- Email: [email protected]
Development of observational learning during school formation in jack mackerel Trachurus japonicus juveniles
Behavioural Processes 103 (2014) 52–57
Contents lists available at ScienceDirect
Behavioural Processes journal homepage: www.elsevier.com/locate/behavproc
Development of observational learning during school formation in jack mackerel Trachurus japonicus juveniles Kohji Takahashi ∗ , Reiji Masuda, Yoh Yamashita Maizuru Fisheries Research Station, Kyoto University, Nagahama, Maizuru, Kyoto 625-0086, Japan
a r t i c l e
i n f o
Article history: Received 19 April 2013 Received in revised form 11 October 2013 Accepted 27 October 2013 Available online 9 November 2013 Keywords: Behavioural ontogeny Carangidae Conditioning Life history strategy Social learning
a b s t r a c t We assessed whether the development of observational learning in jack mackerel Trachurus japonicus juveniles corresponds with that of their schooling behaviour. Schooling behaviour was quantitatively analyzed by nearest neighbour distance and separation angle in two size classes of fish, 20-mm and 40-mm in body length. Observer and non-observer fish with matching sizes were conditioned to pellets by temporarily stopping aeration. Observer fish were provided with five observation trials of other individuals feeding near an air stone when aeration was stopped. After the observation trial, fish were conditioned to pellets with the stop of aeration, and then the learning process was evaluated by the increase in the association with the feeding area when aeration was stopped. In 20-mm fish, which were at an immature stage of schooling behaviour, there was no difference in the learning process between observer and nonobserver fish. In contrast, 40-mm fish were confirmed to have a well-developed schooling behaviour, and the observer learnt the feeding area more efficiently than the non-observer. This study provides evidence that observational learning develops along with the development of the social interaction. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Marine fish generally experience a major habitat shift in their life histories. Habitat shifts increase the encounter rate of novel experiences such as previously unencountered prey items and predators. Thus, rapid adaptation to novel stimuli would increase the chance of survival in migrating juveniles. Learning would help juveniles to acquire appropriate responses to the novel environment through their own experiences; therefore, learning capability can be one of the most important abilities for adapting to habitat shifts. Our previous studies investigated the relationship between the ontogenetic change of learning capability and habitat shift in marine fish using jack mackerel Trachurus japonicus juveniles (Takahashi et al., 2010, 2012a). The spatial learning capability of jack mackerel juveniles increases dramatically at ca. 50 mm in body length (BL), corresponding with the migration from offshore to coastal waters (Takahashi et al., 2010). Moreover, Takahashi et al. (2012a) compared the learning capability conditioning to surface or mid-water structures between 40-mm and 60-mm BL juveniles. The 40-mm fish were conditioned to the surface structure more rapidly than the mid-water structure, whereas the 60-mm fish learned the mid-water structure easier than the surface one. These studies showed the qualitative change in learning capability during
∗ Corresponding author. Tel.: +81 773 62 5512; fax: +81 773 62 5513. E-mail address: [email protected] (K. Takahashi). 0376-6357/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.beproc.2013.10.012
their migration from an offshore pelagic to a coastal neritic life, implying that jack mackerel juveniles have an ontogenetic change in learning capability as part of their life history strategy during their habitat shift. On the other hand, about 50% of teleost fishes in the world form schools at least for a part of their life history (Shaw, 1978). Many fishes do not exhibit social behaviour at an early life stage, but then begin to develop schooling with growth (e.g. Masuda & Tsukamoto, 1999; Ishizaki et al., 2001; Masuda et al., 2003; Fukuda et al., 2010; Masuda, 2011); i.e. they experience a shift in social environment. In a school of fish, each individual is likely to utilize social learning (Brown & Laland, 2011). For example, prey location can be learnt through observations of associating shoal mates to feeding sites without the energetic expenditures of food searching (Brown et al., 2003). The anti-predator behaviour would be enhanced through observations of conspecifics being attacked without risk to the individual (Arai et al., 2007). Researchers have found that fish acquire survival skills by social learning in various life history contexts, such as predator avoidance (Brown & Laland, 2001; Kelly et al., 2003), orientation behaviour (Warner, 1988; Fukumori et al., 2010), feeding information (Reader et al., 2003; Webster & Laland, 2008), and mate selection (Witte & Nobel, 2011). Jack mackerel form large schools after their recruitment to rocky reef habitat (Masuda, 2008; Masuda et al., 2008), and therefore they have opportunities to acquire information from conspecifics when they are schooling. In our previous study, jack mackerel juveniles at the schooling stage acquired information through observational
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57
53
Fig. 1. Schematic drawings of the tank arrangement in experiment 2: (a) demonstrator, (b) observer, and (c) non-observer fish. The tanks were placed side by side, and a black board was put between the demonstrator and observer tank except during the observation period.
learning (Takahashi et al., 2012b); i.e. in jack mackerel at 50 mm BL, fish that observed other individuals feeding at aeration in an adjacent tank were conditioned to aeration as a stimulus to initiate feeding more quickly than fish that did not observe this conspecific behaviour. Jack mackerel juveniles have a shift in social interaction at the early juvenile stage. Hatchery-reared jack mackerel juveniles begin to form schools at ca. 12 mm BL, and their schooling behaviour gradually develops (Masuda et al., unpub.). Observational learning would be more useful for their acquisition of feeding information in social environments than learning in isolation, and therefore may well be related to the development of schooling behaviour. The present study compared the observational learning capability of two size groups of jack mackerel juveniles corresponding with the developmental stage of the social environment.
2. Materials and methods 2.1. Fish Fertilized jack mackerel eggs were obtained from the spawning of wild and local origin broodstocks maintained at the Maizuru Fisheries Research Station (MFRS) of Kyoto University after the injection of human chorionic gonadotropin (0.6 IU/g body weight). A single batch of eggs was stocked in four 500-L transparent polycarbonate tanks supplied with filtered sea water at the MFRS as described in Takahashi et al. (2012a). The fish hatched on 1 June 2010. After hatching, larvae were provided with rotifers Brachionus plicatilis from the opening of the mouth to ca. 8 mm BL, Artemia sp. Nauplii from 6 to 12 mm BL, and dry pellets (N400 and N700, Kyowa Hakko Bio Co., Ltd., Tokyo, Japan from 10 to 30 mm BL; Otohime S1, Marubeni Nisshin Feed Co., Ltd., Tokyo, Japan from 30 to 40 mm BL) in accordance with growth. Each feeding was conducted twice a day in the morning and evening to satiation. The water temperature was maintained at around 25 ◦ C. Two size groups of jack mackerel were used to examine the schooling behaviour and observational learning as follows: mean ± standard deviation (SD) of the small group, defined as the 20-mm group, were 15 ± 1 mm BL and 18 ± 2 mm BL for schooling behaviour and observational learning experiments, respectively, and those of large juveniles, defined as the 40-mm group, were 38 ± 3 mm BL and 38 ± 3 mm BL, respectively. After the experiment, the fish were kept in the MFRS as broodstocks. All the experiments were performed according to the guidelines of the Regulation on Animal Experimentation at Kyoto University.
