[advances in marine biology] advances in marine biology volume 60 volume 60 || social aggregation in...

C H A P T E R F O U R

Social Aggregation in the Pelagic

Zone with Special Reference to Fish

and Invertebrates

David A. Ritz*,1, Alistair J. Hobday†, John C. Montgomery‡

and Ashley J.W. Wardy

Contents1. Introduction 163

2. Aggregation Principles and Features in Pelagic Ecosystems 166

2.1. Origins of sociality 170

2.2. Significance and benefits of social aggregation 171

2.3. Structure and functions of social aggregations 176

2.4. Association patterns within aggregations 183

2.5. Sensing the behaviour of neighbours 184

2.6. Social networks 190

3. Technology Breakthroughs in Experimental and Observational

Methods 192

3.1. Video and motion analysis software 192

3.2. Optical plankton counters and holography 197

3.3. Acoustic technology 198

3.4. Electronic tags 203

3.5. Future technology challenges 204

4. Theoretical Developments in Social Aggregation 205

5. Social Aggregation, Climate Change and Ocean Management 208

6. Conclusion 211

6.1. Do reviews stimulate new work? 211

6.2. Future needs and synthesis 212

Acknowledgements 214

References 214

* School of Zoology, University of Tasmania, Hobart, Australia† Wealth from Oceans Flagship, CSIRO Marine and Atmospheric Research, Hobart, Tasmania, Australia‡ Leigh Marine Laboratory, University of Auckland, New Zealandy School of Biological Sciences, University of Sydney, Sydney, New South Wales, Australia1 Corresponding author: Email: [email protected]

Advances in Marine Biology, Volume 60 © 2011 Elsevier Ltd

ISSN: 0065-2881, DOI: 10.1016/B978-0-12-385529-9.00004-4 All rights reserved.

161

mailto:[email protected]

http://dx.doi.org/10.1016/B978-0-12-385529-9.00004-4

Abstract

Aggregations of organisms, ranging from zooplankton to whales, are an

extremely common phenomenon in the pelagic zone; perhaps the best known

are fish schools. Social aggregation is a special category that refers to

groups that self-organize and maintain cohesion to exploit benefits such as

protection from predators, and location and capture of resources more effec-

tively and with greater energy efficiency than could a solitary individual. In

this review we explore general aggregation principles, with specific reference

to pelagic organisms; describe a range of new technologies either designed

for studying aggregations or that could potentially be exploited for this pur-

pose; report on the insights gained from theoretical modelling; discuss the

relationship between social aggregation and ocean management; and specu-

late on the impact of climate change. Examples of aggregation occur in all

animal phyla. Among pelagic organisms, it is possible that repeated co-

occurrence of stable pairs of individuals, which has been established for

some schooling fish, is the likely precursor leading to networks of social

interaction and more complex social behaviour. Social network analysis has

added new insights into social behaviour and allows us to dissect aggrega-

tions and to examine how the constituent individuals interact with each

other. This type of analysis is well advanced in pinnipeds and cetaceans, and

work on fish is progressing. Detailed three-dimensional analysis of schools

has proved to be difficult, especially at sea, but there has been some prog-

ress recently. The technological aids for studying social aggregation include

video and acoustics, and have benefited from advances in digitization, minia-

turization, motion analysis and computing power. New techniques permit

three-dimensional tracking of thousands of individual animals within a single

group which has allowed novel insights to within-group interactions.

Approaches using theoretical modelling of aggregations have a long history

but only recently have hypotheses been tested empirically. The lack of syn-

chrony between models and empirical data, and lack of a common framework

to schooling models have hitherto hampered progress; however, recent

developments in this field offer considerable promise. Further, we speculate

that climate change, already having effects on ecosystems, could have dra-

matic effects on aggregations through its influence on species composition

by altering distribution ranges, migration patterns, vertical migration, and

oceanic acidity. Because most major commercial fishing targets schooling

species, these changes could have important consequences for the depen-

dent businesses.

Key Words: social aggregation; pelagic zone; marine; association; social

networks; technology; climate change; modelling

162 David A. Ritz et al.

1. Introduction

The marine pelagic zone is defined as the water column, usually in theopen sea. Further divisions of the water column into epipelagic and mesope-lagic can be made; however, here we use the term generally. It differs fromthe coastal marine domains with regard to ecological patterns; high alphadiversity, low beta diversity; apparent lack of keystone predators; few exam-ples of trophic cascades; and little apparent competition for space. Themarine pelagic environment represents 99% of the biosphere volume (Angel,1993). In addition to supplying more than 80% of the fish consumed byhumans (Pauly et al., 2002), pelagic ecosystems account for almost half of thephotosynthesis on Earth (Field et al., 1998). Just as productivity in the pelagicocean is not uniform, individuals are not distributed evenly, and clustering isthe norm. Because of the lack of geological substrate, as in coastal regions,many pelagic species are highly mobile as individuals or populations. In thisreview, we focus on examples from species living in the upper 200 m, whichis also known as the euphotic zone.

Animals need to eat to survive, and in mobile pelagic ecosystems thismeans finding prey. However, the average concentration of resources in theworld’s oceans is insufficient for growth and survival of a variety of marinespecies, ranging from planktonic larvae to top predators (Steele, 1980; Levin,1992; Genin et al., 2005). Therefore, their survival depends on encounteringdense patches of prey that, in the case of zooplankton, form aggregationsthat vary in size along a continuum of spatial scales from 107 to 101 m(Fig. 4.1) (Haury et al., 1978; Mackas et al., 1985, Nicol, 2006).

Steele’s (1980) analysis showed that the patchiness resulting from aggrega-tion increases with trophic level (Fig. 4.2). This seems to be a consequenceof the fact that the higher the trophic level, the less the response to thedetailed structure of the local environment, and a greater ability to use large-scale ocean features such as currents or fronts. The higher the trophic level,the less are the organisms dependent on short-term events such as storms,which markedly affect phytoplankton production, and active behaviour playsa more dominant role in generating patchiness.

This prey aggregation, in turn, aggregates their predators at the samelocations. But why is phytoplankton, the base of the food chain, patchy?The main limitations on primary production are physical and chemical (i.e.light and nutrient concentrations). Variations in the distribution of lightand nutrients occur both temporally and spatially in the ocean. The higherthe trophic level, the lower the physical environment plays in drivingspatial variability of standing stock, and the more behavioural processesassume importance (Steele, 1980; Folt and Burns, 1999). A challenge for thepredators then is first to locate these patchy prey aggregations and to remain

163Social Aggregation in the Pelagic Zone with Special Reference to Fish and Invertebrates

within them until it is no longer profitable to continue feeding. Area-restricted search patterns for food are widespread phenomena among pelagicpredators from copepods to whales indicating that many predators areadapted to find and exploit aggregated prey (Steele, 1980; Leising andFranks, 2000; Leising, 2001; De Robertis, 2002). While aggregation is ubiq-uitous at all scales in pelagic ecosystems, it is not simply a passive processwhere individuals gather together to exploit a food source and separate oncethe food has been eaten. The numerous additional benefits of group livingensure that groups of many different species remain cohesive for non-feedingperiods though membership may change. These benefits are usually reportedas protection from predators, facilitation of foraging and feeding, access tocentralized information, energy saving and facilitation of mate finding andreproduction (Wilson, 1975; Ritz, 1994; Hamner and Parrish, 1997;Heppner, 1997; Krause and Ruxton, 2002).

Persistent animal aggregation has been called a central problem in eco-logical and evolutionary theory (Levin, 1997; Flierl et al., 1999) because ofthe apparently conflicting requirements of short-term selfishness and longer-term group benefits. It may be that the study of the ‘social histories ofgenetic aggregations and organelle symbioses’ can resolve this dilemma(Frank, 2007). We contribute to the analysis of social aggregation by

Figure 4.1 The Stommel diagram, overlain to show the scales that can be sampled with

various platforms, and features such as fronts. From Kaiser et al. (2005), with permission from

Oxford University Press.


reviewing the social behaviour of invertebrates and fish living in the upper200 m of the pelagic environment, but where appropriate, we use examplesfrom marine birds and mammals. This review builds on Ritz (1994), andthus we restricted the present review to post-1994 discoveries except wherereference to earlier papers is necessary for clarity or because of previousomission. Because the scope of this review has been expanded to includefish and, where appropriate, other vertebrates, relevant pre-1994 papers arealso included for these groups. We explore general aggregation principles(Section 2), describe a range of new technologies and provide examples ofthe insights gained from their use (Section 3), and from theoretical modelling(Section 4). In Section 5 we discuss the relationship between social aggrega-tion and ocean management and speculate on the possible impact of climatechange. Since this review complements Ritz (1994), we also examinewhether the post-1994 literature on the subject of social aggregation indi-cates if the earlier review stimulated research in directions identified as beingparticularly worthy of further study. We did this by using search terms associ-ated with the previously identified gaps for the subsequent period. We con-clude with areas ripe for further research to advance understanding of socialaggregation (Section 6).

We note that review papers offer an opportunity for synthesis, com-parison, gap analysis and identification of new areas for attention. Explicitguidelines to achieve these objectives in a repeatable and transparentfashion have been codified for medical reviews by Roberts et al. (2006),who also note that ecological reviews often fail to measure up to thesecriteria. Many of these criteria helped to shape this review, but in particu-lar, identification of the sources of evidence and how they were obtainedallows assessment as to whether the material included is likely to becomprehensive with respect to a topic of interest. Depending on the

Figure 4.2 Patchiness resulting from aggregation increases with trophic level. Modified

from Steele (1980).


presentation of this material, this allows repeatability in future. We performeda comprehensive search for relevant material using several search engines: ISIWeb of Science, Google Scholar, Science Daily using the following terms:Social aggregation in pelagic environments; group dynamics; three-dimensional analysisof pelagic aggregations; modelling pelagic aggregations; pelagic aggregations and oceanmanagement; pelagic aggregations and climate change, and the contractions of thesewords. We did not to restrict our literature search to specific journals, as wewere concerned we might miss important insights and contributions injournals covering alternative disciplines. Additional materials were obtainedfrom reference lists in papers located using our search procedure, ourpersonal reference collections, and from discussion with expert colleagues. Inthis way we accessed relevant breakthroughs in the study of social insects andhumans. Grey literature is difficult to access with traditional search tools (e.g.Biological Abstracts), but increasing use of the Internet allows searching usingthe same keywords for posted grey literature.

2. Aggregation Principles and Features in

Pelagic Ecosystems

Before concentrating on social aggregation, some general pointsabout aggregation are relevant. For example, the importance of aggregationfor energy transfer is often ignored. This energy transfer can be trophic, orspatial, connecting habitats and allowing biological processes to be enhancedin ‘non-productive’ areas. Hydrodynamic patterns can concentrate resources(Alldredge and Hamner, 1980) while migrating animals cause cross-habitatredistribution of carbon and nutrients (Young et al., 1996). Furthermore ithas been shown that schooling animals, by their swimming actions, are animportant source of fine-scale turbulence in the ocean (Huntley and Zhou,2004). They found that estimated rates of kinetic energy production by ani-mal schools are all of the same order, i.e. 1025 W kg21, irrespective of size(see Table 4.1).

Based on these data it appears that animal-induced turbulence is compara-ble in magnitude to energy dissipation resulting from major storms. In fact,according to Dewar et al. (2006), the biosphere generates enough power tostir the ocean. More recent work by Katija and Dabiri (2009) shows thatsuch fine-scale turbulence is primarily dissipated as heat. These authors high-light an alternative mechanism of mixing originally suggested by Darwin(1953), which depends on animal shape and ‘drift volume’, i.e. the volumeof fluid that migrates with the animal as it swims. Importantly, the drift vol-ume of adjacent animals in an aggregation may increase the effective size oftheir combined boundary layers, enhancing the possibility of vertical mixing.


The disadvantage of group living includes predator attraction, local deple-tion of food resources, competition for food and spread of disease (Parrishand Edelstein-Keshet, 1999; Hoare and Krause, 2003) and the trade-offs havebeen examined using a range of evolutionary models. These studies oftenadvocate greater integration between empirical work, theoretical and model-ling approaches (see Parrish and Edelstein-Keshet, 1999).

