Introduction

To collect fishes for research purposes from rivers and streams, various sampling techniques are used such as cast netting, line fishing and traps. In addition to these methods, electrofishing has become a popular technique due to its convenience and efficiency [1–4]. However, many studies have demonstrated that electroshock treatment has a detrimental effect on fishes (e.g. [5, 6]). Electroshocking often causes spinal damage, reduced growth, and increased mortality in fishes, particularly salmonid fishes (e.g. reviewed in [5, 7]). These studies have also found that mortality and injury rates differ considerably depending on the type of current [continuous direct current (DC), pulsed DC, or alternating current (AC)], wave form, voltage, and duration of shock (salmonids: [8–13]; white sturgeon: [14]; Atlantic herring: [15]). Pulsed DC or AC electrofishing is more damaging than DC electrofishing [6, 8]. Hence, many authors suggest investigating the use of electrofishing to catch freshwater fish to be of high precedence (e.g. [6, 8, 14, 15]). Quantifying the damage resulting due to different fishing techniques, including electrofishing, would be invaluable in gauging the benefits from the viewpoint of animal welfare.

When capturing fishes, researchers should also be concerned about the effects of capture on the physiological stress response in fish. Stress, via the action of cortisol, may result in indirect mortality of fish due to deterioration of immune function (e.g. [16–18]). Similarly, freshwater fishes that have encountered stressors often suffer from increased susceptibility to diseases [19]. Despite extensive research on the effects of electrofishing on physical damage and fish mortality [5], the stress response to electrofishing has been less studied. Consistent results show that regardless of fish species, physiological indicators of stress such as plasma cortisol, glucose, and lactic acid markedly increase after electroshocking (e.g. [20–24]). However, these studies have not compared fish stress responses resulting due to DC and AC electrofishing. Taking into account studies on the negative effects of electricity on fish and the injuries and physiological stress induced by the fishing technique, electrofishing seems to be an improper method for catching freshwater fishes. However, these studies have seldom addressed whether electrofishing causes more physical damage or induces more severe stress responses than any other available sampling methods used to collect freshwater fishes (but see [25]). Irrespective of the method utilized, capture and handling will induce stress in fish. In addition, improper capturing and/or poor handling will often cause physical damage to the fish. Considering these facts, only the most efficient, least stressful, and least damaging method for fish capture should be employed.

The aim of our study was to investigate the stress hormone response of the ayu Plecoglossus altivelis to different fishing methods. P. altivelis is a typical amphidromous fish in East Asia that spends most of its life in freshwater rivers or streams. It is also a commercially important species for freshwater fisheries [26–28]. Net casting and electrofishing have been widely used as catching methods by researchers who have studied the ecology and aquaculture of P. altivelis (electrofisher: e.g. [29–31]; cast net: e.g. [32–34]). However, no studies have investigated the effect of cast netting and electrofishing on the physiological responses, physical damage, and mortality in P. altivelis. Therefore, we assessed the stress hormone responses of P. altivelis to different catching methods (DC electrofishing, AC electrofishing, and net casting) by measuring the changes in serum cortisol concentrations at 0, 24, and 48 h after capture. We included 24 and 48 h time periods to evaluate both acute and chronic damage to P. altivelis. The catch efficiency, injury rates, and 48-h mortality rates observed in the different catching techniques were also compared. From these results, we show the most effective and the least damaging method for collecting P. altivelis in streams with respect to stress, physical damage, and efficiency.

Materials and methods

Experimental procedure

Young P. altivelis captured in a river discharging into Lake Biwa were obtained from a trader. The fish were reared from May to July 2009 in two outdoor stock ponds [0.9 × 4.2 × 0.9 m (width × length × depth), 0.6 m in water depth] at the Ueda Station, National Research Institute of Fisheries Science, Fisheries Research Agency, Ueda, Japan. Each pond was continuously supplied with water from a natural stream through an inlet. Fish density was approximately 220 fish/m3 (500 fish per pond) or 4.6 kg/m3. During this rearing period, the fish were maintained under natural photoperiod at 15.2–21.5 °C. They were fed three times a day with a commercial diet using automatic feeders until they reached approximately 110 mm in standard length, SL and 17 g in body mass, BM. Note that the breeding season for P. altivelis is October–November, and thus all the fish used for these experiments were sexually immature.

