Radio-transmitted electromyogram signals as indicators of swimming speed in lake trout and brown trout

Swimming speed and average electromyogram (EMG) pulse intervals were highly correlated in individual lake trout Salvelinus namaycush ( r ²=0·52–0·89) and brown trout Salmo trutta ( r ²=0·45–0·96). High correlations were found also for pooled data in both lake trout ( r ²=0·90) and brown trout of the Emå stock ( r ²=0·96) and Lærdal stock ( r ²=0·96). The linear relationship between swimming speed and average EMG pulse intervals differed significantly among lake trout and the brown trout stocks. This successful calibration of EMGs to swimming speed opens the possibility of recording swimming speed of free swimming lake trout and brown trout in situ . EMGs can also be calibrated to oxygen consumption to record energy expenditure.     

Importance of electrode positioning in biotelemetry studies estimating muscle activity in fish

Red and white axial muscle activity of adult Atlantic salmon Salmo salar was examined using conventional electromyography (EMG x ) and activity radio-transmitters (EMG i ) at 0·5 and 0.7 body lengths (L) along the body of the fish. Critical swimming trials were conducted and maximum sustainable speeds (U_crit) were unaffected by the presence of electrodes, being 1·51 ± 21 m s⁻¹ (3.33 ± 0.34 L s⁻¹) ( n =44). Regardless of longitudinal position of the electrodes within the musculature, both EMG x s and EMG i s indicated increasing red muscle activity with increasing swimming speed, whereas white muscle fibres were recruited only at speeds > 86±5% U_crit. Telemetered EMG i signals indicated that muscle activity varied significantly for electrodes implanted at different longitudinal positions along the fish ( P < 0·001). These results suggest that electrode placement is an important influence affecting the signals obtained from radio transmitters that estimate activity and location should be standardized within biotelemetry studies to allow accurate and consistent comparisons of activity between individuals and species. Optimal location for electrode placement was determined to be in the red muscle, towards the tail of the fish (0·7 L ).     

Endurance swimming of diploid and triploid Atlantic salmon

When groups of diploid (mean ±  s . e . fork length, L _F) 33·0 ± 1·4 cm and triploid (35·3 ± 0·5 cm) Atlantic salmon Salmo salar were forced to swim at controlled speeds in a carefully monitored 10 m diameter 'annular' tank no significant difference was found between the maximum sustained swimming speeds ( U _ms, maintainable for 200 min) where the fish swam at the limit of their aerobic capability. Diploids achieved 2·99 body lengths per second (bl s⁻¹)(0·96 m s⁻¹) and triploids sustained 2·91 bl s⁻¹(1·02 m s⁻¹). The selection of fish for the trials was based on their ability to swim with a moving pattern projected from a gantry rotating at the radius of the tank and the selection procedure did not prove to be significant by ploidy. A significant difference was found between the anaerobic capabilities of the fish measured as endurance times at their prolonged swimming speeds. During the course of the experimentation the voluntary swimming speed selected by the fish increased and the schooling behaviour improved. The effect of the curvature of the tank on the fish speeds was calculated (removing the curved effect of the tank increased the speed in either ploidy by 5·5%). Implications of the endurance times and speeds are discussed with reference to the aquaculture of triploid Atlantic salmon.     

Field-based measurements of oxygen uptake and swimming performance with adult Pacific salmon using a mobile respirometer swim tunnel