Aggregations in the pelagic ecosystem may occur as a result of severalprocesses:

1. Passive aggregation including the concentrating effects of circulationsuch as fronts from river plumes, Langmuir circulation and internal waves(Flierl et al., 1999; Banas et al., 2004), and over abrupt topographies,such as the shelf break and seamounts (Boehlert and Genin, 1987), andcoral reefs (Genin et al., 1988, 1994).

2. Active and non-social aggregation including independent attraction ofconspecific individuals to a food resource (e.g. copepods, Poulet andOuellet, 1982); or to a light source (Yen and Bundock, 1997); predatorsmay gather at the same natural features (Klimley et al., 2003; Hobdayand Campbell, 2009) as well as artificial structures such as fish aggrega-tion devices (FADs) (Freon and Dagorn, 2000).

3. Active and social aggregation that includes groups that ‘self-organize’and maintain cohesion because of the many derived benefits (Ritz,1994; Krause and Ruxton, 2002). Parrish and Edelstein-Keshet (1999)

Table 4.1 Kinetic energy production (Ep(W kg21)) by a range of schooling species

Species Mass (kg) Abundance

(no. m23)

Speed

(m s21)

η Ep (W kg21)

Euphausia superba 0.0002 30000 0.05 0.11a 2.6431025

Engraulis japonicus 0.002 1294 0.09 0.22 1.4131025

Engraulis mordax 0.010 115 0.14 0.24 1.0631025

Sardinops saqax 0.033 29.4 0.19 0.26 1.3131025

Clupea harengus 0.30 4.7 0.35 0.30 3.9031025

Pollachius virens 2.30 0.25 1.05b 0.39 2.7031025

Thunnus albacares 77 0.0035 1.59 0.44 4.4931025

Tursiops truncatus 21.5 0.0010 3.35c 0.85c 2.803l025

Thunnus thynnus 318 6.53 1024 1.30 0.48 5.3631025

Orcin us orca 1645 8.93 1025 3.95c 0.87c 4.0331025

Physeter macrocephalus 19850 4.53 1026 2.08d 0.83e 6.7731025

afrom Torres (1984).baverage swimming speed of free-swimming saithe schools (Pedersen, 2001).cdirect measurement of cruising speed and propulsive efficiency (Fish, 1998).dfrom Rice (1989).eapproximated from measurements on the white whale Delphinapterous leucas (Fish, 1998).Reproduced with permission from Huntley and Zhou (2004).


define social animal aggregations as those that ‘self-organize’ asopposed to aggregations that form in response to external cues e.g.light or food.

It is this active social aggregation that is the focus of this review. Thissubset of aggregation is sometimes termed congregation (Turchin, 1997)and occurs in a range of invertebrates and vertebrates as shown in Fig. 4.3.

Figure 4.3 Examples of aggregations of invertebrates, fish and marine mammals. (A)

Schooling krill, Nyctiphanes australis; (B) mysids, Paramesopodopsis rufa; (C) squid, Sepioteuthis

sepiodea; (D) school of Real Bastard Trumpeter, Mendosoma lineatum; (E) school of northern

bluefin tuna, Thunnus thynnus; (F) pod of dolphins, Tursiops truncatus. (A) Photo by Rudi

Kuiter; (B) photo by Jon Bryan; (C) photo by Ruth Byrne; (D) photo by Ron Mawbey; (E) photo

by Bill Pearcy; (F) photo by Simon Talbot.


Ritz (1994) presented many examples of aggregations of pelagicinvertebrates (his Table 1) with their spatial and temporal attributes.Figure 4.4 shows the spatial and temporal scales of aggregation ofAntarctic krill (Euphausia superba) and illustrates the range of descriptiveterms applied to groups of this species (see also Box 4.1).

Temporalscale

Spatialscale

ConcentrationMths Macro>1000 km

PatchesWks Meso 10–1000 km

Super-swarms

: few 1000 g m–3

t:100-250mm

Swarms: few 10 to several

100gm–3

Irregular forms

: few 100s g

Non-aggregateforms

HrsMicro <10km

l: up to several kmt: 1-20 ml: several 10s m

m–3

t: 10 cm: <0.1 g m–3

Cohesive aggregations

Dispersed aggregations

Non-aggregatedforms

Layers andScattered forms

:10gm–3 (approx):t: largel: many km

ρ ρρ

ρρ

Figure 4.4 Nomenclature of aggregations of Antarctic krill, Euphausia superba. ρ5 density;

t5 thickness; l5 length. Reproduced with permission from Miller and Hampton (1989).

Box 4.1

Terms used to define aquatic invertebrate groups (after Ritz, 1994, and

Folt and Burns, 1999) and social groups of aquatic vertebrates (after

Pitcher and Parrish, 1993; Shane et al., 1986) as used in this review.

Swarm: used here to mean a discrete integrated social group with mem-

bers evenly spaced, but not polarized (aligned in the same direction).

School: discrete integrated social group in which members are polarized

and displaying synchrony of movement. A school need not always imply

that all individuals are facing the same direction; social squid can swim

both backwards and forward.

Shoal: a (usually) larger grouping within which are contained subgroups

conforming to the definitions of swarm and school.

Pod (primary group): term confined to smallest social groupings of ceta-

ceans that remain intact for days or weeks.

Herd (secondary group): temporary (minutes or hours) aggregations of

primary groups of cetaceans.


In the rest of this review we will concentrate on social aggregation.Sociality means living in groups which, in turn, creates and intensifies twoopposing forces: on the one hand cooperation with conspecific neighbours,and on the other competition for local resources (Frank, 2007). The formercan increase efficiency and aid in competition with other groups; the lattermay promote conflict. This antagonism between cooperation and competi-tion is a recurring theme in studies of social aggregation. The title of Frank’s(2007) paper, ‘All of life is social’, reflects the fact that multicellularity arosethrough genetic aggregations and cellular symbioses. But what factors makesocial aggregation, as exemplified by fish or krill schools, such a conspicuousfeature of the pelagic zone? Before addressing this question we consider howsociality may have arisen.

2.1. Origins of sociality

Aggregation occurs in all phyla but where does sociality begin? In one sense‘all of life is social’ (Frank, 2007), in that multicellularity owes its existence toamalgamation and symbiotic cooperation of single-celled organisms.According to the argument outlined below, animals (or cells) must be able torecognize conspecifics to the extent that stable networks develop. How doanimals recognize each other? Recognition of self versus non-self must havearisen early in the evolution of multicellularity, perhaps to protect these cellu-lar aggregations against invasion by competing neighbours (Frank, 2007).There is evidence that olfactory cues discriminated by the major histocom-patibility complex are important in vertebrates (Krause and Ruxton, 2002;Villinger and Waldman, 2008). It has been reported recently that there is ananalogous system in invertebrates (Cadavid et al., 2004).

Recent work on social insects has led to the postulation that develop-ment of complex societies arose initially through natural group dynamics,i.e. not a genetic selection for particular traits favouring social behaviour buta tendency towards network development within groups (Fewell, 2003).Relatively simple connections between individuals in a group can createpatterns of behaviour of increasing complexity in the same way as simpledecision rules create complex behaviour in computer-generated aggrega-tions. If true, this suggests that networks are an essential precursor to socialbehaviour.

It has been suggested that the repeated co-occurrence of stable pairsmay have been an important prerequisite for the evolution of cooperativebehaviour and reciprocal altruism (Milinski, 1987; Croft et al., 2005).Relatively simple connections between individuals in a group can createpatterns of behaviour of increasing complexity. Organized societies occur-ring within many different taxa may have arisen through the agency ofthese simple local interactions self-organizing into global networks(Glance and Huberman, 1994; Fewell, 2003). It has long been recognized


that inclusive fitness operates within groups whose members share genes(Hamilton, 1964). However, it is now suggested that multilevel selectionoperates not just because members of groups are related but that denselyconnected networks exist within aggregations (Fewell, 2003). This per-mits rapid and efficient information transfer and flexible responses. It isimportant to explain that multilevel selection does not imply that somegroups are more successful (fitter) than others and contribute more groupsto the next generation. Instead group-level selection implies that the fit-test groups contribute the most individuals to the next generation.Behavioural traits possessed by individuals need to influence the behav-iour of others in the group to be relevant to group selection (Krause andRuxton, 2002).

Behavioural characteristics leading to social networks are not necessarilyrestricted to vertebrates. According to Webster and Fiorito (2001), sociallyguided behaviour, conforming to a framework developed for social verte-brates, can be found in a wide range of non-insect invertebrate phyla.However, among marine taxa, only Crustacea, Gastropoda and Cephalopodadisplayed behaviour typical of social learning, i.e. the acquisition of novelbehaviour due to observation of, or interaction with, a conspecific. Thismight indicate more sophisticated social behaviour in these groups comparedto ‘lower’ invertebrate animals.

2.2. Significance and benefits of social aggregation

The commonly stated benefits of group living are facilitation of foragingand feeding; reproductive facilitation (including sharing parental care inthe case of some cetaceans), protection from predators and energy saving(Wilson, 1975; Ritz, 2000; Krause and Ruxton, 2002). Other authorsadd maintenance of position in the environment (Clutter, 1969), habitatdefence (Hurley, 1977) and access to centralized information (Parrishet al., 2002). Benefits of synchronous breeding and release of youngwithin aggregations of mysids are described by Johnston and Ritz (2001).The benefits listed above are not divorced from one another; in fact anultimate advantage of aggregation may be energy saving in the broadestsense, i.e. efficiency in foraging, food capture, locomotion, protectionfrom predators, etc. It is likely that any adaptation that conserves energywill be favored by selection and fixed into the genetic blueprint (Ritz andSwadling, 2006). Thus if energy can be conserved at the same time asefficient food gathering and escape from predation, this will ultimately beadvantageous for the species. Cohen and Ritz (2003) found that mysidsin small uncohesive groups were more likely to expend energy (tailflips)when exposed to a threat in the form of a fish kairomone than those inswarms. A kairomone is a chemical substance released externally that ben-efits the recipient without benefitting the emitter. The energy saving


benefits of swarm membership by mysids were confirmed by Ritz et al.(2001) who found that larger swarms expended less energy than smallerones, which, in turn, saved more energy than un-aggregated individuals(Ritz, 2000). The conversion of these energy savings into enhanced fit-ness is assumed to follow, but empirical evidence is lacking to date.

The foregoing discussion raises some important questions, e.g. do dif-ferent forces promoting patchiness and social aggregation dominate at dif-ferent scales (Flierl et al., 1999)? Patterns apparent at large scale are notindependent of those at small scale and vice versa. Large populations ofindividuals can self-organize into pattern-generating aggregations (Parrishand Edelstein-Keshet, 1999; Parrish et al., 2002; Viscido et al., 2004).Many of the patterns seen in real-life schools and swarms are apparent incomputer generated models based on a few simple rules (see Boids: http://www.red3d.com/cwr/boids/ and Efloys: http://arieldolan.com/ofiles/eFloys.html). An unresolved question is ‘are all of these emergent patternsevolutionarily advantageous?’ Can we reconcile the short-term selfishnessof individuals with the maximum group benefit of maintaining a cohesiveunit? There is potential for great complexity of trade-offs and constant ten-sion determining decision-making. Examples can be found among krill inthe decision to migrate vertically, and whether it is advantageous to schoolwhile doing so (De Robertis, 2002; Burrows and Tarling, 2004).Silversides (Menidia menidia) changed their schooling behaviour accordingto light levels during periods of twilight, which appeared to be associatedwith predation threat and availability of food (Major, 1977). Verticallymigrating mesopelagic fish may apparently elect to form schools whenlight levels at night are high enough to favour visual predators but not oth-erwise (Kaartvedt et al., 1996; see Fig. 4.5 reproduced from their paper).This suggests there are some disadvantages to schooling, e.g. increased visi-bility to and attraction of the attention of predators; also decreased percapita share of food resources.