2.2. Experiment 1: schooling behaviour A large acrylic tank (length × width × height: 900 × 450 × 450 mm) was used for the evaluation of schooling behaviour. The tank was filled with sea water to a depth of 400 mm, and the water was changed every two trials. A white vinyl sheet was put under the tank to facilitate video recording, and the tank was separated from the observer by a black sheet. A video camera (HDR-CX 550; Sony, Tokyo, Japan) was installed above the experimental tank to record fish behaviour during the experiment. The schooling behaviour was compared between 20-mm and 40-mm fish juveniles. Experiments were conducted using groups consisting of four juveniles, and six groups were used for each size class. Fish were introduced into the experimental tank, and were allowed to acclimatize for 10 min. After acclimatization, fish behaviour was recorded with a video camera using a remote control to eliminate the effect of an observer. Upon completion of the experiment, fish body length was measured after the fish were anaesthetized using MS-222. Schooling behaviour was analyzed using video recordings at 5–6 min and 10–11 min after the start of recording. Nearestneighbour distance (NND) and separation angle (SA) were measured from video images for each individual. NND was calculated by dividing the measured value by the mean fish body size in a group. The measurement was conducted 20 times at 5-s intervals, and these data were averaged for each group. NND was compared between 20-mm and 40-mm fish by Welch’s ANOVA because data did not show homogeneity of variance. SA was compared between 20-mm and 40-mm fish by ANOVA. Furthermore, the SA of each size was compared to 90◦ , the value expected if movement was in a random direction, by one sample t-test. 2.3. Experiment 2: observational learning Three transparent glass tanks (length × width × height: 300 × 200 × 300 mm for 20-mm fish and 450 × 300 × 300 mm for 40-mm fish) were used in the experiment as the demonstrator, observer, and non-observer tanks (Fig. 1). Demonstrator and observer tanks were covered with black vinyl sheets except for on one side. These tanks were arranged to face each other on the uncovered sides of the tanks, and a removable black board was placed between these two tanks except during an observation trial. Both tanks were set in a water bath, and sea water filled the space between the observer and the demonstrator tanks. The non-observer tank was covered on all sides by black vinyl
54
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57
Fig. 2. Average indices of schooling behaviour for 20-mm (grey) and 40-mm (white) fish in experiment 1: (a) nearest-neighbour distance and (b) separation angle. Bars indicate standard deviation (n = 6).
sheets. Experimental tanks were provided with a depth of 200 mm circulating sea water. An air stone was set at the centre of each tank. Aeration was remotely controlled and was turned on except during the experiment. Tanks were separated by black sheets to reduce disturbance from the experimenter. A video camera set at 600 mm above one of the experimental tanks allowed the recording of the fish behaviour during the experiment. Demonstrator fish were conditioned to the stop of aeration as a conditioned stimulus with feeding pellets as an unconditioned stimulus. In the conditioning trials, 1 min after the stop of aeration, approximately 12 pellets (20-mm fish: N700, 40-mm fish: Otohime S1) subdivided into three portions were dropped near the air stone from above the tank. The demonstrators were trained until they showed a prominent response to aeration without pellets over 10 trials in a day. Two groups of four fish were used as demonstrators for each size of fish (average BL of demonstrator for 20 mm: 18 ± 3 mm, 40 mm: 35 ± 2 mm). The learning capability was compared between the observers and non-observers. Groups of four individuals were introduced into the observer and non-observer tanks on the previous day, and were allowed to acclimatize overnight. A few pellets were provided to confirm acclimatization prior to initiating the experiment for observer and non-observer fish, which was started if the fish ate these initial pellets. When the observer fish foraged the pellets, the black board between the observer and demonstrator tank was removed 30 min before the start of observation trials. Observers and demonstrators were attracted to each other when the black board was removed, indicating that they were visible to each other.
In the observation trial, observer fish were provided with a view of demonstrators aggregating in the feeding area when aeration was stopped, as had occurred in the conditioning trials of demonstrator. Since the demonstrators continued to aggregate in the feeding area when aeration was stopped, this suggested that the conditioned behaviour of the demonstrators was not biased by the presence of observers. Five observation trials were conducted at 30-min intervals. After five observation trials, the black board was replaced, and then 30 min later, conditioning trials were started. Observers and non-observers were conditioned in the same manner as demonstrators; i.e. 1 min after the stop of aeration, pellets were dropped near the air stone. The behaviour of observers and non-observers was video recorded for 2 min from 1 min before the stop of aeration to 1 min after. This process was defined as one trial, and 10 trials, separated by an interval of at least 30 min, were conducted from 12:00 to 18:00 with eight replications for each treatment and control. Fish BL was measured after anaesthesia using MS-222 when fish had accomplished the conditioning trials. The number of fish in the feeding area was used as an index of learning; the feeding area for the 20-mm and 40-mm juveniles was defined as the 50 × 50 mm and 100 × 100 mm areas surrounding the air stone, respectively. The number of fish in the feeding area was counted every 5 s during each 1-min period either with or without aeration in each trial. Aggregating behaviour to the feeding area for each trial was calculated as the aggregation index, by subtracting the number of fish in the feeding area with aeration from that without aeration. The aggregation indices during the first to tenth trials were compared by a two-way repeated measures
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57
55
Fig. 3. Average increase of the number of fish in the feeding area from the pre- to post-stimuli period in experiment 2: (a) 20-mm and (b) 40-mm fish. Symbols represent observer (䊉) and non-observer fish (). Bars indicate standard deviation (n = 8).
ANOVA with Bonferroni correction (˛ = 0.0083) between observation treatments (observer vs. non-observer fish in each size) or body sizes (20-mm fish vs. 40-mm fish in each treatment). 3. Results 3.1. Experiment 1: schooling behaviour The NND of 20-mm juveniles was 9.8 ± 4.3 BL (mean ± SD) and that of 40-mm juveniles was 2.0 ± 0.7 BL, the latter being significantly smaller than the former (Welch’s ANOVA: F1 , 5.21 = 20.8, p < 0.01; Fig. 2(a)). The SA of 20-mm juveniles was 77.9 ± 8.7◦ and that of 40-mm fish was 25.6 ± 16.1◦ , the latter being significantly smaller than the former (ANOVA: F1 , 11 = 6.65, p < 0.05; Fig. 2(b)); both values were significantly smaller than 90◦ (one sample t-test: 20-mm fish df = 5, t = −2.77, p < 0.05; 40-mm fish df = 5, t = −4.16, p < 0.01).