Novel field measurements of critical swimming speed ( U _crit) and oxygen uptake (  M o₂) in three species of adult Pacific salmon Oncorhynchus spp. up to 3·5 kg in body mass were made using two newly designed, mobile Brett-type swim tunnel respirometers sited at a number of field locations in British Columbia, Canada. Measurements of U _crit, which ranged from 1· 68 to 2·17 body lengths s⁻¹, and maximum M o₂, which ranged from 8·74 to 12·63 mg O₂ kg⁻¹ min⁻¹ depending on the species and field location, were judged to be of similar quality when compared with available data for laboratory-based studies. Therefore high quality respirometry studies were possible in the field using adult wild swimming salmonids. In addition, the recovery of wild adult Pacific salmon from the exhaustive U _crit swim test was sufficiently rapid that swimming performance could be repeated with <1 h of recovery time between the termination of the initial swim test and the start of the second test. Moreover, this repeat swimming performance was possible without routine M o₂ being reestablished. This result suggests that wild adult salmon are capable of carrying a moderate excess post-exercise oxygen consumption without adversely affecting U _crit, maximum M o₂ or swimming economy. Such capabilities may be extremely important for timely migratory passages when salmonids face repetitive hydraulic challenges on their upstream migration.     

A biotelemetry system recording fish activity

A biotelemetry system is described for obtaining, transmitting and recording the electromyograms (EMGs) produced in muscle activity of free-swimming fish as quantitative indicators of overall fish activity. The radiotransmitters used come in the form of cylindrical packages having two sensing electrodes, all fully implantable in the fish body cavity. EMGs are transmitted as radio pulses with the intensity of muscular activity determining the intervals between pulses. The packages also contain temperature sensors and fish temperatures are transmitted with every 32nd pulse. Transmitted EMG pulses are detected, 'measured' and stored by a single portable receiver (Model SRX_400, Lotek). Data can be subsequently transferred to a computer (which can also be portable) for storage, processing and statistical analysis. Transmitter battery life can be in excess of 7 months, permitting laboratory or field studies of long duration. Transmitter package implantation surgery requires a mid-ventral incision and internal securing of transmitter and sensing electrodes. Surgical silk, cyanoacrylate tissue adhesives, and polydioxanone (PD), a synthetic absorbable suture, were all tried as means of incision closure. The most effective was PD alone. Trials of the system consisted of forced swims by transmitter-equipped rainbow trout Oncorhynchus mykiss Walbaum. The data obtained provided an inverse linear relation between forced swim speed and EMG pulse interval. Trials were conducted at intervals over periods up to 2 months. Fish showed neither distress, nor difficulty in swimming up to maximum speeds of 60 cm s ⁻¹ (fish lengths 41.0, 44.4 cm).     

Short‐term freshwater exposure benefits sea lice‐infected Atlantic salmon

Atlantic salmon Salmo salar were infected with sea lice Lepeophtheirus salmonis (0·08 ± 0·007 sea lice g⁻¹) over a period of 4 h. Both infected and non‐infected fish were swim tested in sea water (SW) and fresh water (FW). The ventral aorta of each fish was fitted with a Doppler cuff in order to measure cardiac output, stroke volume and heart rate during swim testing. Blood samples were taken at rest and after exercise. Critical swimming speed of infected fish in SW (2·14 ± 0·08 body lengths, bl s⁻¹) was significantly lower ( P  < 0·05) than infected fish switched to FW (2·81 ± 0·08 bl s⁻¹) and non‐infected fish in SW (2·42 ± 0·04 bl s⁻¹) and FW (2·61 ± 0·08 bl s⁻¹). Cardiac and blood results indicated infected fish exposed to FW did experience stress, but relief from osmotic and ionic distress probably reduces energy expenditure, allowing the increase in performance. As the performance of sea lice‐infected fish improved upon transfer to FW, it is likely that heavily infected salmonids do return to FW to restore compromised osmotic and ionic balance, and remove sea lice in the process.     

The angle of attack of the body of common bream while swimming at different speeds in a flume tank

The body attack angle of common bream Abramis brama varied with swimming speed and was best described by θ = −3·32 (±0·24) − 9·23 (±0·54)e^{− u} ( r ² = 0·56, P  < 0·0001; ±1 s . e . given in parentheses), where θ is the body attack angle and u is swimming speed. The hypothesis that neutrally buoyant fishes may swim with body attack angles deviating increasingly from 0° as the swimming speed decreases is supported.     