Active and social groups are those that ‘self-organize’ (characteristicgroup patterns that arise from decentralized behaviour), and maintaincohesion because the derived benefits outweigh the costs (Ritz, 1994;Krause and Ruxton, 2002). As defined in Box 4.1, social aggregationsinclude swarms, schools, shoals, pods and herds. Regardless of the termi-nology, social aggregations can be recognized by their coordinated move-ment, persistence in time, reactions of individuals to the group and by thefitness benefits provided by mutual attraction.

Animal aggregations occur where large numbers of individuals are tobe found gathered in close spatial and temporal proximity. Examples ofaggregations can be found in virtually all animal phyla (Parrish andEdelstein-Keshet, 1999; Parrish et al., 2002). Such aggregations may con-sist entirely of conspecifics; however, there are also many examples ofmultispecies aggregations, e.g. in fish (Allan, 1986), and crustaceans


http://www.red3d.com/cwr/boids/

http://www.red3d.com/cwr/boids/

http://arieldolan.com/ofiles/eFloys.html

http://arieldolan.com/ofiles/eFloys.html

(Ohtsuka et al., 1995), and even mixtures of the two taxa (McFarlandand Kotchian, 1982). Aggregations may form in response simply to thedistribution of resources in the habitat, but the term ‘social aggregation’refers specifically to groups of individuals that are brought together, eitherwholly or in part, by social attraction.

Social aggregations are an extremely common phenomenon with con-siderable ecological and economic importance. For example, well over halfthe world’s fishes, including the overwhelming majority of the commer-cially harvested species, form social aggregations at some stage during theirlives (Shaw, 1978). Theoretically, a social aggregation could consist at aminimum of two individuals (e.g. spawning cuttlefish); however, mostmarine social aggregations are substantially larger than this, indeed somemay often encompass billions of individuals (e.g. krill). The scale of suchaggregations can be dramatic � DeBlois and Rose (1996) reported shoalsof cod (Gadus morhua) of over 10 km in length, migrating groups of mullet(Liza aurata and L. saliens) stretch for over 100 km (Probatov, 1953), whileRadakov (1973) estimated the volume of some Atlantic herring (Clupeaharengus) shoals at up to 5 km3. The concept of the optimal group size hasbeen the subject of considerable theoretical debate (see Sibly, 1983;

Figure 4.5 Echogram showing abrupt changes in vertical distribution of krill, planktivor-

ous (Norway Pout) and piscivorous (Pearlside) fish in response to changes in light penetra-

tion. Reproduced from Kaartvedt et al. (1996).


Giraldeau and Gillis, 1985; Giraldeau and Caraco, 2000); however, fewempirical data exist to test theoretical predictions (Willis, 2008). In marineenvironments, many of the larger aggregations that may be observed areso-called free-entry systems, where group members have little or no ability(and indeed little incentive) to control group membership. In these cases,group sizes are highly dynamic and groups frequently split and reformaccording to the context, a property which has led to them being describedas ‘fission�fusion’ societies (Couzin, 2006). ‘Restricted-entry’ groups arefar less common in the marine environment, and the examples that existare all of vertebrates, perhaps most notably the small aggregations of coralreef fishes (Sale, 1971; Forrester, 1991; Whiteman and Cote, 2004) andsocial groups of cetaceans (Gowans et al., 2001; Lusseau 2003; Hartmanet al., 2008).

There exists considerable variation, both within and between species, insociality. Some authors draw a distinction between facultative and obligatesociality; however, such distinctions are somewhat arbitrary since the extentto which any individual manifests social attraction is likely to vary withontogenetic stage and with context (Ritz, 1994). Nonetheless, some ani-mals exhibit considerable stress if separated from conspecifics: Atlantic her-ring that have been experimentally isolated from conspecifics have beenreported to die as a result (Gerasimov, 1962). While this is an extremeexample, many social species do manifest stress-related changes in behaviourand/or physiology if isolated. For example, Ritz et al. (2003) showed thatheart rate of an individual Antarctic krill was high when isolated but slowedsignificantly when it was tethered at normal schooling distance from a con-specific and was presumably able to access social cues. The extent to whichmarine animals form social aggregations may also be highly dependentupon the environment. For example, in heterogeneous environments, shoalcohesion often decreases (Mochek, 1987). Many fish, amongst them cod(Gadus morhua), sergeant majors (Abudefduf saxatilis) and grey snappers(Lutjanus griseus), exhibit shoaling behaviour when in mid-water, but theshoals break up towards the bottom of the water column or when in near-shore areas (Pavlov and Kasumyan, 2000). Furthermore, shoals of fishcharacteristically break up at night as light intensity decreases (Higgs andFuiman, 1996).

While some species consistently form aggregations throughout theirlives, many others are more social during some stages of their life historythan others. The larvae of many pelagic fish typically do not begin toshoal until metamorphosis (Fuiman and Magurran, 1994). The importanceof the development of the central nervous system in relation to schoolingbehaviour in larval and juvenile striped jack (Pseudocaranx dentex) washighlighted by Masuda and Tsukamoto (1999), who reported the emer-gence of mutual social attraction among individuals at around 12 mm inlength. Interestingly, Antarctic krill first begin to show social attraction at


around the same length (Hamner et al., 1989), even though final size isvery different. The expression of strong social attraction towards conspeci-fics during the juvenile phase is a common pattern in fishes before becom-ing increasingly solitary as they grow (Pavlov and Kasumyan, 2000).Indeed, many marine species form aggregations during vulnerable early lifestages, since grouping behaviour is often suggested to confer valuable anti-predator benefits upon group members (Ritz, 1994). The opposite trend ismanifest in spiny lobsters where juveniles are typically solitary and adultsvery often live in groups (Ratchford and Eggleston, 1998). Butler et al.(1999) demonstrated that attraction to conspecific chemical cues in spinylobsters only occurs once individuals reach a given size.

The larvae (and indeed the adults) of many marine species aggregate asplankton; however, it is arguable whether this is to any great extent dueto social attraction. Banas et al. (2004) consider that the interactionbetween individuals in many zooplankton swarms (as opposed to schools)is of secondary importance. For example, Leising and Yen (1997) contendthat density of copepod swarms in their experiments was controlled byavoidance reactions to chance close-range encounters. The same authorsfound five species of copepod to be insensitive to proximity of conspeci-fics except at very close range. Extrapolating this view, in their modellingstudies, Banas et al. (2004) regard a swarm not as an interaction betweenindividuals, but between each animal and its local stimulus field. Ofcourse, this interpretation does not necessarily reduce the importance ofsocial attraction. Swarms can be social and density-dependent or non-social and density-independent. Most modelling studies of swarming/schooling behaviour are based on a resolution of forces of attraction,repulsion and alignment (Couzin, 2006). The degree to which individualsrespond to each other and over what distance is probably a function ofsensory capability but also the need to maintain a hydrodynamic ‘terri-tory’ (or flow field) that is an essential part of the feeding current, and bydistortion of which an individual gains information about approaches byother animals.

Social aggregation in general is less well studied in pelagic inverte-brates, despite their amenable sizes for experimental laboratory work. Thelabel ‘plankton’, with its connotations of passivity, has probably seriouslyhindered the study of social behaviour of ‘true’ zooplankton and micro-nekton (Ritz, 1994). The term plankton is not a taxonomic unit andencompasses a huge diversity of species and forms. It seems likely that aspectrum of capacity to manifest social behaviour exists in which mostcopepods would occupy one end, and live in swarms, with little socialinteraction between individuals except to ensure that empty space aroundthemselves is maintained. On the other end, mysids and euphausiidsexhibit a full range of social interaction with neighbours. Unfortunately,many authors still group the larger more active constituents, e.g. krill and


mysids, with the smaller ones, e.g. up to the size of most copepods, andascribe little in the way of social interaction. Alldredge et al. (1984)claimed that zooplankton swarms with a nearest neighbour distance(NND) of more than a few body lengths were rare in nature. Later results(Jiang et al., 2002) explain why this is true for copepods. Larger inter-individual distances of 7�14 body lengths ensure that hydrodynamicinteractions between neighbours are minimized so that feeding currentsand detection of nearby individuals are not compromised. Jiang et al.(2002) showed that copepods gain no energetic benefits when in closeproximity to conspecifics. In contrast, O’Brien (1989) showed that mysidand euphausiid NNDs were on average 1�2 body lengths, and energeticbenefits of swimming close to neighbours were demonstrated by Ritz(2000) and Patria and Wiese (2004), and communication benefits byWiese (1996).

The characteristics of aggregations of zooplankton are sometimes sug-gested to vary considerably between different species (Banas et al., 2004).Despite these assertions, there exist many similarities between the aggrega-tion behaviour of many different marine species. For example, Hamner(1985) and O’Brien and Ritz (1988) describe behaviour of swarms andschools of krill and mysids as being strongly reminiscent of fish schools.Escape manoeuvres of groups of the two taxa require a high degree of syn-chrony between individuals, probably requiring a combination of vision,chemoreception and mechanoreception. Furthermore, the possibility of anyclear distinction between the aggregation behaviour of copepods and thatof mysids and krill seems unlikely since even copepods (Labidocera pavo) canbe found forming schools (Omori and Hamner, 1982) which surelyrequires some inter-individual coordination. Differences in NNDs betweenaggregations of copepods and krill/mysids may be due to the differentfeeding methods, lack of any energetic benefit in close alignment in the for-mer and/or the differences in the reliance on vision. The eyes of mostcopepods and other non-Malacostracan crustacean zooplankters do nothave lenses and do not have an image-forming capacity (Eloffson, 1966).However, it is perhaps significant to note that Labidocera have ‘remarkableeyes’ (Omori and Hamner, 1982) that have a dorsal pair of spherical lensesserving a single mobile eyecup (Land, 1988). Land suggests that scanningmovements of the rhabdoms in the eyecup are concerned with visual detec-tion of conspecifics.

2.3. Structure and functions of social aggregations

From an evolutionary point of view, it is predicted that for aggregationsto persist, the benefits of group membership must, on average, outweighthe costs. The benefits and costs of grouping generally are dealt withextensively in other texts (Krause and Ruxton, 2002) and so here we


provide only limited pelagic examples to illustrate the commonality withnon-pelagic taxa and to emphasize social benefits and costs.

According to Flierl et al. (1999), many groups in marine pelagic sys-tems, with the exception of marine mammals, seem to be large, relativelytransient and loosely knit. This is not necessarily the case for fish schoolsor schools of mysids and euphausiids, where the aggregation can be long-lived even if membership changes over time (O’Brien, 1988; Parrish andEdelstein-Keshet, 1999). Social grouping implies some appreciation byindividuals of their membership of the group and its consequences. Inother words, individuals respond to conspecifics in a density-dependentmanner (Flierl et al., 1999; Burrows and Tarling, 2004; Hensor et al.,2005; Grunbaum, 2006). A group formed initially by random encountersmay grow by density-dependent interactions between members witheventual size being determined by mean payoff to individuals in thegroup (Parrish and Edelstein-Keshet, 1999). An example might be infood acquisition whereby all group members are disadvantaged if thegroup outgrows the food resource and individual rations decrease. Socialorganization then is an emergent property (group characteristic arisingfrom decentralized interactions) that develops from the network structure(see below).

Research on the influence of the social group on metabolic processesincluding growth, feeding efficiency and energy use has been neglected(Ritz, 2002). Ritz (2000) predicted that growth rate in schools or swarmsof krill and mysids was likely to be much greater than in isolated indivi-duals or small groups; however, it was some years before evidence wasforthcoming. Atkinson et al. (2006) showed that when growth of freshlycaught krill is recorded by the instantaneous rate method, a techniquethat allows insight into the recent social behaviour, the maximum valueswere much higher than most of the previously published rates. Thesehigh rates are likely due to the benefits of feeding within a school imme-diately before capture. Analysis by Ritz (1997) showed that food capturerate by mysid swarms of different sizes varied significantly when per capitafood availability was constant and inhomogeneous. Furthermore Ritz(2000) noted that ingestion rates of Antarctic krill measured in freshlycaught animals by faecal egestion were 3 times higher than those fed inlaboratory tanks (Pakhomov et al., 1997). He suggested that the differencewas due to the former having fed within aggregations immediately beforecapture, whereas the krill in laboratory tanks rarely form aggregations(Kawaguchi et al., 2010). It appears that krill (and mysids) not only saveenergy in social groups but may also feed more efficiently (see also Ritz,1997). The latter is a well-established principle in fish. For example, fishare known to forage more successfully on spatially variable food patcheswhen searching in social groups (Ryer and Olla, 1992). The issue of foodcompetition in aggregations may be more acute for vertebrates than for


invertebrates. Ritz (2000) made the point that fish in schools in the seaare more likely to be food limited than particle-feeding coastal mysids.