trials, F1 , 154 = 48.4, p < 0.001; treatments × trials, F1 , 154 = 0.00, p > 0.1; Fig. 3(a)). However, in 40-mm juveniles, there was a significant difference between the observer and non-observer treatment in the aggregation index from the first to the tenth trial (two-way repeated measures ANOVA with Bonferroni correction: treatments, F1 , 52.3 = 55.1, p < 0.001; trials, F1 , 154 = 34.2, p < 0.001; treatments × trials, F1 , 154 = 1.55, p > 0.1; Fig. 3(b)). Comparing the two size groups, the aggregation index of 40-mm juveniles was significantly higher than that of 20-mm fish in observer fish (two-way repeated measures ANOVA with Bonferroni correction: treatments, F1 , 8.25 = 29.6, p < 0.001; trials, F1 , 154 = 41.0, p < 0.001; treatments × trials, F1 , 154 = 0.66, p > 0.1), whereas there was no such difference in non-observer fish (two-way repeated measures ANOVA with Bonferroni correction: treatments, F1 , 5.94 = 3.59, p > 0.1; trials, F1 , 154 = 38.8, p < 0.001; treatments × trials, F1 , 154 = 0.31, p > 0.1). 4. Discussion
3.2. Experiment 2: observational learning There was no significant difference between observer and nonobserver fish in the aggregation index from the first to the tenth trial in the 20-mm juveniles (two-way repeated measures ANOVA with Bonferroni correction: treatments, F1 , 3.83 = 4.14, p > 0.1;
The 20-mm juveniles showed an SA slightly smaller than 90◦ , although they maintained their distance from each other. This result suggests that the 20-mm juveniles are at an immature stage of schooling behaviour. On the other hand, the 40-mm juveniles are considered to have reached the schooling stage, because they
56
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57
showed much closer NND and SA. Striped jack Pseudocaranx dentex juveniles, another carangid fish, start schooling behaviour at 12 mm total length (TL), whereas NND decrease with growth until 30 mm TL (Masuda & Tsukamoto, 1998). This finding supports the results of the present study; i.e. jack mackerel juveniles at 20 mm were still in the developing stage of schooling, while those at 40 mm BL had completed it. Aggregation to the feeding area was not different between the observation and control treatment for 20-mm fish, whereas the 40-mm fish showed an increase in aggregation in the feeding area through the observation trial of responding conspecifics. This implies that the observational learning capability of jack mackerel develops from 20 mm to 40 mm BL. Further, the present study suggests that the development of observational learning capability coincides with that of schooling behaviour. Fish without full development of schooling, i.e. 20-mm juveniles, would have fewer opportunities to observe other individuals, and therefore observational learning should be a low priority in their life. Indeed, jack mackerel at this stage are typically found associating with jellyfish and other floating objects either singly or in a small school (Masuda et al., 2008; Masuda, 2009). Meanwhile, 40-mm juveniles with fully developed schooling would have their life dependent on social behaviour. Various information on prey items and predators would be transmitted between individuals through observational learning in the social environment. Thus, the function of the observational learning capability would be to optimize the behaviour of jack mackerel juveniles at the stage of high dependence on schooling. Whereas previous studies suggested that the learning capability of juveniles changes in accordance with their environmental change from habitat shifts (Masuda & Ziemann, 2000; Makino et al., 2006; Hawkins et al., 2008; Takahashi et al., 2010, 2012a), the present study revealed that such an ontogenetic change also occurs with a change in social environment. There was no difference in the learning process in non-observer fish between 20- and 40-mm juveniles. This indicates that the ability to learn the location of the feeding area would not improve between these sizes, which is in agreement with the results of Takahashi et al. (2010) who showed that there was no difference in spatial learning capability measured by a Y-maze between 20-mm and 50-mm juveniles of jack mackerel. This suggests that something related to sociality must have happened during the juvenile stages to induce the changes in observational learning capability. There is a possibility that the change in attraction to other conspecific individuals induced an ontogenetic change in observational learning capability. Our study investigated the key factor of observational learning in striped jack juveniles, and striped jack juveniles could learn the feeding area to observe the foraging conspecific (Takahashi et al., 2013). The result suggested that attraction to conspecific fish is essential for observational learning. Attraction to conspecific fish would also be required for school formation behaviour, e.g. conspecific tracking. The change in the level of attraction may cause the development of observational learning capability and schooling behaviour. Whereas juveniles need basic locomotor and sensory organs for schooling (Hunter & Coyne, 1982; Kohno et al., 1984), the development of the central nervous system would also be necessary (Masuda & Tsukamoto, 1999). Similarly, there is the possibility that the central nervous system changes during the stage in which jack mackerel juveniles attain observational learning capability. Nakayama et al. (2007) investigated the ontogenetic change in schooling behaviour and social transmission using the chub mackerel Scomber japonicus. They showed that the social transmission between groups appeared at 30 days post-hatching, and that the volume of optic tectum dramatically increased prior to this behavioural change. Whereas the central nervous system related to observational learning has not been elucidated, the
developmental process of the nervous system in jack mackerel juveniles from 20 mm to 40 mm may provide a basis for studying the nervous system’s role in observational learning. In previous studies, a change in learning ability was observed in hatchery-reared fish (Masuda and Ziemann, 2000; Makino et al., 2006; Hawkins et al., 2008; Takahashi et al., 2012a); therefore, the timing of ontogenetic change is probably a genetically programmed life history strategy. On the other hand, the social environment of early life stages can have an impact on social behaviour in later stages (Moretz et al., 2007; Salvanes et al., 2007; Arnold & Taborsky, 2010); e.g. the observational learning capability of guppies Poecilia reticulata is influenced by the density of conspecifics in the early juvenile stage (Chapman et al., 2008). Therefore, it is likely that the early social environment is involved in the ontogenetic change of observational learning described in this study. In the present study, we were not able to evaluate which factor induced the ontogenetic change in observational learning because 40-mm fish had experienced a social environment in the rearing tank. The role of the social environmental effect at an early stage in the ontogenetic change in observational learning should be verified by further experiments, e.g. by isolating fish from conspecifics at early stages. In rodents, reliance on social learning differs between relatively social and asocial species or populations (Lupfer et al., 2003; Rymer et al., 2008). In golden hamsters Mesocricetus auratus, although adults in a solitary life did not influence one another’s food preferences, 4-week-old pups showed a significantly enhanced preference for foods that their mother was eating (Lupfer et al., 2003), suggesting that the social learning capability changes according to the shift in sociality. The present study revealed a more or less equivalent change in learning capability with the development of social interactions in conspecific fish. To our knowledge this is the first study to have shown the coincidence of the ontogeny of learning and that of social behaviour in fishes. Because of their abundant behavioural repertories (Brown et al., 2011) and relatively simple central nervous system, fishes are promising study organisms for the fundamental understanding of behavioural processes. Acknowledgements We thank Dr. Yuichi Fukunishi, Dr. Johan Bolhuis, and the reviewer whose comments substantially improved the quality of the manuscript. References Arai, T., Tominaga, O., Seikai, T., Masuda, R., 2007. Observational learning improves predator avoidance in hatchery-reared Japanese flounder Paralichthys olivaceus juveniles. Journal of Sea Research 58, 59–64. Arnold, C., Taborsky, B., 2010. Social experience in early ontogeny has lasting effects on social skills in cooperatively breeding cichlids. Animal Behaviour 79, 621–630. Brown, C., Laland, K.N., 2001. Social learning and life skills training for hatchery reared fish. Journal of Fish Biology 59, 471–493. Brown, C., Laland, K.N., 2011. Social learning in fishes. In: Brown, C., Laland, K.N., Krause, J. (Eds.), Fish Cognition and Behavior. , 2nd ed. Blackwell Publishing Ltd., UK, pp. 240–257. Brown, C., Laland, K.N., Krause, J., 2011. Fish Cognition and Behavior, 2nd ed. Blackwell Publishing Ltd., UK. Brown, C., Markula, A., Laland, K.N., 2003. Social learning of prey location in hatcheryreared Atlantic salmon. Journal of Fish Biology 63, 738–745. Chapman, B.B., Ward, A.J.W., Krause, J.K., 2008. Schooling and learning: early social environment predicts social learning ability in the guppy, Poecilia reticulata. Animal Behaviour 76, 923–929. Fukuda, H., Torisawa, S., Sawada, Y., Takagi, T., 2010. Ontogenetic changes in schooling behaviour during larval and early juvenile stages of Pacific bluefin tuna Thunnus orientalis. Journal of Fish Biology 76, 1841–1847. Fukumori, K., Okuda, N., Yamaoka, K., Yanagisawa, Y., 2010. Remarkable spatial memory in a migratory cardinalfish. Animal Cognition 13, 385–389. Hawkins, L.A., Magurran, A.E., Armstrong, J.D., 2008. Ontogenetic learning of predator recognition in hatchery-reared Atlantic salmon, Salmo salar. Animal Behaviour 75, 1663–1671.
