Does fishing dismantle fish culture and ecosystem structure? Questions about the implications of social learning among fish and fishers

Scientific awareness of social learning, especially among vertebrates, has expanded rapidly in recent decades. That literature suggests that social learning may be a second adaptive mechanism that interacts with and refines genetic adaptation. For an individual fish, learning from others reduces the costs of acquiring experience-based behaviours and minimizes the hazards that arise from imperfect knowledge of local regularities. For a group of fish, social learning facilitates the evolution of time and place behaviours that work in its locality. It spreads those behaviours within the group and to subsequent generations. Thus, social learning enables persistent adaptation at a finer scale than might be possible through genetic processes alone. Strong evidence of genetic differentiation at less than a panmictic scale and persistent local depletions suggests regular, fine-scale system structure. Social learning may play an important role in creating and maintaining this finer-scale structure. Fishers' learned adaptations to the market and natural system usually lead them to target larger/older fish and fish aggregations at familiar times and places. However, older fish are likely to be the principal repository of the time-and-place experience required for local growth, survival, and reproduction, while social aggregations are important schools in which younger fish acquire the experience of older fish. Consequently, if adaptation through social learning is important among fish, there is reason to be concerned that heavy fishing of social learners reduces their abundance, as usually assumed, and impairs the inheritance of the socially learned experience required for persistent local adaptation.     

School for learning: sharing and transmission of feeding information in jack mackerel <Emphasis Type="Italic">Trachurus japonicus</Emphasis> juveniles

We aimed to evaluate the effect of social interaction on learning in juvenile jack mackerel Trachurus japonicus. We first compared the learning process between single fish and groups of fish. Reward-conditioned learning was established in eight trials in both treatments, whereas individuals in the group treatment responded to stimuli more frequently in the feeding area than in the single fish. This implies that information about the feeding area was shared in the group and pursuing other individuals gave them a behavioral advantage for feeding. We then investigated whether information on the feeding area can be transmitted through observation of other individuals in aligned tanks. Fish in the control group required six trials to be conditioned to aeration stimuli and feeding location, whereas those in the observation treatment required only three trials for this learning. This result implies that information on the feeding area was transmitted through visual observations. The present research suggests that sharing and transmission of information occur in schools of jack mackerel. Schooling behavior would thus enable optimization of the foraging behavior in this species.     

Largemouth bass use prior experience,but not information from experienced conspecifics,to avoid capture by anglers

Angler satisfaction is based on the presence of catchable fish, and therefore, understanding the mechanisms driving catchability is important for fisheries biologists and managers. Lure avoidance learning (including via social learning) may be a driver of catchability, but these mechanisms have not been explored in depth. To address this, largemouth bass, Micropterus salmoides (Lacépède), were stocked into four ponds, and two were angled using a green plastic worm. Naïve bass were then introduced into all ponds, and angling commenced using a sequence of three lures (green plastic worm, white plastic worm and in‐line spinnerbait). Naive bass stocked alongside experienced conspecifics were not less catchable, indicating socially learned lure avoidance did not take place. Catch rates declined across sessions until the switch to the spinnerbait, when rates temporarily spiked. This suggests that the similarity of a novel lure to previously experienced lures may determine if a fish avoids angling capture.     

Ontogenetic changes of learning capability under reward conditioning in striped knifejaw Oplegnathus fasciatus juveniles

ABSTRACT:   To elucidate the ontogenetic changes in learning capability of the striped knifejaw Oplegnathus fasciatus , which is a common rocky-shore member in Japan, 28 individuals (2–10 cm standard length [SL]) of laboratory-reared fish under reward conditioning were trained. A single fish was first conditioned to go to one end of two arms of a Y-maze by being given a reward (food pellets), and once the fish has cleared the original problem, the rewarded side was reversed. All fish cleared at least the original problem, although fish at 2–4 cm SL showed lower scores because they learned more slowly or did not clear the reversal problem. Fish showed better learning capability as they grew. Fish of ca. 7 cm SL showed the highest score with quick learning of both the original and reversal problems. Their learning curve (relationship between SL and individual score) was approximated by a quadratic curve with the maximum score at approximately 7 cm SL. It was speculated that fish at its recruitment stage requires a high learning capability to adapt to the new environment of the rocky shore, which is composed of more complex microhabitats compared to their previous pelagic habitat.     