Three outdoor artificial streams [2.2 × 4.2 × 0.9 m (width × length × depth), 0.3 m in water depth] adjacent to the stock ponds were used for catching experiments. Water in both the artificial streams and the stock ponds was supplied from the same natural stream, and thus, turbidity, water temperature, and photoperiodicity were identical between the two reservoirs. At 15:30 on the day before catching, about 130 fish were transferred into the three artificial streams (43 ± 3 fish per stream), using a hand net. As a control experiment, 45 fish were transferred into a box-shaped net cage (1.0 × 1.0 × 1.0 m with an open top) set in a pond identical to the stock ponds. Catching experiments were conducted between 11:00 and 13:00 using an electrofisher (Backpack Electrofisher, model 12-B; Smith-Root; http://www.smith-root.com/) or a cast net (Mitani Gyogu; http://www.mitani-gyogu.jp/index.html). Twelve fish in each pond were captured using DC electrofishing (200 V), AC electrofishing (200 V, 45 Hz, 0.3–6 ms), and a cast net (mesh size 17.8 mm, net length 2.8 m, maximum net diameter 4.0 m) by an expert in both electrofishing and net casting (K.I.). These fish were immediately transferred into a lethal dose of anesthesia (200 mg/l of MS-222; Sigma-Aldrich; http://www.sigmaaldrich.com/japan.html), using a hand net, after which their blood was sampled by cutting the caudal peduncle with a small kitchen knife or scalpel. Twelve fish in the net cage for the control experiments were caught using a hand net and immediately administered a lethal dose of anesthesia. Their blood was sampled as described previously, acquired within 30 s from the start of capture to the time when the 12 control fish were administered anesthesia; a previous study has shown that this sampling method does not elevate serum cortisol levels in P. altivelis [28]. The fish were then measured (SL, to the nearest 0.1 mm) and weighed (BM, to the nearest 0.1 g). Just before the start of the experiment, the temperature of the water in the reservoirs was 21.5 °C and the turbidity was considerably low, at a value of 3.0 mg/l (measured using a turbidity sensor, TCR-5Z; Kasahara Chemical Instruments Co.; http://www.krkjpn.co.jp/handy/TCR-5Z.htm). Fish densities in the experimental streams and in the net cage were approximately 16 and 45 fish/m3, respectively, both of which were far lower than those in the stock ponds (220 fish/m3). Therefore, turbidity and fish density were unlikely to contribute to the stress responses of the fish during experiments (see [19, 28, 35]). In addition, although we were only able to complete one experimental trial for each capture method, we ensured that stream effects were unlikely to be large sources of error by using three identical artificial stream channels.

Catch efficiency by each sampling method was defined as the time from the start of catching to the time when all the 12 fish were administered anesthesia. We did not attempt to catch individual fish, and thus often caught several at once. Three box-shaped net cages, which were identical to the ones used in the control experiments, were prepared in advance to be set in the ponds. The remaining fish in each experimental stream (29 ± 1 fish per stream) were captured using the same catching equipment and immediately transferred into the net cages. Fish in the net cages of control and catching experiments were sampled 24 and 48 h after capturing treatment with a hand net (12 fish per experiment); at this point, their blood samples were collected and body measurements were taken. The sampling method used here was the same as used when the control fish were captured at 0 h after treatment. Wounds, abrasions, and fractures were also checked by observing their bodies. Residual fish in the net cage (alive and dead) were counted to calculate the observed mortality at 48 h. The living fish were then returned to the stock ponds.

Cortisol measurement

Blood was centrifuged at 7,500 rpm for 5 min and the serum was removed and frozen at −30 °C until analysis could be performed. Cortisol measurements were taken on each sample in duplicate using the Cortisol EIA Kit according to the manufacturer’s instructions (EA65, Oxford Biomedical Research Inc.; http://www.oxfordbiomed.com/). Further details on cortisol measurement have been described previously [28].