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57 Hunter, J.R., Coyne, K.M., 1982. The onset of schooling in northern anchovy larvae, Engraulis mordax. CalcCOFI Reports 23, 246–251. Ishizaki, Y., Masuda, R., Uematsu, K., Shimizu, K., Arimoto, M., Takeuchi, T., 2001. The effect of dietary docosahexaenoic acid on schooling behaviour and brain development in larval yellowtail. Journal of Fish Biology 58, 1691–1703. Kelly, J.L., Evans, J.P., Ramnarine, I.W., Magurran, A.E., 2003. Back to school: can antipredator behaviour in guppies be enhanced through social learning? Animal Behaviour 65, 655–662. Kohno, H., Shimizu, M., Nose, Y., 1984. Morphological aspects of the development of swimming and feeding function in larval Scomber japonicus. Nippon Suisan Gakkaishi 50, 1125–1137. Lupfer, G., Frieman, J., Coonfield, D., 2003. Social transmission of flavor preferences in two species of Hamsters (Mesocricetus auratus and Phodopus campbelli). Journal of Comparative Psychology 117, 449–455. Makino, H., Masuda, R., Tanaka, M., 2006. Ontogenetic changes of learning capability under reward conditioning in striped knifejaw Oplegnathus fasciatus juveniles. Fisheries Science 72, 1177–1182. Masuda, R., Tsukamoto, K., 1998. The ontogeny of schooling behaviour in the striped jack. Journal of Fish Biology 52, 483–493. Masuda, R., Tsukamoto, K., 1999. School formation and concurrent developmental changes in carangid fish with reference to dietary conditions. Environmental Biology of Fishes 56, 243–252. Masuda, R., Ziemann, D.A., 2000. Ontogenetic changes of learning capability and stress recovery in Pacific threadfin juveniles. Journal of Fish Biology 56, 1239–1247. Masuda, R., Shoji, J., Nakayama, S., Anaka, M., 2003. Development of schooling behavior in Spanish mackerel Scomberomorus niphonius during early ontogeny. Fisheries Science 69, 772–776. Masuda, R., 2008. Seasonal and interannual variation of subtidal fish assemblages in Wakasa Bay with reference to the warming trend in the Sea of Japan. Environmental Biology of Fishes 82, 387–399. Masuda, R., Yamashita, Y., Matsuyama, M., 2008. Jack mackerel Trachurus japonicus juveniles use jellyfish for predator avoidance and as a prey collector. Fisheries Science 74, 276–284. Masuda, R., 2009. Ontogenetic changes in the ecological function of the association behavior between jack mackerel Trachurus japonicus and jellyfish. Hydrobiologia 616, 269–277.
57
Masuda, R., 2011. Ontogeny of swimming speed, schooling behaviour and jellyfish avoidance by Japanese anchovy Engraulis japonicus. Journal of Fish Biology 78, 1323–1335. Moretz, J.A., Martins, E.P., Robinson, B.D., 2007. The effects of early and adult social environment on zebrafish (Danio rerio) behavior. Environmental Biology of Fish 80, 91–101. Nakayama, S., Masuda, R., Tanaka, M., 2007. Onsets of schooling behavior and social transmission in chub mackerel Scomber japonicus. Behavioural Ecology and Sociobiology 61, 1383–1390. Reader, S.M., Kendal, J.R., Laland, K.N., 2003. Social learning of foraging sites and escape routes in wild Trinidadian guppies. Animal Behaviour 66, 729 –739. Rymer, T., Schradin, C., Pillay, N., 2008. Social transmission of information about novel food in two populations of the African striped mouse, Rhabdomys pumilio. Animal Behaviour 76, 1297–1304. Salvanes, A.G.V., Moberg, O., Braithwaite, V.A., 2007. Effects of early experience on group behaviour in fish. Animal Behaviour 74, 805–811. Shaw, E., 1978. Schooling fishes. American Scientist 66, 166–175. Takahashi, K., Masuda, R., Yamashita, Y., 2010. Ontogenetic changes in the spatial learning capability of jack mackerel Trachurus japonicus. Journal of Fish Biology 77, 2315–2325. Takahashi, K., Masuda, R., Yamashita, Y., 2012a. Stimulus-specific development of learning ability during habitat shift in pre to post-recruitment stage jack mackerel. Journal of Ethology 30, 309–316. Takahashi, K., Masuda, R., Yamashita, Y., 2012b. School for learning: sharing and transmission of feeding information in jack mackerel Trachurus japonicus juveniles. Fisheries Science 78, 269–276. Takahashi, K., Masuda, R., Yamashita, Y., 2013. What to copy: the key factor of observational learning in striped jack (Pseudocaranx dentex) juveniles. Animal Cognition, http://dx.doi.org/10.1007/s10071-013-0686-z. Warner, R.R., 1988. Traditionality of mating-site preferences in a coral reef fish. Nature 335, 719–721. Webster, M.M., Laland, K.N., 2008. Social learning strategies and predation risk: minnows copy only when using private information would be costly. Proceedings of the Royal Society London B: Biological Science 275, 2869–2876. Witte, K., Nobel, S., 2011. Learning and mate choice. In: Brown, C., Laland, K.N., Krause, J. (Eds.), Fish Cognition and Behavior. , 2nd ed. Blackwell Publishing Ltd., UK, pp. 81–107.
