© 2004 by ICES/CIEM International Council for the Exploration of the Sea/Conseil International pour l'Exploration de la Mer
Reaction of juvenile flounder to grid separators
a National Research Institute of Fisheries Engineering Hasaki, Ibaraki 314-0421, Japan
b Chiba Prefecture Fisheries Research Center Chikura, Chiba 295-0024, Japan
*Correspondence to Y. Matsushita: tel: +81 479 44 5952, fax: +81 479 44 6221. e-mail: yoshiki{at}fra.affrc.go.jp.
The reaction behaviour of juvenile Japanese flounder (Paralichthys olivaceus) to towed grids (0.5 x 0.2 m, horizontally, or vertically orientated bars at 10-mm intervals) was observed as a means of understanding fish behaviour in relation to grid selection for a beam trawl fishery in Tokyo Bay. Reaction behaviours were categorized within four patterns by grid types and illumination levels: (i) forward swimming in towed direction; (ii) swimming over the grid; (iii) sticking on the grid; and (iv) passing through the grid. The most dominant reaction pattern was forward swimming, but its ratio was higher for light than for dark conditions. Passing through the grid bars occurred most frequently with horizontal bars. Approximately 40% of tested fish passed through the grid in light conditions, approximately 30% in dark conditions. Most of these fish penetrated bar gaps head first, while a considerable proportion categorized as "forward swimming" kept swimming even though their tails or bodies had partly passed between the bars. It is concluded that penetration of flounders through bar gaps is governed by voluntary actions.
Keywords: bar orientation, grid separator, Japanese flounder, Paralichthys olivaceus, reaction behaviour, tank experiment
Received 13 March 2003; accepted 26 April 2004.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
More than 300 coastal beam trawlers operate in Tokyo Bay, Japan, harvesting a variety of marine organisms varying by area and season. Fine diamond mesh (approx. 25 mm) codends are employed to retain the main target species, conger eel (Conger myriaster), in summer. This is the season, however, when juveniles of flounder species such as Japanese flounder (Paralichthys olivaceus) and marbled sole (Limanda yokohamae) are present on the fishing grounds, and by-catch of these juveniles is a concern. We have been developing by-catch reduction devices for this fishery that exclude juvenile flounders while retaining conger eel.
Our observations of fish behaviour in conventional otter trawl codends in waters outside Tokyo Bay indicate that most flounder species are forced back near the bottom panel in the codend (Matsushita et al., 1998). In addition, Japanese flounder are known to be bottom-dwelling and negatively buoyant (Kawabe et al., 2003). Based on these facts, a codend was designed and constructed with a grid separator (stainless steel, L 100 x H 20 cm, with horizontally or vertically oriented bars) across the bottom part of the codend (Figure 1). Fish entering the codend near the top pass to the back; fish entering low encounter the grid. While grid designs developed in other regions are aimed at repulsing fish penetration (e.g. Isaksen et al., 1992; Kennelly and Broadhurst, 1995; Rose and Gauvin, 2000), the grid in this design was considered as an escape vent with the expectation that juvenile flounder would pass through the bar gaps and exit the codend.
|
The performance of this design has been tested in fishing trials since summer 2000 using a cover-net method. The gear retains most eels, but the exclusion of flounders was not achieved at the desired level (Matsushita, unpublished). To investigate the causes of this poor performance, we tried to observe fish behaviour in the codend with an underwater video camera during actual fishing practices, but year-round turbidity on the fishing grounds inhibited filming. In this study, the reaction behaviour of juvenile flounders to two types of grids was examined in an experimental tank to select the more effective separator grid design with the understanding that fish behaviour is the key to developing effective and practical by-catch reduction devices.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
We obtained juvenile Japanese flounder (99–128 mm total length; 114 mm on average; s.d. 7 mm) artificially hatched at the Chiba Prefecture Tokyo Bay Sea Farming Center. These fish were kept in the holding tank for more than 1 month and fed dry pellets.