Hydroacoustic measurement of swimming speed of North Sea saithe in the field

Saithe Pollachius virens , tracked diurnally with a split-beam echosounder, showed no relationship between size and swimming speed. The average and the median swimming speeds were 1·05 m s⁻¹(±0·09 m s⁻¹) and 0·93 m s⁻¹, respectively. However, ping-to-ping speeds up to 3·34 m s⁻¹ were measured for 25–29 cm fish, whose swimming speeds were significantly higher at night (1·08 m s⁻¹) than during the day (0·72 m s⁻¹). The high average swimming speed could be related to the foraging or streaming part of the population and not to potential weakness of the methodology. However, the uncertainty of target location increased with depth and resulted in calculated average swimming speeds of 0·15 m s⁻¹ even for a stationary target. With increasing swimming speed the average error decreased to 0 m s⁻¹ for speeds >0·34 m s⁻¹. Species identity was verified by trawling in a pelagic layer and on the bottom.     

Endurance swimming of European eel

A long‐term swim trial was performed with five female silver eels Anguilla anguilla of 0·8–1·0 kg ( c . 80 cm total length, L _T) swimming at 0·5 body lengths (BL) s⁻¹, corresponding to the mean swimming speed during spawning migration. The design of the Blazka‐type swim tunnel was significantly improved, and for the first time the flow pattern of a swim tunnel for fish was evaluated with the Laser‐Doppler method. The velocity profile over three different cross‐sections was determined. It was observed that 80% of the water velocity drop‐off occurred over a boundary layer of 20 mm. Therefore, swim velocity errors were negligible as the eels always swam outside this layer. The fish were able to swim continuously day and night during a period of 3 months in the swim tunnel through which fresh water at 19° C was passed. The oxygen consumption rates remained stable at 36·9 ± 2·9 mg O₂ kg⁻¹ h⁻¹ over the 3 months swimming period for all tested eels. The mean cost of transportation was 28·2 mg O₂ kg⁻¹ km⁻¹. From the total energy consumption the calculated decline in fat content was 30%. When extrapolating to 6000 km this would have been 60%, leaving only 40% of the total energy reserves for reproduction after arriving at the spawning site. Therefore low cost of transport combined with high fat content are crucial for the capacity of the eel to cross the Atlantic Ocean and reproduce.     

Swimming performance, oxygen consumption and excess post-exercise oxygen consumption in adult transgenic and ocean-ranched coho salmon

Routine oxygen consumption ( M o ₂) was 35% higher in 1 day starved and 21% higher in 4 day starved adult transgenic coho salmon Oncorhynchus kisutch relative to end of migration ocean-ranched coho salmon. Critical swimming speed ( U _crit) and M o ₂ at U _crit ( M o _2max) were significantly lower in 4 day starved transgenic coho salmon (1·25 BL s⁻¹; 8·79 mg O₂ kg⁻¹ min⁻¹) compared to ocean-ranched coho salmon (1·60 BL s⁻¹; 9·87 mg O₂ kg⁻¹ min⁻¹). Transgenic fish swam energetically less efficiently than ocean-ranched fish, as indicated by a poorer swimming economy at U _crit ( M o _2max     ). Although M o _2max was lower in transgenic coho salmon, the excess post-exercise oxygen consumption (EPOC) measured during the first 20 min of recovery was significantly larger in transgenic coho salmon (44·1 mg O₂ kg⁻¹) compared with ocean-ranched coho salmon (34·2 mg O₂ kg⁻¹), which had a faster rate of recovery.     