Contents lists available at ScienceDirect
Behavioural Processes journal homepage: www.elsevier.com/locate/behavproc
Development of observational learning during school formation in jack mackerel Trachurus japonicus juveniles Kohji Takahashi ∗ , Reiji Masuda, Yoh Yamashita Maizuru Fisheries Research Station, Kyoto University, Nagahama, Maizuru, Kyoto 625-0086, Japan
a r t i c l e
i n f o
Article history: Received 19 April 2013 Received in revised form 11 October 2013 Accepted 27 October 2013 Available online 9 November 2013 Keywords: Behavioural ontogeny Carangidae Conditioning Life history strategy Social learning
a b s t r a c t We assessed whether the development of observational learning in jack mackerel Trachurus japonicus juveniles corresponds with that of their schooling behaviour. Schooling behaviour was quantitatively analyzed by nearest neighbour distance and separation angle in two size classes of fish, 20-mm and 40-mm in body length. Observer and non-observer fish with matching sizes were conditioned to pellets by temporarily stopping aeration. Observer fish were provided with five observation trials of other individuals feeding near an air stone when aeration was stopped. After the observation trial, fish were conditioned to pellets with the stop of aeration, and then the learning process was evaluated by the increase in the association with the feeding area when aeration was stopped. In 20-mm fish, which were at an immature stage of schooling behaviour, there was no difference in the learning process between observer and nonobserver fish. In contrast, 40-mm fish were confirmed to have a well-developed schooling behaviour, and the observer learnt the feeding area more efficiently than the non-observer. This study provides evidence that observational learning develops along with the development of the social interaction. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Marine fish generally experience a major habitat shift in their life histories. Habitat shifts increase the encounter rate of novel experiences such as previously unencountered prey items and predators. Thus, rapid adaptation to novel stimuli would increase the chance of survival in migrating juveniles. Learning would help juveniles to acquire appropriate responses to the novel environment through their own experiences; therefore, learning capability can be one of the most important abilities for adapting to habitat shifts. Our previous studies investigated the relationship between the ontogenetic change of learning capability and habitat shift in marine fish using jack mackerel Trachurus japonicus juveniles (Takahashi et al., 2010, 2012a). The spatial learning capability of jack mackerel juveniles increases dramatically at ca. 50 mm in body length (BL), corresponding with the migration from offshore to coastal waters (Takahashi et al., 2010). Moreover, Takahashi et al. (2012a) compared the learning capability conditioning to surface or mid-water structures between 40-mm and 60-mm BL juveniles. The 40-mm fish were conditioned to the surface structure more rapidly than the mid-water structure, whereas the 60-mm fish learned the mid-water structure easier than the surface one. These studies showed the qualitative change in learning capability during
∗ Corresponding author. Tel.: +81 773 62 5512; fax: +81 773 62 5513. E-mail address: [email protected] (K. Takahashi). 0376-6357/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.beproc.2013.10.012
their migration from an offshore pelagic to a coastal neritic life, implying that jack mackerel juveniles have an ontogenetic change in learning capability as part of their life history strategy during their habitat shift. On the other hand, about 50% of teleost fishes in the world form schools at least for a part of their life history (Shaw, 1978). Many fishes do not exhibit social behaviour at an early life stage, but then begin to develop schooling with growth (e.g. Masuda & Tsukamoto, 1999; Ishizaki et al., 2001; Masuda et al., 2003; Fukuda et al., 2010; Masuda, 2011); i.e. they experience a shift in social environment. In a school of fish, each individual is likely to utilize social learning (Brown & Laland, 2011). For example, prey location can be learnt through observations of associating shoal mates to feeding sites without the energetic expenditures of food searching (Brown et al., 2003). The anti-predator behaviour would be enhanced through observations of conspecifics being attacked without risk to the individual (Arai et al., 2007). Researchers have found that fish acquire survival skills by social learning in various life history contexts, such as predator avoidance (Brown & Laland, 2001; Kelly et al., 2003), orientation behaviour (Warner, 1988; Fukumori et al., 2010), feeding information (Reader et al., 2003; Webster & Laland, 2008), and mate selection (Witte & Nobel, 2011). Jack mackerel form large schools after their recruitment to rocky reef habitat (Masuda, 2008; Masuda et al., 2008), and therefore they have opportunities to acquire information from conspecifics when they are schooling. In our previous study, jack mackerel juveniles at the schooling stage acquired information through observational
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57
53
Fig. 1. Schematic drawings of the tank arrangement in experiment 2: (a) demonstrator, (b) observer, and (c) non-observer fish. The tanks were placed side by side, and a black board was put between the demonstrator and observer tank except during the observation period.
learning (Takahashi et al., 2012b); i.e. in jack mackerel at 50 mm BL, fish that observed other individuals feeding at aeration in an adjacent tank were conditioned to aeration as a stimulus to initiate feeding more quickly than fish that did not observe this conspecific behaviour. Jack mackerel juveniles have a shift in social interaction at the early juvenile stage. Hatchery-reared jack mackerel juveniles begin to form schools at ca. 12 mm BL, and their schooling behaviour gradually develops (Masuda et al., unpub.). Observational learning would be more useful for their acquisition of feeding information in social environments than learning in isolation, and therefore may well be related to the development of schooling behaviour. The present study compared the observational learning capability of two size groups of jack mackerel juveniles corresponding with the developmental stage of the social environment.
2. Materials and methods 2.1. Fish Fertilized jack mackerel eggs were obtained from the spawning of wild and local origin broodstocks maintained at the Maizuru Fisheries Research Station (MFRS) of Kyoto University after the injection of human chorionic gonadotropin (0.6 IU/g body weight). A single batch of eggs was stocked in four 500-L transparent polycarbonate tanks supplied with filtered sea water at the MFRS as described in Takahashi et al. (2012a). The fish hatched on 1 June 2010. After hatching, larvae were provided with rotifers Brachionus plicatilis from the opening of the mouth to ca. 8 mm BL, Artemia sp. Nauplii from 6 to 12 mm BL, and dry pellets (N400 and N700, Kyowa Hakko Bio Co., Ltd., Tokyo, Japan from 10 to 30 mm BL; Otohime S1, Marubeni Nisshin Feed Co., Ltd., Tokyo, Japan from 30 to 40 mm BL) in accordance with growth. Each feeding was conducted twice a day in the morning and evening to satiation. The water temperature was maintained at around 25 ◦ C. Two size groups of jack mackerel were used to examine the schooling behaviour and observational learning as follows: mean ± standard deviation (SD) of the small group, defined as the 20-mm group, were 15 ± 1 mm BL and 18 ± 2 mm BL for schooling behaviour and observational learning experiments, respectively, and those of large juveniles, defined as the 40-mm group, were 38 ± 3 mm BL and 38 ± 3 mm BL, respectively. After the experiment, the fish were kept in the MFRS as broodstocks. All the experiments were performed according to the guidelines of the Regulation on Animal Experimentation at Kyoto University.