A schematic diagram of the testing apparatus is given in Figure 2. A fibreglass tank (5 m long, 1 m wide, and 1 m deep) was set up indoors and filled with seawater to a depth of 40 cm over packed beach sand. Seawater was circulated and filtered at all times except during experiments. Grid models were towed in the tank from one end to the other by winding an electric winch at one end of the tank. Grid models consisted of a steel frame, polyethylene netting (25 mm), and a grid (see Figure 2). Polyethylene netting similar to conventional codend material was put on the steel frame (40 x 50 cm) to model the bottom part of the codend. Two grid models were tested, one with horizontal bars and one with vertical bars. The grids were attached at a 60° attack angle against the towed direction. Dimensions and bar intervals of both grid models were the same: 50 x 20 cm with a 10.1-mm mean interval (s.d. 0.3 mm). We attached a waterproof CCD camera (Kowa Corp. Marine Eye, 
5 cm) to the side netting panel of the model to observe and record fish reaction to the grid on digital videotape. The field of view of the camera covered almost the entire area of the grid.
|
A plastic mesh pipe (
20 cm) was placed on the central part of the netting and three juvenile flounder were released in the pipe for each test. Fish were used only once each. The pipe was removed after flounder were settled on the netting, and the cable was quickly rewound at a constant speed (1.0 m s–1). We carried out a series of tow experiments with two different illumination levels: light conditions (75–135 lux at water surface) and dark conditions (0.02–0.05 lux at water surface). Visual observation beside the tank was also conducted for the series of experiments in dark conditions, since illumination level was lower than the minimum illumination requirement for the CCD camera (0.3 lux). |
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
In all, 225 individuals were tested during the series of experiments (75 for horizontal/light, 45 for horizontal/dark, 60 for vertical/light, and 45 for vertical/dark). Reaction behaviours were categorized within four patterns (Figure 3) by playing videotapes frame-by-frame and/or by visual observation: (i) forward swimming in towed direction; (ii) swimming over the grid; (iii) sticking on the grid; and (iv) passing through the grid. Most fish (212 individuals, 94% of tested fish) showed just one pattern during a trial, since towing duration was approximately only 4 s. However, the remaining fish (13 individuals, 6%) exhibited multiple behaviours, such as "swam over the grid after forward swimming". In these cases, only the final behaviour was considered. Frequencies of each reaction pattern were significantly different by grid types and illumination levels (p < 0.01, chi-square test, n = 225) (Figure 4). The most dominant reaction pattern was forward swimming for all experimental conditions, but its ratio was higher for the vertical bars vs. the horizontal bars, and higher for the light condition compared to the dark condition.
|
|
A high proportion of fish passed through the grid bars with the grid and horizontal bars in place. During testing of the grid with horizontal bars, approximately 40% of tested fish (31 individuals) passed through the grid in the light conditions; approximately 30% of fish (13 individuals) passed through in the dark conditions. Only one individual passed through the grid with vertical bars in light conditions and none in dark conditions. In light conditions, 25 individuals completed the sequence of encounter, penetration, and escape through the grid with horizontal bars. Out of these fish, 20 individuals (80%) penetrated the gaps head first. The rest of the fish (five individuals) first stuck belly-first on the grid and struggled, then finally penetrated their heads into the gaps.
The "sticking on the grid" and "swimming over the grid" reaction patterns occurred most frequently with vertical bars – in dark conditions the former more so than the latter.
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The high frequency of the "forward swimming" behaviour pattern over all experimental conditions suggests that flounder visually recognized and reacted to the approaching grid. High frequencies of "swimming over the grid" in light conditions and "sticking on the grid" in dark conditions also suggest that the fish reacted visually to the approaching grid.
Our results showed that juvenile flounder, which have a horizontally compressed body shape, passed through the grid with horizontal bars and only one fish penetrated the vertical grid. On the other hand, one study on fish penetration behaviour for different shapes and colours of meshes and twines demonstrated that mackerel (Scomber scombrus) and haddock (Melanogrammus aeglefinus), which have laterally compressed body shapes, passed most frequently through the lowest contrast vertical twines both voluntarily and when conditioned. Glass et al. (1993) concluded that the stronger visual stimulus by horizontally orientated twines elicited an avoidance response against twines. Therefore, the likelihood of fish passing through the grid probably relates to bar (twine) orientations and body shapes.
We observed that 13 individuals out of 41 categorized as "forward swimming" before the horizontal grid kept swimming in the direction of the tow even though their tails or bodies had passed partly between the bars. On the other hand, we found that most fish penetrated into bar gaps and passed through the grid head first. We conclude from these observations that penetration of flounders through bar gaps was governed by voluntary actions.
