Home
Search
Collections
Journals
About
Contact us
My IOPscience
Zebrafish (Danio rerio) behavioural response to bioinspired robotic fish and mosquitofish (Gambusia affinis)
This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2013 Bioinspir. Biomim. 8 044001 (http://iopscience.iop.org/1748-3190/8/4/044001) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 62.141.160.13 The article was downloaded on 04/09/2013 at 09:52
Please note that terms and conditions apply.
IOP PUBLISHING
BIOINSPIRATION & BIOMIMETICS
doi:10.1088/1748-3182/8/4/044001
Bioinspir. Biomim. 8 (2013) 044001 (7pp)
COMMUNICATION
Zebrafish (Danio rerio) behavioural response to bioinspired robotic fish and mosquitofish (Gambusia affinis) Giovanni Polverino and Maurizio Porfiri 1 Department of Mechanical and Aerospace Engineering, Polytechnic Institute of New York University, Six MetroTech Center, 11201, Brooklyn, NY, USA E-mail:
[email protected]
Received 5 April 2013 Accepted for publication 22 July 2013 Published 3 September 2013 Online at stacks.iop.org/BB/8/044001 Abstract The field of ethorobotics holds promise in aiding fundamental research in animal behaviour, whereby it affords fully controllable and easily reproducible experimental tools. Most of the current ethorobotics studies are focused on the behavioural response of a selected target species as it interacts with a biologically-inspired robot in controlled laboratory conditions. In this work, we first explore the interactions between two social fish species and a robotic fish, whose design is inspired by salient visual features of one of the species. Specifically, this study investigates the behavioural response of small shoals of zebrafish interacting with a zebrafish-inspired robotic fish and small shoals of mosquitofish in a basic ecological context. Our results demonstrate that the robotic fish differentially influences the behaviour of the two species by consistently attracting zebrafish, while repelling mosquitofish. This selective behavioural control is successful in spatially isolating the two species, which would otherwise exhibit prey–predator interactions, with mosquitofish attacking zebrafish. S Online supplementary data available from stacks.iop.org/BB/8/044001/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
example, bioinspired robots have been used to investigate the behaviour of birds [4], dogs [5], lizards [6], fish [7–11] and rats [12]. Species-specific signalling has been integrated in robotic prototypes to study social behaviour in cockroaches through pheromones administration [13], in squirrels through alarm vocalizations [14] and in honeybees by reproducing the pulsing air currents typical of their dance [15]. In a similar domain, robots have also been utilized to investigate courtship in male bowerbirds [16]. The use of animal models in laboratory studies is a common practice in several research domains related to biomedicine, developmental biology and neurobiology [17]. Zebrafish (Danio rerio) is rapidly emerging as a valid experimental species due to its small size, remarkable
The interdisciplinary field of ‘ethorobotics’ holds promise in aiding hypothesis-driven research on animal behaviour by offering to researchers untapped methodologies to standardize highly customizable experimental protocols and independently control salient variables in the experimental design [1]. Robots with varying degree of biomimicry [2, 3] have been utilized to influence the behaviour of several species across a wide set of experimental paradigms, tailored to influence animal response in a lock-and-key system. For 1
Author to whom any correspondence should be addressed.