2.2. Experiment 1: schooling behaviour A large acrylic tank (length × width × height: 900 × 450 × 450 mm) was used for the evaluation of schooling behaviour. The tank was filled with sea water to a depth of 400 mm, and the water was changed every two trials. A white vinyl sheet was put under the tank to facilitate video recording, and the tank was separated from the observer by a black sheet. A video camera (HDR-CX 550; Sony, Tokyo, Japan) was installed above the experimental tank to record fish behaviour during the experiment. The schooling behaviour was compared between 20-mm and 40-mm fish juveniles. Experiments were conducted using groups consisting of four juveniles, and six groups were used for each size class. Fish were introduced into the experimental tank, and were allowed to acclimatize for 10 min. After acclimatization, fish behaviour was recorded with a video camera using a remote control to eliminate the effect of an observer. Upon completion of the experiment, fish body length was measured after the fish were anaesthetized using MS-222. Schooling behaviour was analyzed using video recordings at 5–6 min and 10–11 min after the start of recording. Nearestneighbour distance (NND) and separation angle (SA) were measured from video images for each individual. NND was calculated by dividing the measured value by the mean fish body size in a group. The measurement was conducted 20 times at 5-s intervals, and these data were averaged for each group. NND was compared between 20-mm and 40-mm fish by Welch’s ANOVA because data did not show homogeneity of variance. SA was compared between 20-mm and 40-mm fish by ANOVA. Furthermore, the SA of each size was compared to 90◦ , the value expected if movement was in a random direction, by one sample t-test. 2.3. Experiment 2: observational learning Three transparent glass tanks (length × width × height: 300 × 200 × 300 mm for 20-mm fish and 450 × 300 × 300 mm for 40-mm fish) were used in the experiment as the demonstrator, observer, and non-observer tanks (Fig. 1). Demonstrator and observer tanks were covered with black vinyl sheets except for on one side. These tanks were arranged to face each other on the uncovered sides of the tanks, and a removable black board was placed between these two tanks except during an observation trial. Both tanks were set in a water bath, and sea water filled the space between the observer and the demonstrator tanks. The non-observer tank was covered on all sides by black vinyl
54
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57
Fig. 2. Average indices of schooling behaviour for 20-mm (grey) and 40-mm (white) fish in experiment 1: (a) nearest-neighbour distance and (b) separation angle. Bars indicate standard deviation (n = 6).
sheets. Experimental tanks were provided with a depth of 200 mm circulating sea water. An air stone was set at the centre of each tank. Aeration was remotely controlled and was turned on except during the experiment. Tanks were separated by black sheets to reduce disturbance from the experimenter. A video camera set at 600 mm above one of the experimental tanks allowed the recording of the fish behaviour during the experiment. Demonstrator fish were conditioned to the stop of aeration as a conditioned stimulus with feeding pellets as an unconditioned stimulus. In the conditioning trials, 1 min after the stop of aeration, approximately 12 pellets (20-mm fish: N700, 40-mm fish: Otohime S1) subdivided into three portions were dropped near the air stone from above the tank. The demonstrators were trained until they showed a prominent response to aeration without pellets over 10 trials in a day. Two groups of four fish were used as demonstrators for each size of fish (average BL of demonstrator for 20 mm: 18 ± 3 mm, 40 mm: 35 ± 2 mm). The learning capability was compared between the observers and non-observers. Groups of four individuals were introduced into the observer and non-observer tanks on the previous day, and were allowed to acclimatize overnight. A few pellets were provided to confirm acclimatization prior to initiating the experiment for observer and non-observer fish, which was started if the fish ate these initial pellets. When the observer fish foraged the pellets, the black board between the observer and demonstrator tank was removed 30 min before the start of observation trials. Observers and demonstrators were attracted to each other when the black board was removed, indicating that they were visible to each other.
In the observation trial, observer fish were provided with a view of demonstrators aggregating in the feeding area when aeration was stopped, as had occurred in the conditioning trials of demonstrator. Since the demonstrators continued to aggregate in the feeding area when aeration was stopped, this suggested that the conditioned behaviour of the demonstrators was not biased by the presence of observers. Five observation trials were conducted at 30-min intervals. After five observation trials, the black board was replaced, and then 30 min later, conditioning trials were started. Observers and non-observers were conditioned in the same manner as demonstrators; i.e. 1 min after the stop of aeration, pellets were dropped near the air stone. The behaviour of observers and non-observers was video recorded for 2 min from 1 min before the stop of aeration to 1 min after. This process was defined as one trial, and 10 trials, separated by an interval of at least 30 min, were conducted from 12:00 to 18:00 with eight replications for each treatment and control. Fish BL was measured after anaesthesia using MS-222 when fish had accomplished the conditioning trials. The number of fish in the feeding area was used as an index of learning; the feeding area for the 20-mm and 40-mm juveniles was defined as the 50 × 50 mm and 100 × 100 mm areas surrounding the air stone, respectively. The number of fish in the feeding area was counted every 5 s during each 1-min period either with or without aeration in each trial. Aggregating behaviour to the feeding area for each trial was calculated as the aggregation index, by subtracting the number of fish in the feeding area with aeration from that without aeration. The aggregation indices during the first to tenth trials were compared by a two-way repeated measures
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57
55
Fig. 3. Average increase of the number of fish in the feeding area from the pre- to post-stimuli period in experiment 2: (a) 20-mm and (b) 40-mm fish. Symbols represent observer (䊉) and non-observer fish (). Bars indicate standard deviation (n = 8).
ANOVA with Bonferroni correction (˛ = 0.0083) between observation treatments (observer vs. non-observer fish in each size) or body sizes (20-mm fish vs. 40-mm fish in each treatment). 3. Results 3.1. Experiment 1: schooling behaviour The NND of 20-mm juveniles was 9.8 ± 4.3 BL (mean ± SD) and that of 40-mm juveniles was 2.0 ± 0.7 BL, the latter being significantly smaller than the former (Welch’s ANOVA: F1 , 5.21 = 20.8, p < 0.01; Fig. 2(a)). The SA of 20-mm juveniles was 77.9 ± 8.7◦ and that of 40-mm fish was 25.6 ± 16.1◦ , the latter being significantly smaller than the former (ANOVA: F1 , 11 = 6.65, p < 0.05; Fig. 2(b)); both values were significantly smaller than 90◦ (one sample t-test: 20-mm fish df = 5, t = −2.77, p < 0.05; 40-mm fish df = 5, t = −4.16, p < 0.01).