Many animals that aggregate socially have been demonstrated to saveenergy while aggregated. Birds (pelicans) in V-formation showed a signifi-cant decrease in heart rate and the energy saving may have been becausethey were able to spend more time gliding and less flapping (Weimerskirchet al., 2001). Fish (trout) have been shown to slalom between experimen-tally generated vortices (Karman gait) using only their anterior axial mus-cles. By synchronizing their body kinematics in this way, they may usevery little energy and gain a hydrodynamic advantage beyond that gainedby simple drafting (Fig. 4.6). Thus it is possible that any favourable hydro-dynamic consequences generated by the aggregation itself or its interactionwith surfaces (ground effect) could be exploited to save energy.

2.3.1. Patchiness in zooplanktonMany authors have documented the uneven or patchy distribution ofzooplankton in marine and freshwater (Hardy and Gunther, 1935; Steeleand Henderson, 1981; Folt and Burns, 1999). It has taken several decadesfor the contribution of plankton behaviour in generating this patchiness

Figure 4.6 (A) Time series of body outlines of trout superimposed on vorticity and

velocity vector plots of the wake produced by a cylinder located in the flow (left to right).

(B) Midlines for seven consecutive tailbeats. (C) Phase between body and vortices where

180� represents slaloming in between vortices and 0� and 360� represent vortex intercep-

tion. After Liao et al. (2003a,b).


to be fully appreciated (Ritz, 1994; Folt and Burns, 1999). These authorsdescribe general principles of invertebrate aggregations, and both papershighlight the importance of biological factors in creating and maintainingplanktonic aggregations. By contrast, Siegel and Kalinowski (1994) pro-vide Table 4.2 listing suggested causes of aggregation by Antarctic krill. Itis particularly noteworthy for how few authors credit behaviour as a pri-mary contributor to aggregation.

Nonetheless, the conclusion that social behaviour is the main driver ofvariability in small-scale density distribution of Antarctic krill is stronglysupported by a comparison of variance spectra of phytoplankton, temper-ature and krill (Weber et al., 1986; Verdy and Flierl, 2009). At small scales,the krill spectrum is flatter than the others indicating that factors otherthan environmental ones are generating the observed patchiness in densitydistribution (Fig. 4.7).

Folt and Burns (1999) list four behavioural mechanisms that can resultin zooplankton patchiness: (1) diel vertical migration, (2) predator avoid-ance, (3) food finding and (4) mating behaviour. Genin (2004) gives fivefurther mechanisms contributing to patchiness by which plankton, micro-nekton and fish can become aggregated above abrupt topographic fea-tures, all driven by (1) ocean currents where long residence upwelledwater enriches primary production which propagates up the food web;(2) daily accumulations where topography blocks morning descent ofzooplankton, e.g. over seamounts; (3) behavioural response of zooplank-ton to upwelling currents; (4) behavioural response to downwelling cur-rents; (5) enhanced population growth by residents due to currentamplification driven by abrupt topographies. These mechanisms do notnecessarily imply any social interaction between individuals but may pro-vide opportunities for closer attraction. Hamner (1988) observed thatalmost any animal behaviour can generate patchiness, but more sophisti-cated behaviour is required for social aggregation.

Recent work by Genin et al. (2005), using sophisticated multibeamacoustic equipment, has demonstrated that zooplankters,5 mm activelyswim against upwelling and downwelling currents in an effort to maintaindepth. In this way, they aggregate at fronts. The value of maintaining depthin this way is not yet clear, but may serve to keep them within food-richzones and prevent them straying into less favourable depths. Any behaviourthat actively or passively leads to individual distributions becoming clumpedcould result in a tendency to remain in a group once the many benefits aremanifested. Swimming against currents of up to 1 cm s21 at rates of .10body lengths s21 (Genin et al., 2005) is energetically expensive butmight be less so if the individuals formed aggregations. Ritz (2000) andRitz unpublished observation have shown that mysid swarms expendbetween 3 and 7 times less energy than small groups of individuals swim-ming uncohesively. A possible explanation is that swimming action by


Table 4.2 Suggested causes of aggregation in Antarctic krill, Euphausia superba

Author Main causes and environmental

limits

Region

Marr (1962) Inflow of the current branches

from the Weddell Sea

South Georgia

Maslennikov (1972) Presence of system currents

flowing in opposite direction

South Georgia

Wolnomicjski et al.

(1978)

Occurrence of eddies South Georgia

Rukusa-Suszczewski

(1978)

No elements of environmental

limits directly influence

distribution

West Atlantic

Mauchline (1980) Environmental parameters and

social behaviour

Theoretical

consideration

Kils (1979) Oxygen limits Laboratory

experiments

Wilek et al. (1981) No elements of environmental

parameters T, S, O2, nutrients,

phytoplankton

West Atlantic

Stein and Rakusa-

Suszczewski (1984)

Topography of bottom, which

influences direction of water

masses and hydrodynamic process

Bransfield

Strait

Kalinowski and Witek

(1985b); Witek et al.

(1988)

Hydrodynamic processes and social

behaviour

West Atlantic

Hampton (1985) Hydrodynamic forces, thermocline Indian sector

Weber and El-Sayed

(1985)

No physical and chemical

variability

Indian sector

Weber et al. (1986) Phytoplankton (2�20 km spatial

scale)

Kalinowski and Witek

(1985a)

Presence of daylight for patch

formation

West Atlantic

Loeb and Shulenberger

(1987)

Temperature (infusion of cold

water) and wind direction

Elephant

Island

Everson and Murphy

(1987)

Hydrodynamic processes (passive

current-borne movement)

King George

Island

El-Sayed (1988) Phytoplankton at scales 2�20 km Southern

Ocean

Murphy et al. (1988) Environmental phenomena and

food availability

Southern

Ocean

Priddle et al. (1988) Environmental phenomena and

active reaction for

South Georgia

disadvantageous environmental

conditions

Bransfield

Strait

(continued)


cohesive groups generates favourable currents that could be exploited bymembers to reduce the cost of forward propulsion and/or to minimize therate of sinking by relatively dense crustaceans. Patria and Wiese (2004)showed that swimming krill (Meganyctiphanes norvegica) generated vortex

Table 4.2 (continued )

Author Main causes and environmental

limits

Region

Maslennikov and

Solyankin (1988)

Variability of hydrological

conditions

West Atlantic

Makarov et al. (1988) Stable eddies and influence of

water masses

West Atlantic

Reproduced with permission from Siegel and Kalinowski (1994).

Figure 4.7 Mean spectral plots for krill, fluorescence and temperature. . After Weber et al.

(1986); http://plankt.oxfordjournals.org/content/14/10/1397.full.pdf


http://www.plankt.oxfordjournals.org/content/14/10/1397.full.pdf

rings behind the tail that could be exploited by following krill for this pur-pose. Note that this benefit would only make sense if the krill were inschool formation, i.e. individuals polarized and neighbours sufficiently closeto perceive and take advantage of currents moving in the same direction asthemselves. Ritz (2000) showed that energetic benefit accrued in mysidswarms, i.e individuals, was not polarized. In this case it may have been thepowerful downdraft generated collectively by the group that induced anupdraft at the margins of the swarm. This could have been used to counter-act sinking. Buskey (1998) reported that mysids frequently changed verticalposition in the school while maintaining their horizontal position in a cur-rent. Energetic benefit may derive from specific relative positions of indivi-duals in groups (Parrish and Edelstein-Keshet, 1999; Svendsen et al., 2003).Thus it would be logical for individuals not to adopt fixed positions withinthe group but to continually make excursions in order to maximize what-ever benefits accrued at any given moment (Ritz, 1997).

2.3.2. Diel vertical migrationDiel vertical migration (DVM) is generally held to represent a trade-offbetween the functions of food gathering and avoiding predators(Kaartvedt et al., 1996). The typical pattern is a dawn descent into deeper,darker levels during the day and an ascent towards the surface beginningaround dusk. Ritz (1994) was the first to suggest that since social aggrega-tion serves the same purposes, species that form swarms or schools mayfind DVM redundant. This could account for the confusing and some-times contradictory evidence for vertical migration in Antarctic krill(Miller and Hampton, 1989). It appears from work by De Robertis(2002) that there are situations in which the normally social euphausiidEuphausia pacifica performs DVM but does not form social aggregations,i.e. one is redundant in the presence of the other. De Robertis suggeststhat DVM may be favoured over social behaviour in open-water zoo-plankton because of well-developed vertical gradients in predation risk.However, Antarctic krill are strongly social in open-water situations andexhibit variable DVM (O’Brien, 1987; Daly and Macaulay, 1991). Furtherevidence is supplied by Kaartvedt et al. (1996). The DVM behaviour offish and krill (mainly Thysanoessa inermis) varied according to the lightconditions in upper shelf waters and the predation risk. In brief intervalsat dawn and dusk, known as ‘anti-predation windows’, there is sufficientlight for planktivorous fish to locate prey, but not enough to render thesefish vulnerable to piscivores. ‘Anti-predation windows’ may occur at othertimes due to the passage of fronts bringing water of higher turbidity. Atthese times planktivorous Norway Pout may migrate vertically to forageon krill, whereas when the light penetration through overlying waters isgreater, the planktivores remain in safer, deeper layers but out of contactwith their prey. It is unclear whether the planktivores migrate in schools.


These examples serve to illustrate the great flexibility of behavioural strat-egies and the fitness benefits they may confer.

2.4. Association patterns within aggregations

Association patterns within and between groups of marine animals are sel-dom entirely random. At a very basic level, multispecies aggregations arevery often dominated by a small number of species (Krause et al., 1996;2005), suggesting either species-level recognition and an association prefer-ence for conspecifics, or that this emerges passively from similarities inactivity patterns within but not between species (Conradt and Roper,2000). Beyond this, group membership may be structured by phenotype;fish shoals for example frequently assort by body length, species, colour andparasite load (Krause et al., 2000). The fact that invertebrate aggregationscommonly consist of a narrow individual size range (Watkins et al., 1992)suggest that they too possess a certain level of cognitive ability permittingpairing with other morphologically similar individuals (Wilson andDugatkin, 1997). Watkins and Murray (1998) reported that biological char-acteristics between adjacent swarms of Antarctic krill (Euphausia superba)vary in individual size, maturity stage, moult and feeding state. Young et al.(1994) show that Daphnia clones have differing swarm-forming tendenciesand suggest that clone-mates can recognize each other probably by chemi-cal cues. This appears to be a fruitful subject for further research.

Any advantages of group fidelity may be partly related to the fact thatindividuals associate with others of similar phenotype. In addition to this,research has shown that individuals manifest social association preferencesfor some conspecifics over others, in the absence of any clear phenotypicdifferences (Milinski et al., 1990; Ward and Hart, 2003). ‘Familiarity’, asthis subgrouping phenomenon is known, acts to enhance the benefits ofshoaling, further reducing the per capita risk of predation (Chivers et al.,1995) and improving foraging performance (Ward and Hart, 2005). In con-sequence, it might be expected that this would act in concert with othermechanisms (Conradt and Roper, 2000) to stabilize association patternsover time. Yet the data gathered from field studies on this topic are equivo-cal; it appears that there is no general rule regarding shoal fidelity amongfree-ranging schooling fish. Evidence exists for shoal fidelity in three-spinesticklebacks, Gasterosteus aculeatus (Ward et al., 2002), and Klimley andHolloway (1999) reported the co-occurrence of individual yellowfin tunain time and space. By contrast, banded killifish (Fundulus diaphanus)appeared to show no consistent shoal fidelity (Hoare et al., 2000) despiteshowing a strong tendency to assort with fish of similar phenotypic charac-teristics. Of the few other studies in which the movement of marked fishamong and between shoals have been followed, Helfman (1984) found lowshoal fidelity among yellow perch (Perca flavescens), and Hilborn (1991)


showed that schools of skipjack tuna (Katsuwonus pelamis) mixed rapidly andwere not composed of the same individuals for more than a few weeks (seealso Willis and Hobday, 2007). There are even fewer examples of experi-ments testing shoal fidelity in invertebrate aggregations, although Twininget al. (2000) demonstrated homing behaviour and shoal fidelity in a mysid(Mysidium gracile).