1748-3182/13/044001+07$33.00
1
© 2013 IOP Publishing Ltd
Printed in the UK & the USA
Bioinspir. Biomim. 8 (2013) 044001
Communication
shoaling tendency, high stocking density, fast reproduction rate and short intergeneration time [18–20]. The sequencing of zebrafish genome has also favoured its use as an alternative animal model to rodents for investigating functional and dysfunctional processes [17, 21]. Scientific research on the zebrafish animal model has fostered parallel advancements in the development of enabling technologies for high-throughput data acquisition [22]. In [7, 9, 11], we have demonstrated that zebrafish exhibit a marked and consistent preference for a bioinspired robotic fish, whose shape, aspect ratio, locomotion and colour pattern were inspired by a fertile female subject. While the experimental observations in [7, 9, 11] addressed zebrafish response to systematic variations of the robotic fish visual aspect and behaviour, the experimental subjects were allowed to interact only with conspecifics or robotic stimuli. In this study, we seek to extend the results of [7, 9, 11] by investigating the behavioural response of small shoals of zebrafish to the concurrent feedback offered by the zebrafish-bioinspired robotic fish and a second social fish species of comparable size and similar housing requirements to zebrafish. Mosquitofish (Gambusia affinis) was selected as the second species for their similarities in shoaling behaviour [23–25] and habitat needs to zebrafish [18, 24, 25]. Differently from zebrafish [26], mosquitofish exhibit aggressive attitudes towards other similar-sized aquatic species [27–30], are highly competitive in exploiting the environmental resources [24, 31] and their presence in non-native areas is often associated with a negative impact on the indigenous animal communities [25, 32, 33]. The extraordinary adaptive capacity to novel environments of mosquitofish is likely a determinant factor for its successful colonization of new habitats, despite the presence of competitors and potential predators [34, 35]. Recent studies have indicated an uncontrolled increase in mosquitofish populations in areas native to zebrafish [36], while significant aggression phenomena of mosquitofish towards zebrafish have been documented in [26]. Notably, in contrast with observations on zebrafish [7], recent results in [37] have shown that mosquitofish are repelled by a mosquitofish-inspired robotic fish, independently of its aspect ratio or swimming depth. The following predictions were expected to be met during the study: (i) zebrafish should maintain a consistent preference for the robotic fish even in the presence of mosquitofish; (ii) mosquitofish should display aggressive behaviours towards zebrafish and (iii) zebrafish and mosquitofish should interact differently with the robotic fish.
The two populations were procured from an online aquaria source (LiveAquaria.com, Rhinelander, WI, USA) and were acclimated for two weeks prior to the beginning of the experiments. According to their mean body length, zebrafish and mosquitofish used in this study were considered young adult sexually mature [38, 39], see figure 1. Individuals of this age are known to display prominent shoaling tendencies in either species [24, 38, 40]. While only female mosquitofish were selected for the experiments due to their gregarious tendency and to avoid the aggressive sexual activity observed in male subjects [40–43], zebrafish sex ratio was kept approximately one to one for the identical shoaling tendency between genders [44]. Zebrafish and mosquitofish were housed in separate tanks covered on three sides with opaque blue plastic panels to prevent visual interactions between species. For each species, fish density in the housing tanks was less than 0.33 fish l−1. The water temperature and pH of the housing tanks were maintained at 26 ± 1 ◦ C and 7.2, respectively. Illumination was provided by full spectrum fluorescent lights for 10 h per day, from 9 am to 7 pm, in accordance with the circadian rhythm of these fish species [24, 45]. Fish were fed with commercial flake food (Hagen Corp., Nutrafin max, USA) once a day, after the conclusion of the daily experimental session.
2. Materials and methods
2.2. Robotic fish
The experiments described in this work were approved by the Polytechnic Institute of New York University (NYU-Poly) Animal Welfare Oversight Committee AWOC-2011-101.
The robotic fish prototype considered in this study has already been described in [7, 9–11, 46]. It is made of a rigid acrylonitrile butandiene styrene plastic body shell, an oscillating tail section actuated by a servomotor and a passive mylar caudal fin. The robot parts were designed in SolidWorks and printed on a rapid prototyping machine (Stratasys, Dimension SST, USA). The robot is 15 cm long, 4.8 cm tall and 2.6 cm wide, which correspond to the smallest
(a)
(b)
Figure 1. Photograph of a wild-type zebrafish (a) and a female mosquitofish (b).