trials, F1 , 154 = 48.4, p < 0.001; treatments × trials, F1 , 154 = 0.00, p > 0.1; Fig. 3(a)). However, in 40-mm juveniles, there was a significant difference between the observer and non-observer treatment in the aggregation index from the first to the tenth trial (two-way repeated measures ANOVA with Bonferroni correction: treatments, F1 , 52.3 = 55.1, p < 0.001; trials, F1 , 154 = 34.2, p < 0.001; treatments × trials, F1 , 154 = 1.55, p > 0.1; Fig. 3(b)). Comparing the two size groups, the aggregation index of 40-mm juveniles was significantly higher than that of 20-mm fish in observer fish (two-way repeated measures ANOVA with Bonferroni correction: treatments, F1 , 8.25 = 29.6, p < 0.001; trials, F1 , 154 = 41.0, p < 0.001; treatments × trials, F1 , 154 = 0.66, p > 0.1), whereas there was no such difference in non-observer fish (two-way repeated measures ANOVA with Bonferroni correction: treatments, F1 , 5.94 = 3.59, p > 0.1; trials, F1 , 154 = 38.8, p < 0.001; treatments × trials, F1 , 154 = 0.31, p > 0.1). 4. Discussion
3.2. Experiment 2: observational learning There was no significant difference between observer and nonobserver fish in the aggregation index from the first to the tenth trial in the 20-mm juveniles (two-way repeated measures ANOVA with Bonferroni correction: treatments, F1 , 3.83 = 4.14, p > 0.1;
The 20-mm juveniles showed an SA slightly smaller than 90◦ , although they maintained their distance from each other. This result suggests that the 20-mm juveniles are at an immature stage of schooling behaviour. On the other hand, the 40-mm juveniles are considered to have reached the schooling stage, because they
56
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57
showed much closer NND and SA. Striped jack Pseudocaranx dentex juveniles, another carangid fish, start schooling behaviour at 12 mm total length (TL), whereas NND decrease with growth until 30 mm TL (Masuda & Tsukamoto, 1998). This finding supports the results of the present study; i.e. jack mackerel juveniles at 20 mm were still in the developing stage of schooling, while those at 40 mm BL had completed it. Aggregation to the feeding area was not different between the observation and control treatment for 20-mm fish, whereas the 40-mm fish showed an increase in aggregation in the feeding area through the observation trial of responding conspecifics. This implies that the observational learning capability of jack mackerel develops from 20 mm to 40 mm BL. Further, the present study suggests that the development of observational learning capability coincides with that of schooling behaviour. Fish without full development of schooling, i.e. 20-mm juveniles, would have fewer opportunities to observe other individuals, and therefore observational learning should be a low priority in their life. Indeed, jack mackerel at this stage are typically found associating with jellyfish and other floating objects either singly or in a small school (Masuda et al., 2008; Masuda, 2009). Meanwhile, 40-mm juveniles with fully developed schooling would have their life dependent on social behaviour. Various information on prey items and predators would be transmitted between individuals through observational learning in the social environment. Thus, the function of the observational learning capability would be to optimize the behaviour of jack mackerel juveniles at the stage of high dependence on schooling. Whereas previous studies suggested that the learning capability of juveniles changes in accordance with their environmental change from habitat shifts (Masuda & Ziemann, 2000; Makino et al., 2006; Hawkins et al., 2008; Takahashi et al., 2010, 2012a), the present study revealed that such an ontogenetic change also occurs with a change in social environment. There was no difference in the learning process in non-observer fish between 20- and 40-mm juveniles. This indicates that the ability to learn the location of the feeding area would not improve between these sizes, which is in agreement with the results of Takahashi et al. (2010) who showed that there was no difference in spatial learning capability measured by a Y-maze between 20-mm and 50-mm juveniles of jack mackerel. This suggests that something related to sociality must have happened during the juvenile stages to induce the changes in observational learning capability. There is a possibility that the change in attraction to other conspecific individuals induced an ontogenetic change in observational learning capability. Our study investigated the key factor of observational learning in striped jack juveniles, and striped jack juveniles could learn the feeding area to observe the foraging conspecific (Takahashi et al., 2013). The result suggested that attraction to conspecific fish is essential for observational learning. Attraction to conspecific fish would also be required for school formation behaviour, e.g. conspecific tracking. The change in the level of attraction may cause the development of observational learning capability and schooling behaviour. Whereas juveniles need basic locomotor and sensory organs for schooling (Hunter & Coyne, 1982; Kohno et al., 1984), the development of the central nervous system would also be necessary (Masuda & Tsukamoto, 1999). Similarly, there is the possibility that the central nervous system changes during the stage in which jack mackerel juveniles attain observational learning capability. Nakayama et al. (2007) investigated the ontogenetic change in schooling behaviour and social transmission using the chub mackerel Scomber japonicus. They showed that the social transmission between groups appeared at 30 days post-hatching, and that the volume of optic tectum dramatically increased prior to this behavioural change. Whereas the central nervous system related to observational learning has not been elucidated, the
developmental process of the nervous system in jack mackerel juveniles from 20 mm to 40 mm may provide a basis for studying the nervous system’s role in observational learning. In previous studies, a change in learning ability was observed in hatchery-reared fish (Masuda and Ziemann, 2000; Makino et al., 2006; Hawkins et al., 2008; Takahashi et al., 2012a); therefore, the timing of ontogenetic change is probably a genetically programmed life history strategy. On the other hand, the social environment of early life stages can have an impact on social behaviour in later stages (Moretz et al., 2007; Salvanes et al., 2007; Arnold & Taborsky, 2010); e.g. the observational learning capability of guppies Poecilia reticulata is influenced by the density of conspecifics in the early juvenile stage (Chapman et al., 2008). Therefore, it is likely that the early social environment is involved in the ontogenetic change of observational learning described in this study. In the present study, we were not able to evaluate which factor induced the ontogenetic change in observational learning because 40-mm fish had experienced a social environment in the rearing tank. The role of the social environmental effect at an early stage in the ontogenetic change in observational learning should be verified by further experiments, e.g. by isolating fish from conspecifics at early stages. In rodents, reliance on social learning differs between relatively social and asocial species or populations (Lupfer et al., 2003; Rymer et al., 2008). In golden hamsters Mesocricetus auratus, although adults in a solitary life did not influence one another’s food preferences, 4-week-old pups showed a significantly enhanced preference for foods that their mother was eating (Lupfer et al., 2003), suggesting that the social learning capability changes according to the shift in sociality. The present study revealed a more or less equivalent change in learning capability with the development of social interactions in conspecific fish. To our knowledge this is the first study to have shown the coincidence of the ontogeny of learning and that of social behaviour in fishes. Because of their abundant behavioural repertories (Brown et al., 2011) and relatively simple central nervous system, fishes are promising study organisms for the fundamental understanding of behavioural processes. Acknowledgements We thank Dr. Yuichi Fukunishi, Dr. Johan Bolhuis, and the reviewer whose comments substantially improved the quality of the manuscript. References Arai, T., Tominaga, O., Seikai, T., Masuda, R., 2007. Observational learning improves predator avoidance in hatchery-reared Japanese flounder Paralichthys olivaceus juveniles. Journal of Sea Research 58, 59–64. Arnold, C., Taborsky, B., 2010. Social experience in early ontogeny has lasting effects on social skills in cooperatively breeding cichlids. Animal Behaviour 79, 621–630. Brown, C., Laland, K.N., 2001. Social learning and life skills training for hatchery reared fish. Journal of Fish Biology 59, 471–493. Brown, C., Laland, K.N., 2011. Social learning in fishes. In: Brown, C., Laland, K.N., Krause, J. (Eds.), Fish Cognition and Behavior. , 2nd ed. Blackwell Publishing Ltd., UK, pp. 240–257. Brown, C., Laland, K.N., Krause, J., 2011. Fish Cognition and Behavior, 2nd ed. Blackwell Publishing Ltd., UK. Brown, C., Markula, A., Laland, K.N., 2003. Social learning of prey location in hatcheryreared Atlantic salmon. Journal of Fish Biology 63, 738–745. Chapman, B.B., Ward, A.J.W., Krause, J.K., 2008. Schooling and learning: early social environment predicts social learning ability in the guppy, Poecilia reticulata. Animal Behaviour 76, 923–929. Fukuda, H., Torisawa, S., Sawada, Y., Takagi, T., 2010. Ontogenetic changes in schooling behaviour during larval and early juvenile stages of Pacific bluefin tuna Thunnus orientalis. Journal of Fish Biology 76, 1841–1847. Fukumori, K., Okuda, N., Yamaoka, K., Yanagisawa, Y., 2010. Remarkable spatial memory in a migratory cardinalfish. Animal Cognition 13, 385–389. Hawkins, L.A., Magurran, A.E., Armstrong, J.D., 2008. Ontogenetic learning of predator recognition in hatchery-reared Atlantic salmon, Salmo salar. Animal Behaviour 75, 1663–1671.