Data analyses

Effect of the sampling methods and time since treatment on serum cortisol concentrations was examined using two-way analysis of variance (ANOVA). Since a significant interaction was detected (see Results), Dunnett’s post hoc tests were used to differentiate the cortisol levels between fish captured by DC electrofishing, AC electrofishing, or cast net and control fish at each sampling time. Tukey’s honestly significant difference (HSD) post hoc tests were applied when changes in cortisol levels were assessed for each sampling method. To satisfy the normal distribution requirements for parametric tests, outliers with defined values greater or less than 1.5 times the interquartile range were removed prior to analyses. As a consequence, 2 % of the total values were removed. The mean SL and BM of P. altivelis were 112.7 mm (range = 71.0–148.0 mm, n = 129) and 17.1 g (range = 4.5–41.3 g), respectively. There were no significant differences in SL and BM among the treatment groups (SL: one-way ANOVA, F 11,116  = 1.16, p = 0.32; BM: F 11,116  = 1.08, p = 0.38). The observed 48-h mortality was compared using Fisher’s exact test. All analyses were conducted using SPSS 16.0 (SPSS Inc.; http://www-01.ibm.com/software/analytics/spss/). All given p values are two-tailed and α was set at 0.05.

Results

Significant interactions between sampling methods and time since treatments were detected for cortisol concentrations (two-way ANOVA, F 6,117  = 3.92, p = 0.001), indicating that stress hormone responses were different among the sampling methods at the different time points. Therefore, cortisol levels were compared between different sampling methods and control treatments at each sampling time and changes in cortisol levels were assessed for each sampling method.

Immediately after capture, cortisol concentrations were significantly higher in fish caught using a cast net than control fish (Dunnett’s post hoc test, p < 0.0001; Fig. 1). The fish caught by DC and AC electrofishing did not have different cortisol concentrations compared to the control fish (p > 0.10 for both capture treatment). It took a shorter time to catch 12 fish using DC electrofishing (20 s) compared to AC electrofishing (45 s). More time was required when a cast net was used (840 s). The cortisol concentrations in fish captured by cast net were significantly higher than the levels found in control fish 24 h after capture (p = 0.04; Fig. 1). Fish caught by AC electrofishing exhibited greater cortisol levels than the control fish 48 h after capture (p = 0.05; Fig. 1). Although not significant, a similar trend was seen in fish captured by cast net at the 48-h time point (p = 0.07).

Fig. 1
figure 1

Serum cortisol concentrations in Plecoglossus altivelis 0, 24, and 48 h following DC electrofishing, AC electrofishing, cast netting, or control treatment. Bars represent mean + SE. Numerals in bars show sample sizes. ***p < 0.0001, *p ≤ 0.05, † p = 0.07, compared with control at each sampling time (Dunnett’s post hoc test)

Full size image

Cortisol levels declined significantly with time following capture treatment in both the DC electrofishing (Tukey HSD post hoc test, 0 vs. 24 or 48 h: p < 0.02, 24 vs. 48 h: p > 0.99) and cast netting groups (0 vs. 24 or 48 h: p < 0.0001, 24 vs. 48 h: p = 0.58; Fig. 1). This trend was also observed in fish that received the control treatment (0 vs. 24 h: p = 0.01, 0 vs. 48 h: p = 0.08, 24 vs. 48 h: p = 0.69). Cortisol concentrations after AC electroshocking decreased from 0 to 24 h, but increased again from 24 to 48 h (0 vs. 24 h: p = 0.0005, 24 vs. 48 h: p = 0.02; 0 vs. 48 h: p = 0.43; Fig. 1). The observed 48-h mortality of fish exposed to AC electroshocking (dead/total fish = 3/40, 7.5 %) was significantly higher than that obtained in fish exposed to both DC electroshocking (0/45, 0 %) and cast netting (0/45, 0 %; Fisher’s exact test, p = 0.03). Furthermore, one of the 40 fish with AC electroshocking exposure exhibited spinal injury.