Learned predator recognition and antipredator responses in fishes

Early research into the causes of geographical variation in antipredator behaviour in fishes revealed that population differences have an underlying genetic basis. However, evidence from a variety of fish species suggests that learning plays an important role in the development of antipredator responses. Here, we consider the opportunities for learning during the three main stages of a predator–prey interaction: detection, recognition and assessment, and attack avoidance. Much of the evidence for learning is based on the recognition and assessment stage of the predator–prey interaction, but this may reflect methodological biases. We also examine the relative roles of different sensory cues, in particular, vision and olfaction, and the importance of individual vs. social learning. We provide evidence that visual predator recognition skills are largely built on unlearned predispositions, whereas olfactory recognition typically involves experience with conspecific alarm cues. Populations display variation in their propensity to learn, and we predict that ecological factors are likely to mediate the balance between individual and social learning.     

Demand feeding and welfare in farmed fish

Following the development of demand-feeding systems, many experiments have been conducted to explore feeding motivation and feed intake in farmed fish. This work aims to review a selection of studies in the field, focusing on three key factors, related to demand feeding and fish welfare. Firstly, we outline how demand feeders should be considered when developing feed management strategies for improving welfare in production conditions. Secondly, via laboratory demand-feeding experiments, we show self-feeding activities depend not only on feeding motivation and social organisation, but also on individual learning capacity and risk-taking behaviour. Thirdly, we report encouraging results demonstrating that when presented with two or more self-feeders containing complementary foods, fish select a diet according to their specific nutritional requirements, suggesting that demand feeders could be used to improve welfare by allowing fish to meet their nutritional needs.     

Role des individus conditionnes dans l'initiation a l'auto-nourrissage et dans l'adaptation a la captivite du loup,Dicentrarchus labrax, 0+ sauvage,en periode hivernale

It is shown that the presence of pre-conditioned fish in a wild naive group reared under limiting ecological conditions (winter) facilitates the adaptation of the group to captivity by accelerating the learning of self-feeding and minimizing rearing stress and mortality. The reasons for this phenomenon, which has implications for aquaculture and fishing, are discussed and related to the learning capacity of this species when imitating pre-conditioned fish.     

Habituation and conditioning in gilthead sea bream (Sparus aurata): Effects of aversive stimuli,reward and social hierarchies

To tailor the farming environment to a fish species, we should understand the species‐specific responses to stimuli, including the degree of adaption and learning. Groups of gilthead sea bream were given a delay Pavlovian conditioning regime using a conditioning stimulus (CS) of light flashes signalling arrival of food. Controls were exposed to light flashes unrelated to feeding. Fish in both treatments showed an initial fear response of moving away from the CS combined with reduced swimming speed. In subsequent trials, the Control fish largely habituated the fleeing response but sustained to respond by reducing the swimming speed. The Conditioning fish also stopped to escape from the CS, but opposed to the Control group they gradually increased their swimming speed in response to the CS. In addition, the number of fish in the feeding/CS area increased and became similar to basal level after around 16 trials. A small and variable proportion of the fish displayed black vertical bands on their body and territorial behaviour, and a social hierarchy could interfere with the processes of habituation and conditioning. The swimming speed of the fish increased with number of dark individuals, but this was not found during the CS and the light stimulus thus seemed to overrule the effect of territorial behaviour. The persistent negative response to light flashes in the Control suggests that fish seemingly adapted to repetitive stressors are still in a state of alertness. The change in the response to light shows the potential for rewarding aversive stimuli to reduce stress.     

Using fish models to investigate the links between microbiome and social behaviour: The next step for translational microbiome research?