A fishing trial using the grid separator with vertical bars was carried out at night-time in another beam trawl fishery in Japan (Tokai et al., 1996). Approximately 70% of frog flounder (Pleuronichthys cornutus) passed through a grid when the ratio of body height to bar distance was 0.5, and all frog flounder were retained when the ratio of body height to bar distance was more than 1.0. This suggests that selection by the grid was mostly governed by a mechanical property of the grid. On the other hand, no fish in dark conditions and only one fish in light conditions passed through the grid with vertical bars in our experiment, which is quite different from what might have been expected judging from the results of Tokai et al. (1996), since body heights of flounder ranged from 5 mm to 7 mm (5.8 mm on average, s.d. 0.6 mm) against a bar interval of 10.1 mm. This difference in grid selection was, of course, because of differences in experimental condition and gear configuration, but also because of difference in fish behaviour and partial simulation of the whole capture process in beam trawl fishing. Frog flounder and Japanese flounder may have different behaviour characteristics because they have different feeding habits; frog flounder feed on benthos (Minami, 1982), whereas Japanese flounder are piscivorous (Yasunaga, 1988). In addition, we used reared flounder that had not been allowed to acclimate to the tank conditions prior to the experiment, while fish are adapted to their habitat in actual fishing conditions. These differences may affect grid selection. In addition, we simulated the process of the grid approach only once for each fish. Most fish stimulated by a grid into "forward swimming" in a trawlnet will encounter the grid when exhausted. Isaksen et al. (1992) observed fish behaviour around a Nordmore grid and reported that fish swam in front of the funnel for some time before becoming exhausted and pinned against the grid. Likewise, we believe that when the model is towed over a long period the fish will be exhausted, and will show three other behaviour patterns. Thus, the "forward swimming" behaviour seen in this study may be neglected when considering the real capture process. Ratios of flounders passing through the grid with horizontal bars were estimated as 66% in light conditions and 40% in dark conditions if consistent ratios of reaction patterns were assumed throughout multiple encounters to the grid.
Applying our results to the beam trawl fishery in Tokyo Bay, the grid design with horizontal bars is more likely to enhance the exclusion of small flounders. However, the following tank experiment conditions encourage flounder to penetrate the grid and contribute to differences between tank trials and fishing trials: (i) all fish encounter the grid; (ii) no other catch obstructs the grid opening; and (iii) visibility is higher than in fishing grounds. We believe that fishing gear modifications can overcome the first two differences and improve correlation between tank and fishing grounds. The introduction of funnel netting and larger grids to improve the frequency of fish encounters with the grid, along with a large mesh separator panel to intercept large catches which may accumulate on the grid, are possible modifications to improve grid selection.
|
Acknowledgements |
|---|
We are grateful to Naoshi Makino and Hisanori Mita of Chiba Prefecture Tokyo Bay Sea Farming Center for providing juvenile flounders for the experiment, and Michael Pol of Massachusetts Division of Marine Fisheries for his valuable comments on the first draft.
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Glass C.W., Wardle C.S., Gosden S.J. (1993) Behavioral studies of the principles underlying mesh penetration by fish. ICES Marine Science Symposia 196:92–97.
Isaksen B., Valdemarsen J.W., Larsen R.B., Karlsen L. (1992) Reduction of fish by-catch in shrimp trawl using a rigid separator grid in the aft belly. Fisheries Research 13:335–352.[CrossRef][Web of Science]
Kawabe R., Nashimoto K., Hiraishi T., Naito Y., Sato K. (2003) A new device for monitoring the activity of freely swimming flatfish, Japanese flounder Paralichthys olivaceus. Fisheries Science 69:3–10.[CrossRef][Web of Science]
Kennelly S. J. and Broadhurst M. K. (1995) Fishermen and scientists solving bycatch problems: examples from Australia and possibilities for the Northeastern United States. Solving Bycatch: Considerations for Today and Tomorrow pp. 121–128 Alaska Sea Grant College Program Report No. 96-03, University of Alaska, Fairbanks. 322 pp.
Matsushita Y., Inoue Y., Shida M., Nojima Y. (1998) Fish behaviour in the codend of small trawl. Abstracts for the Meeting of the Japanese Society of Fisheries Science, 23–27 September 1998(Japanese Society of Fisheries Science, Tokyo) 212 pp.
Minami T. (1982) The early life history of a flounder Pleuronichthys cornutus. Bulletin of the Japanese Society of Scientific Fisheries 48:369–374.[Web of Science]
Rose C. and Gauvin J.R. (2000) Effectiveness of a rigid grate for excluding Pacific halibut, Hippoglossus stenolepis, from groundfish trawl catches. Marine Fisheries Review 62:61–66.
Tokai T., Omoto S., Sato R., Matuda K. (1996) A method of determining selectivity curve of separator grid. Fisheries Research 27:51–60.[CrossRef][Web of Science]
Yasunaga Y. (1988) Studies of physiology and ecology of larvae and juveniles of plaice Paralichthys olivaceus. Bulletin of National Research Institute of Fisheries Engineering 9:9–164.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||