2.1. Animals and housing Two populations of fish (45 wild-type zebrafish and 45 mosquitofish) were used for this study in December 2011. 2
Bioinspir. Biomim. 8 (2013) 044001
Communication
zebrafish and mosquitofish. Each experimental shoal included three conspecifics that were randomly selected, without replacement, each day across the experimental campaign (N = 15). In the first condition (zm), the robotic fish was absent and we investigated the interaction between zebrafish and mosquitofish. In the second condition (Rm), we studied the interaction between mosquitofish and the robotic fish. In the third condition (Rzm), we analysed the interaction among zebrafish, mosquitofish and the robotic fish. Each experimental condition consisted of 45 trials executed in three repetitions of 15 trials each. After each trial, fish were isolated from their population to assure that the same individuals were not tested multiple times within each repetition. In conditions Rm and Rzm, the robotic fish was anchored on one side of the experimental tank at a 45◦ angle to the longitudinal axis of the tank using a thin plastic rod of diameter 4 mm [7, 9, 11]. The plastic rod was painted in ivory (the base robot colour) and anchored to an external support to control the vibrations induced by the robot motion. The robotic fish was anchored 1 cm below the water surface. To ensure a homogeneous visual background, an identical rod was inserted on the opposite side of the experimental tank. The robot position was systematically alternated between the two sides of the tank during each experimental condition and the temporal distribution of trials was counterbalanced across conditions, thus reducing the risk of bias in the data and limiting potential confounds arising from the time of day of each trial, respectively.
Figure 2. Experimental apparatus used in this study as pictured by the overhead webcam, showing a sample tracking of three mosquitofish marked as red vectors.
possible size that is needed for the robot to autonomously swim and maintain zebrafish aspect ratio. Notably, this robotic fish was found to elicit the highest attraction in zebrafish as compared to robots with smaller or larger aspect ratios and an identical prototype with homogenous grey coloration [7]. Namely, alterations in the colour and shape of the zebrafish-inspired robotic fish did not result in fish aversive responses, yet they have reduced or eliminated the robot’s attraction. The same robotic prototype was considered in [37] to study mosquitofish response to a robot whose colouration was inspired by mosquitofish. Therein, it was found that the aversion of mosquitofish for the robotic fish increased as its aspect ratio was skewed towards elongated body shapes. As in [7, 9, 11], the colour pattern of the robotic fish was inspired by relevant phenotypic features known to elicit social attraction in zebrafish. Specifically, the saturated yellow pigment and the magnified blue stripes were selected for their attractive effect on zebrafish observed through computeranimated images [47, 48] and experiments on different phenotypic varieties of zebrafish [44]. The presence of a flexible element allowed for a sinuous undulation of the tail mimicking the carangiform/subcarangiform swimming pattern characteristic of the two target species [49, 50]. The tail-beat frequency and tail-beat amplitude of the robot were set at 2 Hz and 1.5 cm, respectively [7, 9, 11].
2.5. Data acquisition The preference space of each fish shoal was ascertained by virtually partitioning the experimental tank into 1 cm bins along its longitudinal axis and then partitioning it into three regions; the two stimulus areas were of an equal length of 20 cm from the aquarium wall side and the central region was 34 cm long. Spatial preference and cohesion of the fish shoals were extracted by using a custom script in MATLAB (Mathworks Inc.), in which 30 snapshots were captured for each experimental trial at 10 s intervals following [7, 11, 37]. Ninety fish positions were obtained for each trial in condition Rm, while 180 fish positions were extracted for each trial in conditions zm and Rzm (90 per species). In each snapshot, fish were individually localized and their one-dimensional (1D) position along the longitudinal axis was computed. The positions obtained for each trial were then averaged in each experimental condition and used for separately estimating the spatial preference of zebrafish and mosquitofish for any of the three tank regions, as the absolute frequency of fish appearance. In order to measure the shoals’ cohesion and the interaction between individuals of the two species, two distinct measures were defined as follows: the intra-specific average nearest neighbour distance (intra-ANND) and the interspecific average nearest neighbour distance (inter-ANND). The first parameter was computed by averaging the 1D
2.3. Apparatus Three cameras were used to simultaneously film the experimental trials: an overhead webcam (Logitech, Webcam Pro 9000) and two external digital video cameras (Canon, Vixia HG20, Japan), mounted 100 cm above the water surface and 20 cm away from the short edges of the tank, respectively, see figure 2. The webcam provided a bird’s-eye view of the experimental tank. The video cameras were used to facilitate the visual identification of each species by manually creating a montage of all three views for independent verifications. 2.4. Procedure Experiments were performed in isolation under controlled conditions. The fish were allowed to acclimate in the experimental tank for 10 min [51, 52]. After the acclimatization period, a 5 min experimental session was filmed to score the fish position in the tank (observation period). A total of three experimental conditions were performed to investigate the preference space and the cohesion in shoals of 3
Bioinspir. Biomim. 8 (2013) 044001
Communication
Figure 4. Histograms of the intra-ANND (top) and inter-ANND (bottom) between zebrafish and mosquitofish shoals for each experimental condition. Condition Rm, in which only mosquitofish were present, is excluded from the inter-ANND. Zebrafish data are depicted with white histograms and black histograms refer to mosquitofish. Error bars are standard errors. Means not sharing a common superscript are significantly different (Fisher’s PLSD, p < 0.05); superscripts are independent between top and bottom histograms.