Recent research has revealed surprisingly important connections between animals’ microbiome and social behaviour. Social interactions can affect the composition and function of the microbiome; conversely, the microbiome affects social communication by influencing the hosts’ central nervous system and peripheral chemical communication. These discoveries set the stage for novel research focusing on the evolution and physiology of animal social behaviour in relation to microbial transmission strategies. Here, we discuss the emerging roles of teleost fish models and their potential for advancing research fields, linked to sociality and microbial regulation. We argue that fish models, such as the zebrafish (Danio rerio, Cyprinidae), sticklebacks (?Gasterosteidae), guppies (Poeciliidae) and cleaner–client dyads (e.g., obligate cleaner fish from the Labridae and Gobiidae families and their visiting clientele), will provide valuable insights into the roles of microbiome in shaping social behaviour and vice versa, while also being of direct relevance to the food and ornamental fish trades. The diversity of fish behaviour warrants more interdisciplinary research, including microbiome studies, which should have a strong ecological (field‐derived) approach, together with laboratory‐based cognitive and neurobiological experimentation. The implications of such integrated approaches may be of translational relevance, opening new avenues for future investigation using fish models.     

When fish shoals meet: outcomes for evolution and fisheries

The mechanisms underlying the social structure of free‐ranging fish shoals have received little attention in comparison to functional studies on shoaling. Recently, however, a number of investigations, both in the marine and in the freshwater environment, have begun to address the underlying mechanisms by concentrating on interactions between free‐ranging shoals. The rates of shoal encounters can influence the opportunities for individual assortment by phenotype and selection of shoal size; act as a constraint on the observed patterns of shoal structure by restricting individual choice behaviour and were found to be high in several small freshwater species (where intershoal distances were small), but lower in marine species where shoals were more dispersed. The duration of encounters may play a role in that it affects both the time available for assessment (of the encountered shoal) as well as that for exchange of individuals. Scarce published information on the outcome of shoal encounters suggests that the outcome of shoal encounters is influenced by shoal composition but not by shoal size. Individual behaviour may have evolved to maintain shoal size when shoals encounter, but when fish populations are depleted by fishing, this trait can exacerbate range and stock collapse. Furthermore, an understanding of the dynamics of shoal encounters has important consequences for the evolution of reciprocal altruism and the transmission of information through social learning within populations. Finally, information on encounter rates between shoals and the number of individuals that are exchanged on such occasions could be important for making predictions about the spread of disease through fish populations.     

Nondestructive evaluation of fish meat using ultrasound signals and machine learning methods

In this study, we propose a new method for the nondestructive measurement of the fat content and texture of fish meat using machine learning and the bag of features approach. We employed two machine learning methods, that is, a self-organizing map (SOM) and radial basis function (RBF) network. The SOM was applied to symbolize the pattern of the frequency spectrum extracted from ultrasound signals and to generate key features for the bag of features technique. The RBF network was applied to estimate the fat content and texture of fish meat from the bag of features histogram. We verified the accuracy of the fat content and texture estimations given by the proposed method through a series of experiments. The results showed that the fat content and texture of fish meat was estimated more accurately using the proposed method than by the conventional approach.     

Effect of fishing effort on catch rate and catchability of largemouth bass in small impoundments

Largemouth bass Micropterus salmoides (Lacepède) catch rates decline with sustained fishing effort, even without harvest. It is unclear why declines in catch rate occur, and little research has been directed at how to improve catch rate. Learning has been proposed as a reason for declining catch rate, but has never been tested on largemouth bass. If catch rate declines because fish learn to avoid lures, periods of no fishing could be a management tool for increasing catch rate. In this study, six small impoundments with established fish populations were fished for two May to October fishing seasons to evaluate the effect of fishing effort on catch rate. Closed seasons were implemented to test whether a 2‐month period of no fishing improved catch rates and to determine whether conditioning from factors other than being captured reduced catch rate. Mixed‐model analysis indicated catch rate and catchability declined throughout the fishing season. Catch rate and catchability increased after a 2‐month closure but soon declined to the lowest levels of the fishing season. These changes in catch rate and catchability support the conclusion of learned angler avoidance, but sustained catchability of fish not previously caught does not support that associative or social learning affected catchability.     