On a larger scale, studies of migrating fish populations do suggest a ten-dency for stable association patterns to develop (McKinnell et al., 1997; Hayand McKinnell, 2002). McKinnell et al. (1997) reported that steelhead trout(Oncorhynchus mykiss) form long-term associations at sea throughout theirmigration. Furthermore, Hay and McKinnell’s (2002) remarkable study ofthe movements of more than half a million Pacific herring (Clupea pallasii)over a period of 14 years concluded that individuals formed stable temporaland spatial associations. Nonetheless, it would be difficult to conclude thatfamiliarity alone is entirely responsible for the observed patterns. Groupfidelity in fish migrations may be influenced by any of several differentmechanisms, including kin- or population-specific recognition (Quinn andTolson, 1986), activity synchronization (Conradt and Roper, 2001), popula-tion-specific migration traditions (Warner, 1988) or pheromonal attraction(Baker and Montgomery, 2001). Alternatively, because migrating fish tend toremain in large, temporally stable shoals, as opposed to the smaller and looseraggregations characteristic of many of the shallow water species studied bybehavioural ecologists, patterns of association in migrating fish may beexplained simply by long-term shoal cohesion.

2.5. Sensing the behaviour of neighbours

The sensory dimensions of social aggregation include the sensory basis ofgroup formation (about which little is known), and the behaviour of indi-viduals, and hence the group, under different contexts such as feeding andpredation threat. Understanding the sensory basis of behaviour is one ofthe first steps to understanding the underlying neural algorithms drivingthe behaviour in individuals and hence the ‘behaviour’ of the aggregation.The focus of this section will be to review the sensory basis of fish school-ing, as one of the most important, and highly coordinated, examples ofsocial aggregation. This discussion will then be extended to consider thesimilarities and differences between fish schooling, and invertebrate aggre-gations, and briefly the sensory communication within marine mammalaggregations.

The sensory basis of fish behaviour has been reviewed in Montgomeryand Carton (2008). Vision, lateral line and hearing are the strongest can-didates for schooling coordination although olfaction may also play animportant role. Olfaction is certainly implicated in schooling coordinationfor group spawning. Pheromones play a role in maturation timing, and


spawning synchrony by ensuring that male spawning readiness is linked tofemale maturation and spawning. Observations of spawning behaviour inthe sparid (Pagrus auratus) also describe males following females prior toand during their vertical spawning run (JM, personal observation).However, for the more demanding expressions of schooling coordination,the high temporal and spatial resolution characteristics of visual andmechanosensory (lateral line and acoustic) senses are required. The sen-sory basis of schooling behaviour during feeding will first be considered,followed by the more challenging context of tight schooling coordinationin response to predator threat. Both examples provide interesting casestudies in the conflicting demands of mid-water camouflage and schoolingcommunication and coordination.

In most coral and temperate reef habitats, schools of plankton-feedingfish form over reefs. The reefs provide hydrodynamic conditions that canconcentrate and transport plankton to the schools, and in some cases thereef also provides shelter from predators of the fish. This discussion con-centrates on those species such as mackerel (Trachurus? sp.) that are oftenreef associated, but that depend more on schooling rather than reef shelterfor predator defence (a parallel example exists for mysids; see Flynn andRitz, 1999). These species also depend on the strategy of reflective cam-ouflage to reduce their visibility to predators in open water (Denton,1970, 1980; Johnsen and Sosik, 2003). Clearly camouflage and visualcommunication are conflicting requirements; visual communication sig-nals must to some extent undermine effective camouflage. Reflectivecamouflage also only operates effectively within strict physical constraints,including fish orientation. When the camouflage is compromised bychange in orientation, or a sudden turn, the resulting visual stimulus canprovide a basis for visual communication.

A mirror in the water column will be difficult to see, only if it is verticaland far enough away from the surface to be in a vertically symmetricallight field. Only under these conditions will the reflected light match thebackground space light and reduce the contrast between target and spacelight and hence the effective visual range of detection. The reflective cam-ouflage of pelagic fishes operates on a similar principle (Denton, 1970,1980). Reflective platelets in the skin and the scales of the fish are orientedparallel to the dorso-ventral axis of the fish so that when the fish is in itsnormal orientation they form an array of vertical mirrors. This reflectivecamouflage works best when the fish is horizontal in the anterio-posterioraxis, with its dorsal surface pointing to the most intense downwelling light.Deviations from this position increase the target contrast against back-ground space light for an observer (Dare, 2008). The relevance of theseconsiderations for social aggregation is that the demands for effective forag-ing can over-ride camouflage and provide schooling conspecifics withvisual signals that convey information on feeding opportunity and success.


Reflective camouflage is a strategy used by many larger mobile (hencemuscular) fish, but zooplankton (including ichthyoplankton) typicallyemploy an alternate strategy based on transparency. Transparent prey are bydefinition hard to see; however, sighting of transparent prey can beimproved by viewing them just beyond angles in Snell’s window (Janssen,1981). Snell’s window is a phenomenon by which an underwater viewersees everything above the surface through a cone of light of width of about96�. This phenomenon is caused by refraction of light entering water andis governed by Snell’s law. The area outside Snell’s window will either becompletely dark or will show a reflection of underwater objects (Janssen,1981). Tilting up by approximately 40� to achieve this forward sightingline improves feeding success, but necessarily increases the predator fishcontrast against the background and hence visibility to conspecifics andpredators. Moreover, the slight flaring of the gills and opening of themouth associated with suction feeding also provides a clear visual signal, orlight flash, to other school members conveying feeding success. In essence,by attending to visual contrast and flashes, members of a school can gaininformation on which parts of the school are feeding and how successfully.It is debatable as to whether this represents a signal in the sense of beingpurposely sent, but if one was looking for a benefit to the sender, the shiftof the school towards favourable feeding areas may provide some ‘safety innumbers’ advantage for them. Finally, from a sensory perspective thisbehaviour is almost exclusively visually mediated. The polarization of theschool facing into the current (positive rheotaxis) would likely be visuallymediated, though maybe with a lateral line component (Montgomeryet al., 1997), but the dynamics of schooling coordination in relation to tar-geting food patches would be exclusively mediated by visual signals. Bycontrast, the tight coordination of schools under predation threat requiresmore complex multimodal sensory communication.

Under the threat of predation, observations and studies of complexschooling behaviour (Pitcher, 1993; Pitcher and Parrish, 1993) tend toevoke descriptions of the school as a ‘super organism’. Discrete ‘beha-viours’ of the school are observable such as ‘splits, vacuoles and flashexpansion’. The sensory basis of these ‘school behaviours’ is extremelyhard to study, but is almost certainly due to the sensory detection of anattack by the fish on the ‘front line’ via the visual looming stimulus of thepredator, and/or the associated pressure pulse of a lunging strike. Thebehaviour of the school will result from the way in which this informa-tion propagates into the school, both directly, and as a result of the avoid-ance response of the ‘front line’ fish. Thus, communication betweenneighbours is likely central to the schooling response to attack, but it isalso key to understanding the schooling coordination under the threat ofpredation. For this latter case, there is good experimental evidence andtheoretical reasoning to support the active involvement of both vision and


the octavolateralis sensory systems (lateral line and hearing) (Partridge andPitcher, 1980; Gray and Denton, 1991; Montgomery et al., 1995;Faucher et al., 2010).

The main sensory requirements for coordination relate to initiating andmaintaining the close association of individuals, combined with highly effec-tive collision avoidance. We know rather little about the sensory mechanismsunderpinning initiation of tight schooling formations. The perception ofthreat by individual members of the school and the communication of thatthreat may simply be a function of an abrupt change in direction of those fishaware of the threat and a reduction in NND in that part of the group. Anabrupt change in behaviour in one part of the school could propagatethrough the school visually, but it is also feasible that acoustic communicationis involved. This could be both passive on the part of the signaller, resultingsimply from the abrupt turn or acceleration (Gray and Denton, 1991;Denton and Gray, 1993), but could also include active acoustic communica-tion. Some schooling fish such as mackerel and tuna produce sound throughstridulation of the gill rakers; however, the behavioural context of this soundproduction is not known (Allen and Demer, 2003). Once initiated, themaintenance of close NND is mediated by both visual and lateral line stimuli.Partridge and Pitcher (1980) have shown that blinded fish can still swim information and Faucher et al. (2010) have shown that in some fish the lateralline may be necessary for tightly coordinated schooling behaviour. Liao(2007) also provides an excellent discussion of the theoretical rationale for ahydrodynamic basis to school structure. The prediction is that a fish locatedbehind and in between two preceding members of the school can takeadvantage of the average reduced velocity associated with the thrust wakes ofthose ahead. In effect, fish in schools can benefit from flow refuging (exploit-ing regions of reduced flow) and vortex capture (harnessing the energy ofenvironmental vortices). Direct experimental determination of vortex cap-ture, associated energetic benefits and its sensory basis have not been done forschooling fish. However, individual fish swimming in a flume have beenshown to use lateral line information to position themselves in an energeti-cally favourable position behind a cylinder (Montgomery et al., 2003), and toentrain to shed vortices from a bluff object in the flow (Liao et al., 2003a,b).Figure 4.6 reproduced from Liao et al.’s paper illustrates this.

Visual communication in schooling has been extensively studied byRowe and Denton (1997) and Denton and Rowe (1994, 1998). Their anal-ysis is that the same substrate of reflective surfaces that provides for mid-water camouflage, supplemented by additional reflective and non-reflectivesurfaces (such as the double yellow reflective dots on the tail: see Fig. 4.9),can provide strong communication signals to nearest neighbours.

Additional reflective and non-reflective surfaces include highly silveredpatches on the tail, the dorsal lateral line which is a non-sensory reflectiveopen canal (Rowe and Denton, 1997) and the bands of reflection that


can mask underlying black stripes (Denton and Rowe, 1998). The com-bination of light reflection and colour changes can provide large changesin appearance for relatively small changes in roll, pitch and yaw. Thesechanges could in theory signal a fish’s movement and/or position relativeto its neighbours. These visual communication signals are thought tocombine with hydrodynamic and acoustic stimuli to provide for theimpressive collision avoidance capability of schooling fish.

Crustacean species such as Antarctic krill (Euphausia superba) and mysidsalso undertake coordinated swimming in formation (O’Brien, 1988; Patriaand Wiese, 2004; Kawaguchi et al., 2010). Wiese (1996) has reviewed theavailable information on the role of vision and mechanoreception in thecontrol of schooling in krill. He concludes that the evidence strongly sup-ports a mechanosensory basis for control of this behaviour. Even thoughblinded krill were unable to school (Strand and Hamner, 1990), this seemsto be simply a result of the inability to orient the body to a fixed verticalaxis in space i.e. to the axis of light from the surface. In contrast, fish cancontinue to school with loss of either vision or lateral line but not whendeprived of both (Partridge and Pitcher, 1980). Yen et al. (2003) and Patriaand Wiese (2004) have also described the vortex wake behind a tethered

Figure 4.8 School of Trachurus novaezelandiae showing double yellow reflective dots on

tail. Photo by John Montgomery.


Figure 4.9 (A) Communities and subcommunities in a dolphin social network. Vertices

shaded in black are all part of one community, while all other vertices are part of the sec-

ond community. The second community is subdivided into three subcommunities repre-

sented by white, light grey and dark grey shading. (B) Social network of a population of

guppies. All guppies from two interconnected pools were collected, marked and released.