Figure 3. Histograms of the mean number of fish appearances in the robot stimulus areas (top) and in central area (bottom) of the experimental tank for each experimental condition. In condition zm, the robotic fish was not present and the frequency refers to the randomly selected left empty stimulus side (results are unaltered if the right side were selected instead). Data on zebrafish are represented with white histograms and black histograms refer to mosquitofish. Error bars are standard errors. Means not sharing a common superscript are significantly different (Fisher’s PLSD, p < 0.05); superscripts are independent between top and bottom histograms.
Interestingly, such frequency was the highest for zebrafish as they were confronted with both the robotic fish and the mosquitofish in condition Rzm. Post-hoc comparisons revealed a significant difference between this condition and all the other experimental conditions for both species. Furthermore, post-hoc comparisons between conditions zm and Rzm showed that, in the presence of the robotic fish, zebrafish spent significantly more time in the robot region as compared to either of the empty stimulus areas when the robotic fish was absent. Conversely, mosquitofish spent less time in the robot region as compared to the empty stimulus areas in the absence of the robotic fish. Notably, when comparing our results with the spatial preference observed in shoals of zebrafish in [11], the frequency with which zebrafish were observed near the robotic fish did not vary in the presence or absence of mosquitofish. The frequency of fish position in the central region was significantly different between conditions (ANOVA—F4, 210 = 11.1; p 0.01), see figure 3. Such frequency was the highest for mosquitofish as they were confronted with both the robotic fish and zebrafish in condition Rzm. For the same condition, the frequency of fish position in the central region was the lowest in zebrafish. Post-hoc comparisons revealed that such highest and lowest values were different from any other experimental condition. The cohesion within shoals of conspecifics was influenced by the experimental condition (ANOVA—F4, 210 = 42.2; p 0.01), see figure 4. In particular, post-hoc comparisons showed that the minimum value of the intra-ANND, corresponding to the strongest cohesion, was attained by zebrafish in condition Rzm and that such value was
distances between each subject comprising the shoal and its conspecific nearest neighbour, while the second index was evaluated by averaging the 1D distances between each subject and its heterospecific nearest neighbour. Notably, the interANND depends on the species considered as a reference; thus, for completeness, data are reported for both mosquitofish and zebrafish as a reference species. 2.6. Data analysis Variation between trials was calculated using a two-way ANOVA for both fish position and shoals’ cohesion. In particular, condition and repetition were considered as independent factors, while the frequency of fish position in both the robot (selected as the left stimulus area for condition zm) and the central regions, along with the ANND, were considered as dependent variables. Data analysis was carried out using Statview 5.0. The significance level was set at p < 0.05. Fisher’s protected least significant difference (PLSD) post-hoc tests were used when a significant main effect of the condition variable was observed.