Socially induced stress and behavioural inhibition in response to angling exposure in rainbow trout

It is well known that fish can learn to avoid angling gear after experiencing a catch‐and‐release event, that is, after a private hooking experience. However, the possible importance of social information cues and their influence on an individual's vulnerability to angling remains largely unexplored, that is, social experience of a conspecific capture. The effects of private and social experience of hooking on the stress response of fish and subsequent catch rates were examined. Hatchery‐reared rainbow trout, Oncorhynchus mykiss (Walbaum), were implanted with heart rate loggers and experimentally subjected to private or social experience of hooking. Private and social experience of angling induced an increased heart rate in fish compared with naïve control fish. While private experience of hooking explained most of the reduced vulnerability to capture, no clear evidence was found that social experience of hooking affected angling vulnerability in fish that had never been hooked before. While both private and social experiences of angling constitute significant physiological stressors for rainbow trout, only the private experience reduces an individual's vulnerability to angling and in turn affecting population‐level catchability.     

The effect of social stress on the Standard Metabolic Rate (SMR) of brown trout, Salmo trutta

The effect of social stress, induced by confinement in pairs, on the SMR of the brown trout, Salmo trutta (L.), was investigated. Fish were confined in pairs under laboratory conditions and allowed to establish social hierarchies, with one fish becoming dominant and the other subordinate. The change in SMR of the subordinate fish was significantly greater than that of their respective dominant. Also, the more aggressive the dominant behaved towards the subordinate with which it was paired, the greater the increase in the SMR of the subordinate fish appeared to be. It is concluded that social stress causes an increase in SMR in subordinate fish and therefore imposes a metabolic disadvantage.     

Attending to spatial social–ecological sensitivities to improve trade‐off analysis in natural resource management

Balancing trade‐offs amongst social–ecological objectives is a central aim of natural resource management. However, objectives and resources often have spatial dimensions, which are usually ignored in trade‐off analyses. We examine how simultaneously integrating social–ecological benefits and their spatial complexities can improve trade‐off analysis. We use Pacific herring (Clupea pallasii, Clupeidae)—an ecologically important forage fish with social, cultural and economic value to communities and commercial fisheries—as a case study. By combining spatial management strategy evaluation with social benefits analysis, we illustrate when policies aimed at aggregate stocks versus spatially segregated substocks of fish fail to balance trade‐offs amongst social–ecological objectives. Spatial measures (e.g. area‐based closures) may achieve some objectives but produce alternative trade‐offs that are sensitive to assumptions about fish population dynamics and social complexities. Our analyses identify policies that are inefficient (e.g. yielding economic costs without producing social or ecological gains), highlight management strategies that generate trade‐offs and indicate when costs are distributed unequally for different user groups. We also point to strategies with outcomes that are robust to spatial uncertainties and reveal research priorities by identifying which performance metrics exhibit sensitivity to spatial ecological assumptions. Collectively, our analyses demonstrate how incorporating social objectives and spatial dynamics into management strategy evaluation can reveal trade‐offs and the implications of management decisions.     

Cichlasoma dimerus responds to refeeding with a partial compensatory growth associated with an increment of the feed conversion efficiency and a rapid recovery of GH/IGFs axis

Many fish species display compensatory growth (CG), a phenomenon by which fasted fish grow faster during refeeding. However, most studies use a group‐housed fish approach that could be problematic in social fish when interaction between individuals is not considered or eliminated. Additionally, the growth hormone (GH)/insulin‐like growth factors’ (IGF‐1 and IGF‐2) axis is implicated in postnatal growth in vertebrates, but its relevance in CG is not fully understood. Thus, the aim of this work was to determine whether CG occurs in a social fish, Cichlasoma dimerus, using an individually held fish approach and secondly, to evaluate the GH/IGFs expression profile during refeeding by 3 days and 3 weeks. C. dimerus showed partial CG. The feed conversion efficiency (FCE) was higher in three‐day‐refed fish, which presented higher GH plasma and mRNA levels than controls but shown no differences in liver and muscle GH receptors (GHR1 and GHR2) and IGFs mRNA levels. Surprisingly, three‐week‐refed fish exhibited GHR1 and IGF‐2 increments, but a reduction in GHR2 expression in muscle. These results show a strong association between GH levels, growth rate and FCE during refeeding, and a long‐lasting effect of refeeding on muscular expression of GHRs and IGF‐2.     