Over the next 2 weeks approximately 20 shoals were captured daily and fish that belonged

to the same shoal were connected in the network. (A) After Lusseau and Newman (2004);

(B) after Couzin et al. (2006).


krill and the latter authors demonstrated the ability of antennular sensors todetect these oscillations at normal schooling distances and entrain a syn-chronous pleopod beat in a following animal. From these and other studies(e.g. Wiese and Ebina, 1995), it appears that the vortex wake generated byswimming krill could be used by followers both as a source of information,both hydromechanical and chemical, about the neighbour ahead, and alsoas a means of reducing their energy expenditure.

Acoustic signals travel much better than visual ones in the ocean, so itis not surprising that highly mobile marine mammals have exploited thiscommunication channel for social purposes (Tyack, 2000). Social com-munication within and between groups of delphinid cetaceans includesnot only vision and tactile stimulation, but also sound production in theform of narrow-band, frequency-modulated signals (whistles) (Dudzinskiet al., 2002). These types of signals are relatively easily localized, withdirectional characteristics which, together with their variability andpower, make them particularly suited to contact calls, the pod-specificdialects of resident Orcas or the signature whistle of bottlenose dolphins(Tyack, 2000). Echolocation by clicks, on the other hand, is thought tobe used more for foraging and navigation (Dudzinski et al., 2002).Communication among groups of Spinner dolphins, which feed mainlyat night, may differ from this general pattern. Apparently these dolphinsdo not use whistles while hunting for prey (Benoit-Bird and Whitlow,2009). Instead they used a series of clicks with the highest click rates justprior to foraging. The authors suggested that this may be a strategy tolimit communication only within the group and to avoid betraying thelocation of rich food resources to other predators (e.g. tuna) which canalso hear whistles but not clicks.

2.6. Social networks

Recent application of social networks analysis to the study of animal behav-iour and ecology has allowed novel and intriguing insights into populationsof aggregating animals (Croft et al., 2008). As Croft et al. (2005) note‘Social network theory can help bridge the gap between interactions at thelocal and global level and provides a framework for the study of sociality’.In a social network analysis a graph is often used to describe the interactionsbetween individuals. These individuals are represented as nodes in thegraph, and the lines joining them represent social ties. Most nodes are con-nected by at least one short path, and nodes in the network with a highnumber of connections are known as hubs. One class of social network, theso-called small-world network, is characterized by a comparatively shortaverage path length between nodes and a large number of hubs or ‘cliques’of interacting individuals. These qualities are important in facilitating thespread of such things as information, genes or even disease across


population. Such small-world networks (Watts and Strogatz, 1998) existwithin large fish aggregations (Croft et al., 2005; Couzin et al., 2007), birds(Ballerini et al., 2008) and in dolphin pods (Lusseau, 2003; Lusseau andNewman, 2004) (Fig. 4.9). The implications of these small-world networks,in information transfer and speed of spreading of disease, are now startingto be seriously studied at a range of levels of biological organization, fromthe individual to the population, in marine mammals.

Arguably the main strength of this method is to examine how individualinteractions scale to structure population-level processes, which allows us toexamine self-organization, and potentially the transmission of genes, infor-mation and disease (Watts and Strogatz, 1998; Latora and Marchiori, 2001;Cross et al., 2004). In the context of the present review, social network anal-ysis allows us to dissect aggregations to look at the structures within, particu-larly how the constituents of an aggregation interact with one another.Many of the aggregations discussed here are examples of ‘fission�fusion’groups (Couzin et al., 2006) where, as the name suggests, aggregations splitinto smaller groups at certain times before later coalescing into larger aggre-gations once more (Pearson, 2009). Aggregations, therefore, are very oftenmade up of a mosaic of smaller functional subgroups, a phenomenon thatmay be elucidated using social network analysis. For example, a study ofsocial networks in bottlenose dolphins (Tursiops truncatus) reported that thepopulation of these animals off the east coast of Scotland was composed oftwo largely separate social units (Wiszniewski et al., 2009).

Indeed, most of the work carried out on social networks in the marineenvironment has focused on pinnipeds (Wolf et al., 2007) and cetaceans(Slooten et al., 1993; Chilvers and Corkeron, 2002; Ottensmeyer andWhitehead, 2003; Lusseau, 2003). This may be easier in these species asthe individuals are large, and can be recognized via photo studies. Suchwork has greatly extended our knowledge of the function of cetaceansocial groups in particular. For example, a study by Lusseau and Newman(2004) identified different sex and age-structured social patterns in a NewZealand dolphin population, and perhaps most interestingly, reported theexistence of key individuals (‘brokers’) within the population that linkedseparate subgroups (Fig. 4.9A). Such findings are important not only forour understanding of the social dynamics of these animals, but behaviour-al studies of association patterns and social networks in these animals offerinsights that may be used to inform key management and conservationdecisions in marine animals (Williams and Lusseau, 2006; Higham et al.,2009; Williams et al., 2009). In the coming years, it is hoped that socialnetwork analysis will enable us to gain greater understanding of the socialdynamics of a broader range of marine animals.

Complex biological structures such as social groups consistently showattributes of networks that have non-random systems of connectivity.Several authors point to the fact that emergent properties are common to


groups of simple, identical units even non-biological ones, e.g. moleculesand spin magnets, and caution that appearance of pattern may not neces-sarily signify adaptive behaviour (Parrish and Edelstein-Keshet, 1999:Parrish et al., 2002). Others stress that an acceptance of social groups asnetworks is an important preliminary to an understanding of fitness ofindividuals and groups (Fewell, 2003). Wilson and Dugatkin (1997) showthat assortative interactions within or among groups can create conditionsof highly non-random variation providing material for group selection.The idea of group selection was an unthinkable and heretical conceptuntil just a few decades ago (Wilson and Dugatkin, 1997). A more mod-ern view is that selection operates on a nested hierarchy of units (seeWilson and Sober, 1994, for a review of this concept). As noted inSection 1, group selection implies that the fittest groups contribute notmore groups to the next generation, but more individuals which are them-selves predisposed to form successful groups.

3. Technology Breakthroughs in Experimental

and Observational Methods

Aggregative behaviour is of profound importance to both the ecologyand economical exploitation of the pelagic environment, but the remote-ness and opacity of the environment has limited the observations thatcan inform behavioural understanding. With regard to exploitation theselimitations may have, in fact, prevented the complete over-exploitation ofimportant commercial fish species such as tuna (Sibert et al., 2006), andnear elimination of other species such as great whales (Clapham and Baker,2003; Roman and Palumbi, 2003). However, technology currently avail-able considerably improves our observational abilities, so due caution toavoid continued non-sustainable exploitation needs to be considered. Inthe following subsections we review insights on social aggregation thathave been generated using modern technology including (i) video andmotion analysis software, (ii) optical plankton counters, (iii) acoustics and(iv) electronic tagging.

3.1. Video and motion analysis software

3.1.1. Historical useIn earlier sections we argue that aggregative behaviour is of profoundimportance to the ecology and economy of the pelagic environment.Aquatic animals aggregate for a variety of reasons including reproduction,defence against predators, food finding and energy savings (Ritz 1994,


2000). Classically, research has focused on the attributes of a group as awhole, with little consideration given to individuals within the group.Recent studies highlight the fact that it is the behaviour of individualsthat collectively determine the behavioural trajectory of the aggregation.Detailed analysis of behaviour of individuals and small groups can thusilluminate the behaviour of the much less tractable larger aggregations(Parrish and Edelstein-Keshet, 1999; Banas et al., 2004). Within anyaggregation relative positions of individual members are not static, butusually change constantly, moving from the centre to the edge and backagain (Hamner and Parrish, 1997). Until recently studying individualswithin the context of the entire swarm or school was difficult, particularlyfor small planktonic animals, and researchers generally concentrated eitheron analysing behaviour in the two-dimensional plane or examiningswarming/schooling individuals under artificial, laboratory conditions.

The common first step in studies of social behaviour in aggregationsis to capture images of the aggregation. Video and motion analyticaltools were first tested in the laboratory, with isolated individuals incontrolled situations. Behaviour of pelagic crustaceans has been recordedin three dimensions using a variety of photographic and video methodsto examine mating (Doall, 1998; Strickler, 1998), escape responses(Buskey et al., 2002) and attack volume (Doall et al., 2002). However,these were usually measured in detail under laboratory conditions wheresmall groups or individuals were isolated from the rest of the aggrega-tion often in a small volume of water. Group attributes in the form ofNNDs, bearings and angles of elevation have been measured using stillphotographs (O’Brien et al., 1986), but these methods provided onlyapproximations of parameters such as velocity. Obtaining three-dimensional trajectories of specific individuals for extended periods wasalso difficult, with data typically generated under highly artificial condi-tions (Parrish et al., 2002).

In the past, most analysis of objects in three-dimensional space(photogrammetry) employed film cameras because only these camerascould provide sufficiently high image quality and geometric reliability.However, digital still cameras and digital video cameras now offer imageresolution approaching that of film and, more importantly, provide animaging geometry that is sufficiently stable for photogrammetric pur-poses. Modern digital video cameras offer the possibility of capturingstereo-video and obtaining accurate three-dimensional measurements ofmoving targets (Osborn, 1997). These early studies also tracked a lim-ited number of individuals, with individuals identified by eye in sequen-tial images. Thus, while a focus has been on the spatial and temporaldescription of aggregations, these early studies have also shown thatenergetic benefits to individuals in aggregations often cannot be inferredfrom analysis of isolated individuals.


3.1.2. Technological developments in video and image analysisIncreased computing power, digital storage capability and modern digitalcameras have allowed an order of magnitude increase in analytical powerand offer new insights into the spatial and temporal mechanistics ofaggregation. A selection of technological aids used in studying zooplank-ton aggregations was provided by Folt and Burns (1999). Here we updatethis list and focus on technology that is particularly applicable to, or wasdeveloped for the purpose of, recording social aggregative behaviour.

A second challenge has been to develop technology to allow observa-tions of aggregations at sea. Kils (1992) described a small, free-driftingsystem for recording in situ predator�prey interactions between juvenileherring schools and copepod swarms. One particular objective of thestudy was to determine how herring schools search and encounter preyunder the reduced visibility conditions in the Baltic caused inter alia byrecent enhancement of phytoplankton blooms. The hardware, known asecoSCOPE, consists of two optical endoscopes mounted on a remotelyoperated vehicle. Microlayers containing the herring schools are locatedfrom 40 m away using a scanning sonar also mounted on the ROV.Images of the predators are collected by one charge coupled device arrayand prey by a second one. The field system is accompanied by a softwarepackage dynIMAGE that allows the user to process images in a way thatcompensates for system swaying caused by microturbulence. Using thissystem, Kils (1992) was able to provide one of the first recordings of fishschools capturing prey in the field, describing copepod captures at a rateof 2.4 s21 by herring feeding within layers containing high (up to850 l21) concentrations of copepods.

Recent advances in digital video and its miniaturization and low costhave resulted in a range of new equipment for in situ and laboratory obser-vations. Kawaguchi et al. (2010) present a method that combines footagecaptured with dual digital stereo-video cameras with a commercially avail-able motion analysis system, WinAnalyse (Intec), which enables sophisti-cated studies of the movement of aquatic animals, including the analysis ofcomplex interactions among individuals in an aggregation. An introductionto the basic stereophotogrammetry, including specifics on calibration andprecision of the hardware and software, is provided. They demonstrate thecapability of this equipment by describing qualitative behaviour of labora-tory schools of Antarctic krill and testing hypotheses about the effects oflight and food. For example, the method allows a comparison of NNDsand swimming speeds in different regions of the swarm. Krill swam inpolarized groups and responded cohesively to objects that produced a sharpcontrast but not to those that were less distinct. Schools broke up whenthey encountered dense phytoplankton patches but aggregated more tightlywhen kept in a white featureless background. Viscido et al. (2004) usedsimilar equipment to compare behaviour of real and simulated fish schools.