3. Results Fish spatial preference for the robot varied across the experimental conditions. Specifically, the frequency of the fish (zebrafish and mosquitofish) position in the robot stimulus area was significantly different across the conditions (ANOVA—F4, 210 = 16.3; p 0.01), see figure 3. 4
Bioinspir. Biomim. 8 (2013) 044001
Communication
significantly different from all the other intra-ANND measurements. Furthermore, the cohesion of mosquitofish confronted with zebrafish was not significantly different in the absence or presence of the robotic fish; namely, intra-ANND in Rzm and zm was not found to be significantly different. Posthoc comparisons also demonstrated that the intra-ANND of mosquitofish increased in the presence of zebrafish, attaining values significantly larger than any intra-ANND observed in zebrafish and indicating a remarkable reduction in their cohesion. A condition effect was also found for the interANND (ANOVA—F3,168 = 21.3; p 0.01), upon excluding condition Rm (involving only mosquitofish) from the analysis, see figure 4. Post-hoc comparisons indicated a significant difference between the largest value of the inter-ANND, corresponding to zebrafish as the reference species in condition Rzm, and any other inter-ANND measurement. In contrast, the lowest value of the inter-ANND was attained in condition zm with mosquitofish as the reference species, and post-hoc comparisons revealed a significant difference between such value and any other inter-ANND observation. Trials were not found to be significantly different across the three repetitions of 15 trials each in either the preference (ANOVA—p = 0.1 and p = 0.06 for the robot and the central region) or the ANND analyses (ANOVA—p = 0.3 and p = 0.2 for the intra-ANND and inter-ANND, respectively), data not shown.
characteristics of aggressive social fish species that are efficient in optimizing their use of environmental resources [24, 31]. Moreover, mosquitofish interactions with other species of similar size are generally competitive, if not predatorial [27–30]. In agreement with these observations, our results indicate that the behavioural responses of zebrafish and mosquitofish in a preference test employing a robotic stimulus are different (see supplementary video available from stacks.iop.org/BB/8/044001/mmedia). We find that mosquitofish are highly cohesive as evidenced by an intra-ANND value of approximately three to four body lengths, see figure 4. Considering this low value compared to the 74 cm of the aquarium length, we thus hypothesize that mosquitofish interact as a group. However, cohesion in mosquitofish becomes significantly weaker when zebrafish are present, independently of the robot’s presence. In line with observations in [26], this result suggests that, in the presence of zebrafish, mosquitofish’ social behaviour is reduced in favour of individual competition or possible predation on zebrafish. This response is further confirmed by the smaller inter-specific distance in mosquitofish towards zebrafish than the opposite, see figure 4. While this quantity alone does not allow for fully dissecting the social interaction between the species from the spatial segregation elicited by the robotic fish, the observed difference suggests that the species differentially interact with each other, whereby the lure of mosquitofish towards zebrafish is not reciprocated by zebrafish. Similarly to our findings in [11], zebrafish consistently display strong cohesion, with intra-ANND values of the order of two or three body lengths. The strongest cohesion corresponds to the simultaneous presence of mosquitofish and zebrafish in the test tank, which is accompanied by an increase in the amount of time spent in the robot stimulus area. Our results suggest that the aggressive behaviour of mosquitofish for zebrafish can be modulated by the robotic fish, which is able to selectively attract/repel zebrafish/mosquitofish as the two species are mixed in an elementary ecological community. The aversion of mosquitofish for the robotic fish is not fully overcome by their attraction towards zebrafish, as evidenced by the fact that the mean time spent by mosquitofish in the vicinity of the robot does not change if zebrafish are present or absent. However, mosquitofish spend more time in the central area of the tank in the presence of zebrafish, while zebrafish spend significantly less time in the centre and more time in the robot stimulus area in the presence of the robotic fish and mosquitofish. In other words, the attractive role played by zebrafish increases the spatial preference of mosquitofish towards the central area without affecting their aversion for the robot, see figure 3. In contrast, the simultaneous attraction of zebrafish towards the robotic fish and their aversion for mosquitofish contribute to the large preference for the robot stimulus area, see figure 3. To the best of our knowledge, this is the first evidence of a robot that can selectively regulate the response of multiple species, suggesting possible venues in conservation studies and alien species control.