Physiological impact of sea lice on swimming performance of Atlantic salmon

Atlantic salmon Salmo salar were infected with two levels of sea lice Lepeophtheirus salmonis (0·13 ± 0·02 and 0·02 ± 0·00 sea lice g⁻¹). Once sea lice became adults, the ventral aorta of each fish was fitted with a Doppler cuff to measure cardiac output ( ̇ ), heart rate ( f _H) and stroke volume ( V _S) during swimming. Critical swimming speeds ( U _crit) of fish with higher sea lice numbers [2·1 ± 0·1 BL (body lengths) s⁻¹] were significantly lower ( P  < 0·05) than fish with lower numbers (2·4 ± 0·1 BL s⁻¹) and controls (sham infected, 2·6 ± 0·1 BL s⁻¹). After swimming, chloride levels in fish with higher sea lice numbers (184·4 ± 11·3 mmol l⁻¹) increased significantly (54%) from levels at rest and were significantly higher than fish with fewer lice (142·0 ± 3·7 mmol l⁻¹) or control fish (159·5 ± 3·5 mmol l⁻¹). The f _H of fish with more lice was 9% slower than the other two groups at U _crit. This decrease resulted in ̇ not increasing from resting levels. Sublethal infection by sea lice compromised the overall fitness of Atlantic salmon. The level of sea lice infection used in the present study was lower than has previously been reported to be detrimental to wild Atlantic salmon.     

Effect of temperature on swimming performance of sea bass juveniles

At four temperatures ( T= 15, 20, 25 and 28° C) swimming performance of Dicentrarchus labrax was significantly correlated with total length (23–43 mm L _T); r²=0.623–0.829). The relative critical swimming speed ( RU _crit= U _crit L _T⁻¹), where U _crit is the critical swimming speed, was constant throughout the L _T range studied. The significant effect of temperature on the relative critical swimming speed was described binomially: RU _crit=−0.0323T²+ 1.578 T −10.588 (r²=1). The estimated maximum RU _crit (8.69 L _T s⁻¹) was achieved at 24.4° C, and the 90% performance level was estimated between 19.3 and 29.6° C.     

Factors affecting river entry of adult Atlantic salmon in a small river

In this study, effects of stock origin, fish size, water flow and temperature on time of river ascent of adult Atlantic salmon Salmo salar were tested. Brood stocks were collected in eight Norwegian rivers situated between 59 and 69° N. The fish were reared to smolts, individually tagged and released in the River Imsa, south-west Norway (59° N). Adults from all stocks approached the Norwegian coast concurrently, but Atlantic salmon ≥70 cm in natural tip length entered coastal water slightly earlier during summer than smaller fish. Atlantic salmon <70 cm, however, ascended the river significantly earlier and at lower water flow and higher water temperature than larger fish. Although largest in size, the fish from the northern populations (62–69° N) ascended the River Imsa almost 1 month earlier than those from the south (59–60° N). They seemed less restricted by the environmental factors than the fish originating from the more southern rivers. There was no apparent trend among years in time of river ascent. Maximum ascent per day occurred at water discharges between 12·5 and 15 m³ s⁻¹ and at water temperatures between 10 and 12·5° C. There was a significant positive correlation between water flow and river ascent during the first part of the upstream run from July to September with best correlation for September, when multiple regression analysis indicated that water temperature had an additional positive effect. Stock origin, fish size and water discharge were important variables influencing the upstream migration of Atlantic salmon in small rivers.     

Swimming behaviour of upriver migrating sea lamprey assessed by electromyogram telemetry

The main subject of this study was the swimming behaviour of upriver migrating sea lamprey, Petromyzon marinus , with particular focus on identification of their swim strategies to overcome areas of difficult passage. A biotelemetry technique (electromyogram telemetry) was used to register muscle activity of the tagged animals. In the 2005 spawning season, five adult sea lampreys were surgically tagged and released in the field. Before release, electromyogram (EMG) records were calibrated with the P. marinus swimming speed in a swim tunnel. Differences between ground speed and swimming speed in the wild suggest that the calibrated CEMG (coded electromyogram) transmitter output corresponds to an activity index, and cannot be properly related to actual swimming speed. This study notes the need to confirm the laboratory calibration curves, to ascertain their use in determining swimming speed of tagged fish in the wild. In 2006, in order to confirm the field results seven adult sea lampreys were tagged, calibrated in the laboratory and released in a 30-m long experimental outdoor canal. The results were similar: observed swimming speed was generally higher when compared with the swimming speed obtained with the EMG signal. In the river, when swimming through slow-flow stretches, sea lampreys maintained a constant pattern of activity, attaining an average ground speed of 0.76 BL s⁻¹ (2.5 km h⁻¹). When sea lampreys encountered rapid flow reaches they alternated between short movements ( c. 67 s) and periods of rest ( c. 99 s). In each swim bout they progressed approximately 14 m; to overcome more difficult obstacles sea lampreys increased their number of burst movements instead of longer or more violent swimming events. About 43% of the time negotiating difficult passage areas was spent in resting by attaching motionless to the substrate with their oral disk.     