K. Takahashi et al. / Behavioural Processes 103 (2014) 52–57 Hunter, J.R., Coyne, K.M., 1982. The onset of schooling in northern anchovy larvae, Engraulis mordax. CalcCOFI Reports 23, 246–251. Ishizaki, Y., Masuda, R., Uematsu, K., Shimizu, K., Arimoto, M., Takeuchi, T., 2001. The effect of dietary docosahexaenoic acid on schooling behaviour and brain development in larval yellowtail. Journal of Fish Biology 58, 1691–1703. Kelly, J.L., Evans, J.P., Ramnarine, I.W., Magurran, A.E., 2003. Back to school: can antipredator behaviour in guppies be enhanced through social learning? Animal Behaviour 65, 655–662. Kohno, H., Shimizu, M., Nose, Y., 1984. Morphological aspects of the development of swimming and feeding function in larval Scomber japonicus. Nippon Suisan Gakkaishi 50, 1125–1137. Lupfer, G., Frieman, J., Coonfield, D., 2003. Social transmission of flavor preferences in two species of Hamsters (Mesocricetus auratus and Phodopus campbelli). Journal of Comparative Psychology 117, 449–455. Makino, H., Masuda, R., Tanaka, M., 2006. Ontogenetic changes of learning capability under reward conditioning in striped knifejaw Oplegnathus fasciatus juveniles. Fisheries Science 72, 1177–1182. Masuda, R., Tsukamoto, K., 1998. The ontogeny of schooling behaviour in the striped jack. Journal of Fish Biology 52, 483–493. Masuda, R., Tsukamoto, K., 1999. School formation and concurrent developmental changes in carangid fish with reference to dietary conditions. Environmental Biology of Fishes 56, 243–252. Masuda, R., Ziemann, D.A., 2000. Ontogenetic changes of learning capability and stress recovery in Pacific threadfin juveniles. Journal of Fish Biology 56, 1239–1247. Masuda, R., Shoji, J., Nakayama, S., Anaka, M., 2003. Development of schooling behavior in Spanish mackerel Scomberomorus niphonius during early ontogeny. Fisheries Science 69, 772–776. Masuda, R., 2008. Seasonal and interannual variation of subtidal fish assemblages in Wakasa Bay with reference to the warming trend in the Sea of Japan. Environmental Biology of Fishes 82, 387–399. Masuda, R., Yamashita, Y., Matsuyama, M., 2008. Jack mackerel Trachurus japonicus juveniles use jellyfish for predator avoidance and as a prey collector. Fisheries Science 74, 276–284. Masuda, R., 2009. Ontogenetic changes in the ecological function of the association behavior between jack mackerel Trachurus japonicus and jellyfish. Hydrobiologia 616, 269–277.
57
Masuda, R., 2011. Ontogeny of swimming speed, schooling behaviour and jellyfish avoidance by Japanese anchovy Engraulis japonicus. Journal of Fish Biology 78, 1323–1335. Moretz, J.A., Martins, E.P., Robinson, B.D., 2007. The effects of early and adult social environment on zebrafish (Danio rerio) behavior. Environmental Biology of Fish 80, 91–101. Nakayama, S., Masuda, R., Tanaka, M., 2007. Onsets of schooling behavior and social transmission in chub mackerel Scomber japonicus. Behavioural Ecology and Sociobiology 61, 1383–1390. Reader, S.M., Kendal, J.R., Laland, K.N., 2003. Social learning of foraging sites and escape routes in wild Trinidadian guppies. Animal Behaviour 66, 729 –739. Rymer, T., Schradin, C., Pillay, N., 2008. Social transmission of information about novel food in two populations of the African striped mouse, Rhabdomys pumilio. Animal Behaviour 76, 1297–1304. Salvanes, A.G.V., Moberg, O., Braithwaite, V.A., 2007. Effects of early experience on group behaviour in fish. Animal Behaviour 74, 805–811. Shaw, E., 1978. Schooling fishes. American Scientist 66, 166–175. Takahashi, K., Masuda, R., Yamashita, Y., 2010. Ontogenetic changes in the spatial learning capability of jack mackerel Trachurus japonicus. Journal of Fish Biology 77, 2315–2325. Takahashi, K., Masuda, R., Yamashita, Y., 2012a. Stimulus-specific development of learning ability during habitat shift in pre to post-recruitment stage jack mackerel. Journal of Ethology 30, 309–316. Takahashi, K., Masuda, R., Yamashita, Y., 2012b. School for learning: sharing and transmission of feeding information in jack mackerel Trachurus japonicus juveniles. Fisheries Science 78, 269–276. Takahashi, K., Masuda, R., Yamashita, Y., 2013. What to copy: the key factor of observational learning in striped jack (Pseudocaranx dentex) juveniles. Animal Cognition, http://dx.doi.org/10.1007/s10071-013-0686-z. Warner, R.R., 1988. Traditionality of mating-site preferences in a coral reef fish. Nature 335, 719–721. Webster, M.M., Laland, K.N., 2008. Social learning strategies and predation risk: minnows copy only when using private information would be costly. Proceedings of the Royal Society London B: Biological Science 275, 2869–2876. Witte, K., Nobel, S., 2011. Learning and mate choice. In: Brown, C., Laland, K.N., Krause, J. (Eds.), Fish Cognition and Behavior. , 2nd ed. Blackwell Publishing Ltd., UK, pp. 81–107.