4. Discussion The results of this study support the evidence in [7, 9, 11] that zebrafish are attracted to a robotic fish, whose design is inspired by zebrafish, while suggesting that the same robot induces aversion in mosquitofish, see figure 3. The experimental evidence that the robot’s bright pigmentation pattern, typical of zebrafish, produces an aversive impact on mosquitofish spatial preference agrees with previous studies showing that mosquitofish, in their native environments, are mostly preyed upon by piscivorous fish of the Centrarchidae family, also known as ‘sunfishes’ [53]. These North-American carnivorous fish are generally characterized by moderately to very bright colour patterns, having a body size considerably larger than mosquitofish and feeding patterns typical of open water predators [53, 54]. Thus, we consider the possibility that the robotic fish used in our study induces an aversive reaction in shoals of mosquitofish when replicating the characteristic features of their sympatric predators, in terms of colour pattern, size and swimming position in the water column. Novelty and habituation are unlikely to play a role in fish preference for the robotic fish due to the selected experimental design and the obtained results. In fact, the experimental protocol featured a 10 min acclimatization session to compensate for novelty effects [52, 55], and the experimental results indicate that trial repetitions were not significantly different between each other in both the behavioural responses of zebrafish and mosquitofish. While zebrafish are typically considered a social freshwater prey species [18], mosquitofish embody all the 5
Bioinspir. Biomim. 8 (2013) 044001
Communication
Acknowledgments
[14] Partan S R, Larco C P and Owens M J 2009 Wild tree squirrels respond with multisensory enhancement to conspecific robot alarm behaviour Anim. Behav. 77 1127–35 [15] Michelsen A, Andersen B B, Storm J, Kirchner W H and Lindauer M 1992 How honeybees perceive communication dances, studied by means of a mechanical model Behav. Ecol. Sociobiol. 30 143–50 [16] Patricelli G L, Uy A C, Walsh G and Borgia G 2002 Sexual selection: male displays adjusted to females response Nature 415 279–80 [17] Miklosi A and Andrew R 2006 The zebrafish as a model for behavioral studies Zebrafish 3 227–34 [18] Engeszer R E, Patterson L B, Rao A A and Parichy D M 2007 Zebrafish in the wild: a review of natural history and new notes from the field Zebrafish 4 21–40 [19] Lawrence C 2007 The husbandry of zebrafish (Danio rerio): a review Aquaculture 269 1–20 [20] Reed B and Jennings M 2010 Guidance on the Housing and Care of Zebrafish (Danio rerio) (West Sussex: RSPCA Research Animals Department) [21] Kuo P D and Eliasmith C 2005 Integrating behavioral and neural data in a model of zebrafish network interaction Biol. Cybern. 93 178–87 [22] Gerlai R 2010 High-throughput behavioral screens: the first step towards finding genes involved in vertebrate brain function using zebrafish Molecules 15 2609–22 [23] Bisazza A and Pilastro A 2000 Variation of female preference for male coloration in the eastern mosquitofish Gambusia holbrooki Behav. Genet. 30 207–12 [24] Pyke G 2005 A review of the biology of Gambusia affinis and G holbrooki Rev. Fish Biol. Fisher. 15 339–65 [25] Pyke G 2008 Plague minnow or mosquito fish? A review of the biology and impacts of introduced Gambusia species Annu. Rev. Ecol. Evol. Syst. 39 171–91 [26] Thompson K A, Hill J E and Nico L G 2012 Eastern mosquitofish resists invasion by nonindigenous poeciliids through agonistic behaviors Biol. Invasions 14 1515–29 [27] Barrier R F G and Hicks B J 1994 Behavioral interactions between black mudfish (Neochanna diversus Stokell, 1949: Galaxiidae) and mosquitofish (Gambusia affinis Baird & Girard, 1854) Ecol. Freshwater Fish 3 93–99 [28] Belk M C and Lydeard C 1994 Effect of Gambusia