Is tilting behaviour at low swimming speeds unique to negatively buoyant fish? Observations on steelhead trout,Oncorhynchus mykiss,and bluegill,Lepomis macrochirus

When swimming at low speeds, steelhead trout and bluegill sunfish tilted the body at an angle to the mean swimming direction. Trout swam using continuous body/caudal fin undulation, with a positive (head-up) tilt angle ( 0 , degrees) that decreased with swimming speed ( u , cm s⁻¹) according to: 0 =(164±96).u^{(−1.14±0.41)} (regression coefficients; mean±2 s.e. ). Bluegill swimming gaits were more diverse and negative (head down) tilt angles were usual. Tilt angle was −3·0 ± 0.9° in pectoral fin swimming at speeds of approximately 0.2–1.7 body length s⁻¹ (Ls⁻¹; 3–24 cm s⁻¹), −4.5 ±2.6° during pectoral fin plus body/caudal fin swimming at 1·2–1·7 L s⁻¹ (17–24cm s⁻¹), and −5.0± 1.0° during continuous body/caudal fin swimming at 1.6 and 2.5 L s⁻¹ (22 and 35cm s⁻¹). At higher speeds, bluegill used burst-and-coast swimming for which the tilt angle was 0.1±0.6°. These observations suggest that tilting is a general phenomenon of low speed swimming at which stabilizers lose their effectiveness. Tilting is interpreted as an active compensatory mechanism associated with increased drag and concomitant increased propulsor velocities to provide better stabilizing forces. Increased drag associated with trimming also explains the well-known observation that the relationship between tail-beat frequency and swimming speed does not pass through the origin. Energy dissipated because of the drag increases at low swimming speeds is presumably smaller than that which would occur with unstable swimming.     

Reduced swimming performance in delta smelt infected with Mycobacterium spp.

Delta smelt Hypomesus transpacificus infected with Mycobacterium spp. swam significantly more slowly (mean ± s.e ., 24±5 ± 1·2 cm s ⁻¹) than uninfected fish (30·0 ± 1·7 cm s ⁻¹). Differences in swimming performance were not attributable to differences in fish size ( L _s or wet mass), condition factor or laboratory holding duration. Similar proportions of non-fatigue-related swimming failure among the uninfected and infected fish indicated that mycobacteriosis did not affect the willingness of delta smelt to swim in the flume. Level of infection, measured for the dominant M. chelonae pathogen using enzyme-linked immunosorbent assay (ELISA), did not affect critical swimming velocity.     

A hydro-mechanical approach to the scaling of swimming performance in the sand flathead Platycephalus bassensis Cuvier: effects of changes in morphological features based on fish size

The swimming performance of Platycephalus bassensis at steady speed was assessed with an emphasis on hydrodynamics. The minimum swimming speed to maintain hydrostatic equilibrium for P. bassensis of 0·271 m total length ( L _T) was calculated to be 1·06 L _T s⁻¹. At this speed, the required lift to support the mass of the fish was equivalent to 6·6% of the fish mass; 82·7% of which was created by the body as a hydrofoil, and the rest of which was created by the pelvic fins as hydrofoils. The minimum swimming speed decreased with the L _T of the fish and ranged from 1·15 L _T s⁻¹ for a fish of 0·209 m to 0·89 L _T s⁻¹ for a fish of 0·407 m. The forward movement per tail-beat cycle ( i.e. stride length) was described with an equation including quantities of morphological and hydro-mechanical relevance. This equation explained that stride length was increased by the effect of turbulence characterized by the Reynolds number and demonstrated the morphological and hydro-mechanical functional design of the fish for maximizing thrust and minimizing drag. The larger span of the caudal fin and caudal tail-beat amplitude was associated with larger stride length, whereas greater frictional drag was associated with smaller stride length.     