Some parallels with krill schools are apparent, e.g. there was a clear rela-tionship between group speed and polarity; in both cases polarized groupsswam faster than non-polarized ones.

This equipment would permit researchers to test in situ, e.g. whetherindividuals at the edge of a swarm are likely to be travelling predomi-nantly in vertical trajectories compared to those in the centre that arehypothesized to be travelling predominantly in a more horizontal direc-tion. This would enable those at the edges to exploit favourable currentsgenerated by the swimming movements of swarm members (Ritz, 2000).

A breakthrough in the three-dimensional analysis of thousands of indivi-duals in a group was reported by Cavagna et al. (2008a,b). These authorsstudied flocks of starlings and devised computerized techniques to identifycorresponding images of individuals from stereoscopic pairs of pictures takenby cameras placed 25 m apart and about 100 m from the birds (Fig. 4.10).This analysis yielded several important conclusions, among them:

i. Each bird interacts with a fixed number of neighbours irrespective oftheir distance. In other words, the birds interacted with neighboursaccording to topological distance are not metric. One of the conse-quences seems to be that birds under attack from predators do not losecohesion when the flock rapidly changes shape, density and direction.

ii. Each bird interacted with a maximum of seven neighbours. This mayrepresent a cognitive limit for starlings although there is some evi-dence that it may be a more widespread limit for other social species.

The techniques described by Cavagna et al. (2008a,b) lend themselvesto the study of aggregations of other social species with the promise ofmore robust estimates of behavioural characteristics. They also reportmethods for removing bias due to individuals at the borders of the aggre-gation. Neglect of this factor can cause erroneous conclusions especiallyin small groups which, hitherto, have necessarily been the subject ofthree-dimensional analysis.

A new approach in visualization of aggregations was recently intro-duced by Myriax Pty Ltd. (http://eonfusion.myriax.com/). It is calledEonfusion and permits aggregations to be readily displayed in four dimen-sions, x, y, z coordinates and time. An example of a krill school is shownin Fig. 4.11, and other examples can be found at http://www.youtube.com/watch?v5CF8pb1a9gvA&feature5 player_embedded.

Recent developments include the video plankton recorder (VPR)(Fig. 4.12) that is essentially a towed underwater video microscope thatimages, identifies, counts and sizes plankton, and other particles in thesize range 100 µm�5 cm (IGBP Science 5).

Some of the newest VPR platforms permit high-speed towing (10knots) out of the wake of the ship, a moored autonomous profiler to obtainhigh-resolution time series of water column plankton, and autonomous


http://www.eonfusion.myriax.com/

http://www.youtube.com/watch?v(CF8pb1a9gvA&feature(player_embedded




Figure 4.10 Upper panels are actual photos of a flock of starlings taken by a pair of ste-

reoscopic cameras placed 25 m apart and about 100 m from the birds. Square boxes indi-

cate corresponding birds in the two pictures. Lower panels are three-dimensional

reconstructions of the flock from four different perspectives. After Cavagna and Giardina

(2008).


underwater vehicles to provide remote spatial sampling of plankton andenvironmental variables (IGBP Science, 2003). Together these image cap-ture systems offer a wide range of potential approaches to gather the pri-mary data on aggregations.

3.2. Optical plankton counters and holography

Since its first appearance in 1992 (Checkley et al., 1997), use of the opti-cal plankton counter has increased exponentially to provide quantitative

Figure 4.11 (A) Visualization of a school of Antarctic krill (Euphausia superba) using

Eonfusion. The positions (x,y,z) were used to derive metrics, e.g. swimming speed, direc-

tion, acceleration and nearest neighbour data. These were used to explore behaviour of

individuals over time. (B) Krill connected to their nearest neighbour (for distances

,50 mm). Reproduced with permission from Eonfusion: Tim Pauly; Australian Antarctic

Division: Rob King, So Kawaguchi; University of Tasmania: Jon Osborn; Georgia Tech: David

Murphy, Jeanette Yen, Donald Webster.


measurements of abundance and sizes of meso-zooplankton. In additionit has evolved to become an important piece of equipment in multidisci-plinary studies for acquiring and displaying data from a range of sensors(Cass-Calay, 2003). Spatial and temporal zooplankton distributions can berecorded by instruments mounted on a range of underwater towed frames(Zhou and Tande, 2002). At present, however, OPCs have limited valuein the study of social aggregations because of issues of avoidance, coinci-dences in measurements, image resolution and depth of field.

A submersible holographic system has been used to visualize theinstantaneous in situ three-dimensional distribution of copepods and parti-cles .10 µm in a 1 l volume (Fig. 4.13; Malkiel et al., 1999). Resultsshow clear evidence of clustering at almost all depths sampled. One greatadvantage of holographic imaging systems over other optical systems istheir ability to resolve small particles, e.g. plankters over a much largersample volume. The authors give as an example the comparison betweena planar imaging system that can resolve a 20 µm object with a depth offield of 0.6 mm at a given wavelength, whereas a holographic systemunder similar conditions would resolve the same object with a depth offield over 100 times larger. This potentially makes holography a valuabletool for in situ studies of pelagic aggregations.

3.3. Acoustic technology

Traditional acoustics involves the detection of sound waves reflectedfrom a target (active) or emitted from a target imbedded with an acoustic

Figure 4.12 The housing for a VPR, with a very streamlined shape, sits ready for deploy-

ment off the back of a research vessel (Gulf of Maine Area Program � GoMA). . http://

www.coml.org/investigating/observing/vprs.


http://www.coml.org/investigating/observing/vprs

http://www.coml.org/investigating/observing/vprs

transmitter. Detection of ‘natural’ sound from a target (passive acoustics)has also shown promise for detection of large objects, and we discuss itbriefly with regard to pelagic species in a following subsection.

Verifying the species or individual represented by the acoustic signalhas been a primary challenge and as a result there have been severalattempts to combine optical and acoustic sensors for the purpose of iden-tifying and quantifying zooplankton in situ. For example, Jaffe et al.(1998) introduced the optical�acoustical submersible imaging system inwhich a digital still camera was mounted on their FISH-TV sonar array.The camera is triggered to capture an image within the sonar beamswhen the target strength exceeds 290 dB. Using this equipment, theauthors reported imaging 375,000 individual zooplankters, many as smallas 1 mm (Genin et al., 2005). In a variation on a theme, Warren et al.(2001) used an analogue video camera to aim their acoustic array,mounted on an ROV, at individual zooplankters, i.e. siphonophores,

Mirror

Mirr

or

Spatial filter

Laser

Relay lenses

Camera

Sampling volume

Collimatinglenses

Expandinglenses

Mirror

Mirror

Figure 4.13 Holocamera: optical setup of in-line holography. Redrawn from Malkiel

et al. (1999).


euphausiids or other taxa. These ‘images’ of individuals can then be usedas for conventional optics to derive measures of in situ school parametersrelated to social aggregation such as NND.

Traditional acoustics have been limited to resolving pelagiczooplankton and fish at distances of 10s to 100s of metres from the ves-sel. A further limitation is that single-beam downward-looking echo-sounders sample only a narrow cone of water beneath the vessel (Coxet al., 2009). Thus they may undersample activity just outside this cone,e.g. predators interacting with target schools. Increasingly, multibeamechosounders (MBEs) have been used that not only broaden thewidth of the swath sampled but also enable direct observation ofvolume and surface area of aggregations leading to more accurate three-dimensional estimates of individuals (Cox et al., 2009). This equipmentis well suited to resolving pelagic aggregations in situ, and has been usedto analyse three-dimensional structure of anchovy and sardine schools(Gerlotto and Paramo, 2003; Gerlotto et al., 2004), and also to studypredator�prey interactions between Atlantic puffins and herring(Axelsen et al., 2001). Cox et al. (2009) used an MBE to study interac-tions between swarms of Antarctic krill and penguins and fur seals.Kaartvedt et al. (2009) showed that detailed in situ behaviour of individ-ual mesopelagic fish could be resolved by multibeam acoustics deployedfrom a stationary vessel (Fig. 4.14). This study revealed a variety ofDVM behaviour amongst myctophid fish including normal, reversemigration and non-migration of some individuals. Techniques usingsubmerged echosounders offer exciting prospects for studying deep-liv-ing aggregations non-intrusively, assuming that the target species are notthemselves responding to the echoes.

The GLOBEC’s Southern Ocean programme employs a Bio-OpticalMulti-frequency Acoustical and Physical Environmental Recorder(BIOMAPER-II) to map abundance and distribution of Antarctic phyto-plankton, zooplankton and especially krill (Wiebe et al., 2002).BIOMAPER can record data from 500 m or more of the water column ata time. It is towed at speeds up to 10 knots and data are fed continuouslyto the surface via conducting cable. The instrument uses a five-frequencysonar system, a VPR and an environmental sensor system that measureswater temperature, salinity, oxygen, chlorophyll and light levels.

The most astounding breakthrough in recent years is that of conti-nental shelf-scale imaging (Makris et al., 2006), that has revealed hugefish shoals covering many square kilometres and containing tens of mil-lions of fish. An example of the trace from the ocean acoustic waveguideremote sensing (OAWRS) system is shown in Fig. 4.15. This techniquerelies on the continental shelf environment acting as an acoustic wave-guide, making it possible to survey areas roughly one million timesgreater than conventional fish-finding methods. The technology utilizes


a moored acoustic source and a towed receiving array. Sound propagatesover long ranges via trapped modes that suffer only cylindrical spreadingloss rather than spherical loss suffered by more conventional sonar. Usingthis technique, unprecedented imaging of fish shoals is possible providingdetails of behaviour such as school formation, fragmentation and move-ment. Once species can be discriminated, this technology offers excitingpotential to look at the interaction between schools as a basic unit ofanalysis, perhaps using the density within a school as a measure of behav-ioral response.

Figure 4.14 (A) Echo traces of individual mesopelagic fish. Graduated scale at right of

picture refers to backscattering strength (dB). (B) Example of three-dimensional move-

ment and swimming speed of a target ascribed to Benthosema glaciale (framed in A) at

380 m depth. Graduated scale at right of picture refers to swimming speed (m s21).

Reprinted with permission from Kaartvedt et al. (2009).


Traditional optical visualization is compromised by poor visibility,which acoustics can also overcome, although acoustic methods can alsobe compromised when the water contains many particles. A promisingnew instrument for use in turbid water is the acoustic camera(DIDSONt � Dual Frequency Identification Sonar, i.e. either 1.8 and1.1 MHz or 0.70 and 1.2 MHz) that has been used to count and identifyfish, and to reveal schooling behaviour of salmon in hatchery ponds(Belcher et al., 2002). Depending on the frequencies, it can image objectsfrom 1 out to 80 m range. The near video quality of the image allowsobservation of fish behaviour in turbid water and at night near naturaland manmade structures. Furthermore, the equipment can be used closeto the banks of rivers or streams where deployment of other acousticinstruments is problematic. Since the sonar beam is emitted perpendicular

(A)

(C)

(B)

(D)

(E)

100 m

α

100 m 100 m

100 m

–6

–9

–12

–15

–18

70

80

90

100

0 1

0 3 6Distance (km)

Dis

tanc

e (k

m)

Dep

th (

m)

9 12

2 3 4 5Range (km)

6 7 8 9

Fish/m3

10.20.050.010.001

β

γ

δ

α β γ δ

Figure 4.15 (A�D) Comparison of OAWRS with conventional fish-finding sonar (CFFS).

A sequence of areal density (fish m22) images taken roughly 10 min apart is shown. The cor-

responding CFFS is overlain in light blue (see colour plate). CFFS position for the given

OAWRS image is indicated by a circle. (E) Range-depth profile of fish volumetric density

(fish m23) measured along the transect in (A�D). After Makris et al. (2006).


to the flow, fish can be counted and sized as they pass through the field ofview. DIDSONt is small and portable requiring only 30 Wof power.