holbrooki on a similar-sized, syntopic poeciliid, Heterandria formosa: competitor or predator? Copeia 1994 296–302 [29] Laha M and Mattingly H T 2007 Ex situ evaluation of impacts of invasive mosquitofish on the imperiled barrens topminnow Environ. Biol. Fish. 78 1–11 [30] Keller K and Brown C 2008 Behavioural interactions between the introduced plague minnow Gambusia holbrooki and the vulnerable native Australian ornate rainbowfish Rhadinocentrus ornatus, under experimental conditions J. Fish Biol. 73 1714–29 [31] Meffe G K and Snelson F F 1989 An ecological overview of poeciliid fishes Ecology and Evolution of Livebearing Fishes (Poeciliidae) ed G K Meffe and F F Snelson (Englewood Cliffs, NJ: Prentice Hall) pp 13–31 [32] Kats L and Ferrer R 2003 Alien predators and amphibian declines: review of two decades of science and the transition to conservation Divers. Distrib. 9 99–110 [33] Lawler S, Dritz D, Strange T and Holyoak M 1999 Effects of introduced mosquitofish and bullfrogs on the threatened California red-legged frog Conserv. Biol. 13 613–22 [34] Rehage S, Barnett B and Sih A 2005 Behavioral responses to a novel predator and competitor of invasive mosquitofish and their non-invasive relatives (Gambusia sp.) Behav. Ecol. Sociobiol. 57 256–66 [35] Ward A and Mehner T 2010 Multimodal mixed messages: the use of multiple cues allows greater accuracy in social
The authors would like to gratefully acknowledge Dr N Abaid for her valuable help on the statistical analysis, Drs S Butail and S Macr`ı for reviewing the manuscript, and V Kopman and F Possemato for the technical support on the figures. The authors would also like to thank Dr L Lloyd, Dr D Duffy, Dr D Duncan, Dr M Imlay and Dr B Masters for the references provided and the useful advices on the state of the art on the topic. This material is based upon work supported by the National Science Foundation under grant no CMMI-0745753 and partially supported by the Honors Center of Italian Universities (H2CU) through a scholarship to GP. The authors would also like to thank the two anonymous reviewers for their careful reading of the manuscript and for giving useful suggestions that have helped improve the work and its presentation.
References [1] Krause J, Winfield A F T and Deneubourg J L 2011 Interactive robots in experimental biology Trends Ecol. Evol. 26 369–75 [2] Lepora N F, Verschure P and Prescott T J 2013 The state of the art in biomimetics Bioinspir. Biomim. 8 013001 [3] Lakhtakia A and Mart´ın-Palma R J 2012 Bioinspiration, Biomimetics, and Bioreplication ed A Lakhtakia and R J Mart´ın-Palma (Bellingham, WA: SPIE) [4] Fernandez-Juricic E and Kowalski V 2011 Where does a flock end from an information perspective? A comparative experiment with live and robotic birds Behav. Ecol. 22 1304–11 [5] Kubinyi E, Miklosi A, Kaplan F, Gacsi M, Topal J and Csanyi V 2004 Social behaviour of dogs encountering AIBO, an animal-like robot in a neutral and in a feeding situation Behav. Process. 65 231–9 [6] Partan S R, Otovic P and Price V L 2011 Assessing display variability in wild brown anoles Anolis sagrei using a mechanical lizard model Curr. Zool. 57 140–52 [7] Abaid N, Bartolini T, Macri S and Porfiri M 2012 Zebrafish responds differentially to a robotic fish of varying aspect ratio, tail beat frequency, noise, and color Behav. Brain Res. 224 545–53 [8] Faria J J et al 2010 A novel method for investigating the collective behaviour of fish: introducing ‘Robofish’ Behav. Ecol. Sociobiol. 64 1211–8 [9] Kopman V, Laut J W, Polverino G and Porfiri M 2013 Closed-loop control of zebrafish response using a bioinspired robotic-fish in a preference test J. R. Soc. Interface 10 20120540 [10] Marras S and Porfiri M 2012 Fish and robots swimming together: attraction towards the robot demands biomimetic locomotion J. R. Soc. Interface 9 1–13 [11] Polverino G, Abaid N, Kopman V, Macr`ı S and Porfiri M 2012 Zebrafish response to robotic fish: preference experiments on isolated individuals and small shoals Bioinspir. Biomim. 7 036019 [12] Takanishi A, Aoki T, Ito M, Ohkawa Y and Yamaguchi J 1998 Interaction between creature and robot: development of an experiment system for rat and rat robot interaction IEEE/RSJ Int. Conf. on Intelligent Robots and Systems (Victoria, BC, Canada) pp 1975–80 [13] Halloy J et al 2007 Social integration of robots into groups of cockroaches to control self-organized choices Science 318 1155–8 6