Sea growth, smolt age and age at sexual maturation in Atlantic salmon

Relationships between growth at sea, smolt size and age at sexual maturation of Atlantic salmon Salmo salar were tested. The fish were offspring of brood stocks sampled in eight Norwegian rivers at latitudes between 59° and 70° N, hatchery reared and released at smolting at the mouth of the River Imsa (59° N). Smolt size influenced the subsequent growth rate of Atlantic salmon. The larger the fish were at release, the slower the yearly length increment at sea. Mean sea age at sexual maturity, measured as proportion of the returning adults attaining sexual maturity at sea age 2 years, was significantly correlated with mean growth rate during the first year at sea and mean smolt size ( r ²= 0·74, P < 0·001). Fish attaining maturity at a relatively high sea age were more fast growing during their first year at sea than those maturing at a younger age. The results indicate that high sea age at sexual maturation is a population-specific characteristic and associated with high early growth rate at sea.     

Lacustrine growth of juvenile pink salmon and a comparison with sympatric sockeye salmon

The growth rates of naturally sympatric juvenile pink Oncorhynchus gorbuscha and sockeye Oncorhynchus nerka salmon were compared in a common lacustrine environment in south‐west Alsaka, an unusual opportunity given the normal disparity in freshwater residence time of these two species. Fork length ( L _F) frequency distributions of juvenile pink salmon caught in the lake during the summer in 1991 and 1999–2003 indicated a growth rate of 0·54 mm day⁻¹, 54% greater than the estimated growth rate of juvenile sockeye salmon sampled from 1958 to 2003 (0·35 mm day⁻¹). Examination of daily growth rings on otoliths indicated that pink salmon in Lake Aleknagik grew an average of 1·34 mm day⁻¹ in 2003 but sockeye salmon grew only 0·63 mm day⁻¹(average specific growth rates were 3·0 and 1·8% day⁻¹, respectively). Pink salmon increased from c . 32 mm L _F and 0·2 g at emergence to 78 mm L _F and 3·0 g within 3–4 weeks. After experiencing these rapid growth rates, the pink salmon appeared to leave the lake by late July in most years. The diets of pink and sockeye salmon in the littoral zone of the lake were very similar; >80% of the stomach contents consisted of adult and pupal insects and the remainder was zooplankton. This high degree of diet overlap suggested that the observed differences in growth rate were not attributable to variation in prey composition.     

Reproductive ecology of cultured fish in the wild

Domestication has been shown to have an effect on morphology and behaviour of Atlantic salmon ( Salmo salar ). We compared swimming costs of three groups of juvenile Atlantic salmon subject to different levels of domestication: (1) wild fish; (2) first generation farmed fish origination from wild genitors; and (2) seventh generation farmed fish originating from Norwegian aquaculture stocks. We assessed swimming costs under two types of turbulent flow (one mean flow velocity of 23 cm s⁻¹ and two standard deviations of flow velocity of 5 and 8 cm s⁻¹). Respirometry experiments were conducted with fish in a mass range of 5–15 g wet at a water temperature of 15° C. Our results confirm (1) that net swimming costs are affected by different levels of turbulence such that, for a given mean flow velocity, fish spent 1·5‐times more energy as turbulence increased, (2) that domesticated fish differed in their morphology (having deeper bodies and smaller fins) and in their net swimming costs (being up to 30·3% higher than for wild fish) and (3) that swimming cost models developed for farmed fish may be also be applied to wild fish in turbulent environments.