Discussion

Numerous studies have documented the negative effects, both physical and physiological, of electrofishing (reviewed in [5]). In addition, many authors have reported the efficiency of electrofishing in comparison to other fish catching methods (e.g. [2–4]). However, based on the reviewed literature, this may be the first study to examine both the harmful effects (stress, injury, and mortality) and the catching efficiency of different catching methods of freshwater fish. Our results found increased cortisol levels in fish caught using a cast net immediately after capture. Cortisol concentrations did not differ between the fish captured by electrofishing (DC and AC) and control fish. Previous studies have also found no significant increase in cortisol levels following exposure to electroshocking ([24, 36, 37]; but see [21]). Elevated cortisol concentrations in fish captured by cast net were probably the result of a continuously stressful situation (840 s needed to catch 12 fish) such as the shadow of a person, a splash produced by the cast net, and/or being netted. In fact, elevations in cortisol levels were generally detected within 10 min after stress exposure (reviewed in [38]). Cast netting in an unconfined natural stream may be less stressful than observed in the artificial streams because P. altivelis may be able to move away from and avoid the cast netter. However, such movements would ensure even lower capture efficiency than observed in this study. Nevertheless, this study revealed that the efficiency of electrofishing was found to be higher than that of net casting. Additionally, in the case of P. altivelis, electrofishing was less stressful compared to cast netting. Thus, electrofishing is certainly a better method for catching P. altivelis in artificial streams, and likely, a preferable technique in natural streams.

Cortisol concentrations significantly decreased in all treatment groups including the controls 24 h after capture. Fish captured by cast net, however, showed higher cortisol levels compared to control fish, and such a trend continued even 48 h after capture. In fish, blood cortisol returns to pre-stress levels within 24 h and generally within 6 h [36, 37, 39, 40]. Similar results were obtained in our studies with P. altivelis (Yada T, Awata S, Tsuruta T and Iguchi K, unpubl. data). If the stressful situation continues then the return rates of cortisol may be slowed [38]. Thus, the relatively high cortisol levels in fish 24 and 48 h after capture might result from other stressors induced during cast netting, such as rubbing of skin mucus and/or scale removal. Average blood cortisol levels in the control fish at 0 h was approximately 50 ng/ml, which is higher than the resting levels previously reported for P. altivelis (<10 ng/ml, [28]). This is probably because of the transfer to a new environment a day before the experiment. Similar results were obtained in a previous study [28].

Both control fish and fish captured by DC electrofishing exhibited low levels of blood cortisol 48 h after capture, similar to the levels seen at 24 h. However, cortisol concentrations in fish captured by AC electrofishing were higher than that in control fish at this time. Spinal damage was observed in one fish captured by AC electrofishing but not in the fish captured by DC electroshocking. Although internal examinations using radiography (e.g. reviewed in [5]) were not conducted, several studies have shown that AC electroshocking causes more spinal injuries than do other electrical settings [6, 8, 25]. Thus, the extent of spinal damage after AC electrofishing may be underestimated in our study. Furthermore, in our study, mortality was observed only in the fish captured by AC electrofishing. Taken together, our data comparing three methods of capturing (AC electrofishing, DC electrofishing, and net casting) suggest that AC electrofishing is the most harmful catching method for P. altivelis, since it may cause spinal damage resulting in chronic stress and high mortality.

To summarize, our results reveal that in the case of P. altivelis, DC electrofishing causes less stress, lower spinal damage, and lower mortality than other methods. Catch efficiency with DC electrofishing was far greater than that of other methods. Therefore, it can be concluded that DC electrofishing is the most effective and the least damaging method for catching P. altivelis in artificial streams, and likely in natural streams. Although more research is required to determine whether DC electrofishing is similarly suitable for capturing other freshwater fishes, our findings may be applicable to small fishes occupying a habitat similar to P. altivelis (i.e. streams). In fact, many researchers use electrofishing to study the life history, distribution, and movement of freshwater fishes in natural streams (Gobiidae: [41]; Cyprinidae: [42]; Cobitidae: [43]; Salmonidae: [44, 45]) including P. altivelis [29, 31]. In addition, as shown in a previous study [37], DC electrofishing may be useful for assessing stress responses in P. altivelis since the fishing method does not lead to an immediate increase in cortisol concentrations following catching.