Effects of size grading on growth and survival of juvenile turbot at two temperatures

Juvenile turbot were size graded into three size groups (mean initial size): Small (3.4 g), medium (7.0 g) and large (10.5 g), and additional fish were held in ungraded (6.6 g) groups. Subgroups (n = 396) of fish were tagged and individual growth rates and social interactions within different size categories were studied in fish reared at 13 and 19 °C. Size grading of juvenile turbot did not improve growth. Specific growth rates (SGR) were mainly size-related, and no differences in SGR or mortality between the experimental groups at both tem-peratures were found. A higher level of social interactions was indicated amongst medium-sized fish than amongst those in the smallest and largest categories. Excess feeding may have been important factors in reducing aggression, so that the growth of the smallest individuals was not suppressed by the larger individuals in the present study. Grading seems to be an unnecessary operation to improve the growth and survival of juvenile turbot (5- 150 g). However, as it was mainly the smallest individuals in each group that died, grading of very small turbot (2-5 g) can be recommended.     

Social networks in elasmobranchs and teleost fishes

Over the last decade, there has been an exponential increase in studies using social network analysis to describe the structure of animal societies. In this synthesis, we examine the contribution of social network analysis towards developing our understanding of the social organization of elasmobranchs and teleost fishes. We review and discuss the current state of knowledge of the mechanisms and functions underpinning social network structure in fishes with particular emphasis on cooperation, familiarity, site fidelity, population structure and the welfare of captive populations. We also discuss important methodological issues (e.g. how to identify and mark fish) and highlight new developments in this area of research and their implications for the study of fish behaviour. Finally, we outline promising future research areas for the application of social network analysis to teleost fishes and elasmobranchs.     

Consumer interest in social sustainability issues of whitefish from capture fisheries in the north‐east Atlantic

Capture fisheries in the north‐east Atlantic account for approximately 10% of all fish consumed from capture fisheries globally. The literature shows that consumers show considerable interest in social sustainability of products in general and of fish specifically. This interest, however, has not yet been investigated for fish from the north‐east Atlantic. The first objective of this study, therefore, was to investigate whether consumers are interested in social sustainability issues of whitefish from the north‐east Atlantic by determining preferences for four social sustainability issues with distinct benefits in relation to a known reference point (i.e. approach to overfishing). The second objective of this study was to determine to what degree case‐specific and general psychographic consumer characteristics explain preferences for these issues. Choice modelling results from an online survey among 457 Dutch consumers show that consumers have the strongest preference for the environmental sustainability issue approach to overfishing. In addition, results on the social sustainability issues showed that consumers prefer fish welfare over product quality, worker safety and local employment, indicating that in this case, consumers place animal benefits over personal, worker and community benefits. The case‐specific psychographic characteristic concern contributed most to explaining preferences for the environmental sustainability issue, whereas the general psychographic characteristic personal relevance contributed most to explaining preferences for the social sustainability issues. This result is likely explained by the principal focus on MSC certification in markets for fish products, which caused consumers to form opinions (e.g. concern) on overfishing, but not on social sustainability.     

Development of fish spatio-temporal identifying technology using SegNet in aquaculture net cages

In marine aquaculture, fish populations constantly decrease throughout the cultivation period because of mortality and escape. Current production management systems provide limited opportunities to count the cultured fish, making it difficult to estimate accurately the fish population in the cage. To overcome this problem, an automatic fish identifying method based on particle tracking velocimetry (PTV) flow visualization technology is proposed in this paper. The proposed method utilizes an image processing unit that extracts individual fish from the acquired image and a motion analysis unit that calculates the motion vector for each individual. Thus, the accuracy of the extraction results in the image processing unit affects the system’s counting results. To validate the efficiency and robustness of the image extraction performed by the image processing unit, individuals were extracted from images using the open-source image deep learning semantic segmentation method (SegNet), which is able to distinguish between the background and foreground in the images via analysis at the pixel level. SegNet is able to improve the image discrimination performance by multiplying the learning paths, and the robustness of the detection results can be ensured by changing the layer structure according to the detection target. Accordingly, the use of SegNet was evaluated in terms of the number of layers and images in the training set. The results of this study indicate that the application of SegNet with PTV technology represents a promising method for the automatic identifying and behavioral tracking of fish in an aquaculture net cage.