Bioinspir. Biomim. 8 (2013) 044001
[36]
[37]
[38] [39] [40]
[41]
[42] [43]
[44]
Communication
recognition and predator detection decisions in the mosquitofish, Gambusia holbrooki Behav. Ecol. 21 1315–20 Raghavan R, Prasad G, Anvar-Ali P H and Pereira B 2008 Exotic fish species in a global biodiversity hotspot: observations from River Chalakudy, part of Western Ghats, Kerala, India Biol. Invasions 10 37–40 Polverino G and Porfiri M 2013 Mosquitofish (Gambusia affinis) responds differentially to a robotic fish of varying swimming depth and aspect ratio Behav. Brain Res. 250 133–8 Buske C and Gerlai R 2011 Shoaling develops with age in zebrafish (Danio rerio) Prog. Neuropsychopharmacol. Biol. Psychiatry 35 1409–15 Krumholz L A 1948 Reproduction in the western mosquitofish, Gambusia affinis affinis (Baird & Girard), and its use in mosquito control Ecol. Monogr. 18 1–43 Polverino P, Liao J C and Porfiri M 2013 Mosquitofish (Gambusia affinis) reference and behavioral response to animated images of conspecifics altered in their color, aspect ratio, and swimming depth PLoS One 8 e54315 Pilastro A, Benetton S and Bisazza A 2003 Female aggregation and male competition reduce costs of sexual harassment in the mosquitofish Gambusia holbrooki Anim. Behav. 65 1161–7 Dadda M, Pilastro A and Bisazza A 2005 Male sexual harassment and female schooling behavior in the eastern mosquitofish Anim. Behav. 70 463–71 Smith C C and Sargent R C 2006 Female fitness declines with increasing female density but not male harassment in the western mosquitofish, Gambusia affinis Anim. Behav. 71 401–7 Snekser J L, Ruhl N, Bauer K and McRobert S P 2010 The influence of sex and phenotype on shoaling decisions in zebrafish Int. J. Comp. Psychol. 23 70–81
[45] Cahill G 2002 Clock mechanisms in zebrafish Cell Tissue Res. 309 27–34 [46] Abaid N, Kopman V and Porfiri M 2012 An attraction toward engineering careers: the story of a Brooklyn outreach program for KuFFFD12 students IEEE Robot. Autom. Mag. 20 31–39 [47] Rosenthal G and Ryan M 2005 Assortative preferences for stripe patterns in danios Anim. Behav. 70 1063–6 [48] Saverino C and Gerlai R 2008 The social zebrafish: behavioral responses to conspecific, heterospecific, and computer animated fish Behav. Brain Res. 191 77–87 [49] Plaut I 2000 Effects of fin size on swimming performance, swimming behaviour and routine activity of zebrafish Danio rerio J. Exp. Biol. 203 813–20 [50] Plaut I 2002 Does pregnancy affect swimming performance of female Mosquitofish, Gambusia affinis? Funct. Ecol. 16 290–5 [51] Cote J, Fogarty S, Weinersmith K, Brodin T and Sih A 2010 Personality traits and dispersal tendency in the invasive mosquitofish (Gambusia affinis) Proc. R. Soc. B 277 1571–9 [52] Wong K et al 2010 Analyzing habituation responses to novelty in zebrafish (Danio rerio) Behav. Brain Res. 208 450–7 [53] Langerhans B R, Layman C A, Shokrollahi M A and Dewitt T J 2004 Predator-driven phenotypic diversification in Gambusia affinis Evolution 58 2305–18 [54] Kraft C E, Carlson D M and Carlson M 2006 Inland fishes of New York (online) version 4.0 (Ithaca, NY: Cornell University) [55] Champagne D L, Hoefnagels C C, de Kloet R E and Richardson M K 2010 Translating rodent behavioral repertoire to zebrafish (Danio rerio): relevance for stress research Behav. Brain Res. 214 332–42
7
