benefit on swimming performance in Atlantic cod (Gadus morhua L.)

Thermal acclimation is frequently cited as a means by which ectothermic animals improve their Darwinian fitness, i.e. the beneficial acclimation hypothesis. As the critical swimming speed (U _crit) test is often used as a proxy measure of fitness, we acclimated Atlantic cod (Gadus morhua) to 4 and 10°C and then assessed their U _crit swimming performance at their respective acclimation temperatures and during acute temperature reversal. Because phenotypic differences exist between different populations of cod, we undertook these experiments in two different populations, North Sea cod and North East Arctic cod. Acclimation to 4 or 10°C had a minimal effect on swimming performance or U _crit, however test temperature did, with all groups having a 10–17% higher U _crit at 10°C. The swimming efficiency was significantly lower in all groups at 4°C arguably due to the compression of the muscle fibre recruitment order. This also led to a reduction in the duration of “kick and glide” swimming at 4°C. No significant differences were seen between the two populations in any of the measured parameters, due possibly to the extended acclimation period. Our data indicate that acclimation imparts little benefit on U _crit swimming test in Atlantic cod. Further efforts need to identify the functional consequences of the long-term thermal acclimation process.     

Altered burst swimming in rainbow trout Oncorhynchus mykiss exposed to natural and synthetic oestrogens

Juvenile rainbow trout Oncorhynchus mykiss were exposed to two concentrations each of 17β‐oestradiol (E2; natural oestrogen hormone) or 17α‐ethinyl oestradiol (EE2; a potent synthetic oestrogen hormone) to evaluate their potential effects on burst‐swimming performance. In each of six successive burst‐swimming assays, burst‐swimming speed (U_burst) was lower in fish exposed to 0·5 and 1 µg l^?1 E2 and EE2 for four days compared with control fish. A practice swim (2 days prior to exposure initiation) in control fish elevated initial U_burst values, but this training effect was not evident in the 1 µg l^?1 EE2‐exposed fish. Several potential oestrogen‐mediated mechanisms for U_burst reductions were investigated, including effects on metabolic products, osmoregulation and blood oxygen‐carrying capacity. Prior to burst‐swimming trials, fish exposed to E2 and EE2 for 4 days had significantly reduced erythrocyte numbers and lower plasma glucose concentrations. After six repeated burst‐swimming trials, plasma glucose, lactate and creatinine concentrations were not significantly different among treatment groups; however, plasma Cl^? concentrations were significantly reduced in E2‐ and EE2‐treated fish. In summary, E2 and EE2 exposure altered oxygen‐carrying capacity ([erythrocytes]) and an osmoregulatory‐related variable ([Cl^?]), effects that may underlie reductions in burst‐swimming speed, which will have implications for fish performance in the wild.     

Heart rates of Atlantic salmon Salmo salar during a critical swim speed test and subsequent recovery

In this study, heart rate (HR) bio-loggers were implanted in the abdominal cavity of 12 post-smolt Atlantic salmon Salmo salar weighing 1024 ± 31 g and acclimated to 12°C sea water. One week after the surgical procedure, a critical swim speed (U_crit) test was performed on tagged and untagged conspecifics, whereafter tagged fish were maintained in their holding tanks for another week. The U_crit was statistically similar between tagged and untagged fish (2.67 ± 0.04 and 2.74 ± 0.05 body lengths s⁻¹, respectively) showing that the bio-logger did not compromise the swimming performance. In the pre-swim week, a diurnal cycle was apparent with HR peaking at 65 beats min⁻¹ during the day and approaching 40 beats min⁻¹ at night. In the U_crit test, HR increased approximately exponentially with swimming speed until a plateau was reached at the final speed before fatigue with a maximum of 85.2 ± 0.7 beats min⁻¹. During subsequent recovery tagged fish could be divided into a surviving group (N = 8) and a moribund group (N = 4). In surviving fish HR had fully recovered to pre-swim levels after 24 h, including reestablishment of a diurnal HR cycle. In moribund fish HR never recovered and remained elevated at c. 80 beats min⁻¹ for 4 days, whereafter they started dying. We did not identify a proximal cause of death in moribund fish, but possible explanations are discussed. Tail beat frequency (TBF) was also measured and showed a more consistent response to increased swimming speeds. As such, when exploring correlations between HR, TBF and metabolic rates at different swimming speeds, TBF provides better predictions. On the contrary, HR measurements in free swimming fish over extended periods of time are useful for other purposes such as assessing the accumulative burden of various stressors and recovery trajectories from exhaustive exercise.