3.4. Electronic tags

Both visual and acoustic in situ observations of pelagic social aggregationtypically require the presence of the scientist or attending vessel. Forlarger pelagic species, an alternative is to implant or attach data recordersthat can transmit data or be recovered at a later time. The primary goal ofmost electronic tagging studies has been to understand the movements ofindividual animals, which with sufficient sample size are assumed to berepresentative of the population (Hobday et al., 2009). The study of indi-vidual behaviour has been a focus, aggregation behavior less often, andsocial behavior rarely. Three types of tag are in common use: archivaltags, satellite transmitting tags and acoustic tags, all of which could pro-vide information on social behaviors in aggregations.

Archival tags are usually internally implanted in an individual, and col-lect a variety of data (depth/pressure, internal/external temperature,light), while the animal is at liberty (Gunn and Block, 2001). Followingrecovery of the animal, data can be processed to determine daily position(Welch and Eveson, 1999; Teo et al., 2004). Errors in calculated positioncan still be in the order of hundreds of kilometers which has preventedinsights into group behaviors (Nielsen et al., 2009). Depth resolution ismore accurate, and so simultaneous movements in depth of tagged ani-mals could be used to determine coherence in group behaviors.

A modification of the basic archival tag has been a range of satellitearchival tags which are externally attached and transmit data continuously(e.g. SPLASH, SPOT; Weng et al., 2005) or detach from the animal aftera period of time and transmit summarized data to satellites (PSAT; Blocket al., 2001; Patterson et al., 2008). Position estimates are similarly coarse,and these tags have been of little use for studies of social aggregation.However, both types of archival tag have been widely used on a range ofpelagic species with social aggregations, including fishes, sharks, marinemammals and birds, and even a few large invertebrates (Nomuru jellyfish)and squid (Gilly et al., 2006), and so potential remains to use these tagswithin their technological constraints to achieve breakthroughs in docu-menting social behaviors.

Acoustic tags transmit while attached to the animal, which can beactively tracked by an attending researcher (Block et al., 1997; Davis andStanley 2002) or monitored using fixed acoustic receivers (Heupel et al.,2006; Hobday et al., 2009). Advantages of using fixed receivers include theincreased time over which multiple individuals can be simultaneously moni-tored subsequent to the tagging event. Thus, there is increased likelihood ofdetecting natural behaviors and recurrent grouping of individuals. The


resolution of the position estimates (tens to hundreds of metres) also allowsgreater certainty that individuals are clustering (Klimley and Holloway,1999).

Aggregations of pelagic fishes are common, either with each other,with floating objects, or with topographic features such as seamounts(Holland et al., 2009). This behaviour is increasingly being exploited byfishermen who each year release thousands of floating FADs throughoutthe oceans (Hallier and Gaertner, 2008). These devices typically attracttunas, but also a range of other species including sharks, billfish and seaturtles (Dagorn and Freon, 1999; Gaertner et al., 2002; Hallier andGaertner, 2008). Ongoing research is aimed at a greater understanding ofthe relationship between aggregation/association of fish at FADs and thepotential impact of fisheries (Hallier and Gaertner, 2008). Evolution ofelectronic tags for this purpose has been rapid and there are plans to testthe feasibility of using acoustic data to determine whether the fish are,indeed, schooling, or if they are responding to the floating structure(Taquet et al., 2007; Soria et al., 2009).

Another recently developed device, the CHAT (CommunicatingHistory Acoustic Transmitter) or Business Card tag, has the potential toexchange data between fish and then remotely transfer archived data fromthe fish to listening stations deployed on the seabed or on buoys (Hollandet al., 2009). Whether or not data on schooling behaviour could be loggedwould depend on successful design of suitable sensors, but would be animportant innovation in monitoring aggregations at sea, as well as contribut-ing vital information to ecologists and fisheries scientists (Holland andDagorn, 2009).

While most electronic tagging studies have a single-species focus,Goni et al. (2009) acoustically tracked juvenile albacore in the Bay ofBiscay and simultaneously collected prey distribution data from echosoun-ders. Although tuna depth distribution did not relate to prey distribution,the combination of these technologies is suitable for developing fine-scaleunderstanding of tuna schooling and foraging behavior, and may yieldinteresting results in future.

3.5. Future technology challenges

Traditional technologies based on visualization have provided most of theinsights regarding pelagic social aggregations to date, with increases inusage of acoustic technology providing recent breakthroughs for a rangeof species. The challenge for understanding pelagic social behavior isdeveloping systems that can be used in the open sea, often remote fromthe observing researcher. Use of electronic tags is likely to be restricted tothe larger species for the foreseeable future, and so information on thesmaller taxa is likely to come from combining technologies, such as


camera systems with traditional acoustics. The resolution of the data intime is sufficient in existing technologies, with millisecond reportingcommonplace. Spatial resolution can be considerably improved, and whilegeolocation may have inherent limitations (Welch and Eveson, 1999;Nielsen et al., 2009), alternative technologies that allow tags to communi-cate will overcome the uncertainty in individual position estimates, byconfirming co-location.

With regard to tags, a number of companies are supplying technology,yet these are generally incompatible. Given the scale of the pelagic ocean,and the infrequent encounters between independently tagged animals,opportunities are likely to be missed if tags cannot communicate.Commercial considerations are important and likely prevent completecompatibility between technologies; however, a middle ground may touse a common signal for ‘chat’ between tags and maintain individual cod-ing for the primary data collection (Grothues, 2009).

4. Theoretical Developments in Social

Aggregation

‘A key purpose of modelling is to distinguish behavioural cause from organizational

effect by studying the consequences of various hypothetical social interaction rules’

Parrish et al. (2002)

We do not propose to give a comprehensive review of theoreticalmodelling of aggregations here. Instead we will describe the kinds ofmodels that have been applied to the problem to date, outline the resultsobtained and identify areas and directions where further work is war-ranted. Levin (1997) discussed conceptual issues posed by modellingaggregations, particularly in relation to scale, and Parrish et al. (2002) pro-vided a recent review of theoretical modelling approaches particularly asthey relate to fish schools. The current state of modelling animal aggrega-tions and its relation to empirical data has been described with great clar-ity by Giardina (2008). In general, three major frameworks have beenused to model animal aggregations (Parrish and Edelstein-Keshet, 1999).These three types differ in (i) spatial and temporal scale of the analysis, (ii)the kind of information individuals use to aggregate and (iii) the mathe-matical complexity (Giardina, 2008).

First, individual-based models (IBMs or Lagrangian) are based onindividual trajectories with attributes such as location, genotype, phenotype,physiological and behavioural status, and allow rule-based responsesto environmental data in order to explore the dynamics that lead to forma-tion of groups. For example, fish joining a school would assume a particular


angular position with respect to another individual and a particulardirection that could be modified by the environment (e.g. current strength)(Adioui et al., 2003). The individual unit could be an individual fish, or aschool of fishes. An aggregation has been shown to result when individualsfollow three simple rules: move in the same direction as neighbours, remainclose to them and avoid collisions (Giardina, 2008).

Second, Eulerian models deal with populations as the base unit. Theytypically consider each unit at a geographically fixed location instead ofsimulating each individual and, in the case of a school, predict emergentgroup properties such as population density. Space is not georeferencedand the number of individuals inside each cell is followed in time.Environmental drivers can also be applied to each unit, which respondsaccording to a set of dynamical equations. This type of model is mostlyapplied when investigating population evolution on long time scales orover large spatial scales, and has made a strong contribution to fisheriesscience.

A third approach relies on discrete simulations using a range of indi-vidual behavioural rules and motion (an example is the use of cellularautomata). This method, it is argued, gives a clear, visual prediction ofhow individual behaviour contributes to that of the group. Rules that donot lead to group formation can be modified or rejected, allowing someform of hypothesis testing. Automatic selection of the rules based on‘reproductive fitness’ occurs in models using genetic algorithms whichhave been used to explore schooling behaviour (Giske et al., 1998).

Lagrangian and Eulerian concepts have been combined (Adioui et al.,2003) and further refined to better forecast three-dimensional movementpatterns of individual salmon in the ELAM (Eulerian�Lagrangian�Agentmethod) (Goodwin et al., 2006). These studies have shown that theELAM framework is well suited to describing large-scale patterns inhydrodynamics and water quality at the same time as much smaller scalesat which individual fish make movement decisions. This ability to simul-taneously handle dynamics at multiple scales allows ELAM models to real-istically represent fish movements within aquatic systems. This methodseems to hold promise for future ecological modelling of fish schools.

An Eulerian approach has been used to model the effects of environ-mental conditions, krill fishery and natural predation on Antarctic krillgrowth, distribution, vertical migration, feeding, etc. (Alonzo andMangel, 2001; Alonzo et al., 2003). These authors used a dynamic state-variable model to predict the effect of changes in predation risk onbehaviour and spatial distribution of Antarctic krill. However, these mod-els did not incorporate the influence of schooling behaviour, which couldbe a key factor affecting krill population abundance (Willis, 2007a). Theproblem has been that, it has been impossible to reproduce schoolingbehaviour of Antarctic krill in laboratory conditions and thus gather


reliable empirical data with which to paramaterize models. Recently,however, routine generation of schooling behaviour in krill in aquariumtanks has been achieved (Kawaguchi et al., 2010) which may allow prog-ress in population modelling. Research on other social aquatic speciessuggests that individual behaviour is so closely tied to that of the aggrega-tion that data extrapolated from individuals isolated from their schoolswill be misleading. A good example is metabolism, where it has beendemonstrated that individuals in aggregations consume far less energy perunit mass than isolated individuals (Ritz, 2000; Ritz et al., 2001).

IBMs and computer-generated artificial life creatures (e.g. boids andlater offshoots such as efloys) have been successful in reproducing the self-organizing characteristics of aggregations of flocking birds and marinespecies, particularly fishes. But are these emergent properties of the groupfunctionally important? Parrish and Edelstein-Keshet (1999) caution that,although simple rules can generate lifelike behaviour, there is no guaran-tee that living systems follow simple rules. More experimental data gainedby tracking individuals in small groups is needed (Parrish and Edelstein-Keshet, 1999). In fact, one clear conclusion from this overview is that thefeedback between empirical observations and modelling, that is so impor-tant to scientific progress, has been seriously hampered by lack of data onthe former (Giardina, 2008).

The theoretical study of animal aggregations has a long history, butattempts to characterize the relationships between individual-level behaviourand group-level patterns have been hampered by lack of a common frame-work to schooling models (Parrish et al., 2002). They argue in favour of adistinction between group- and population-level characteristics that areinevitable consequences when many identical particles become aggregatedin space (epiphenomena or ‘pattern’), and true emergent properties, thatbenefit members because of their membership of the group. Viscido et al.(2007) continued this theme by analysing the factors contributing to fishschool formation and maintenance. In a simulation study they found thatseveral, mostly social, factors were important in giving rise to emergentproperties: notably a repulsion factor is necessary to prevent collisions, aneutral zone must exist in which there is neither attraction nor repulsion, amodest alignment impulse strong enough to induce polarity is necessary,number and weighting of influential neighbours is critically important, as isspeed of motion. A topological response to neighbours, i.e. focal individualinteracting with a fixed number of neighbours irrespective of distance, hasalso been shown in simulations of starling flocks (Ballerini et al., 2008) butonly recently demonstrated empirically (see Cavagna et al., 2010) (see alsoFig. 4.10). Bode et al. (2010) suggested an alternative interpretation of thebehavioural rules governing individuals moving in aggregations that is dis-tance based rather than topological. They hypothesize that an individualunder threat must minimize its ‘oddity’ by increasing the rate at which it


updates its position and orientation within the group, leading to an overallincrease in synchronization and uniformity. This is a subject rich with possi-bilities for future study.

Ecosystem modelling, such as trophic modelling via Ecopath, has alsoignored the implications of social aggregation. However, Willis (2007b) hasrecently examined ways in which to incorporate social aggregation andbehaviour into an ecosystem model of southern bluefin tuna (Thunnus mac-coyii) in the Great Australian Bight (see Fig. 4.16). This simple solution,including a schooling and non-schooling category, may be suitable for dis-crete behaviours, but it unlikely to succeed when a continuum of states ispossible. Given that factors such as feeding success, predator vulnerabilityand reproductive capacity differs across the spectrum of solitary to aggre-gated individuals, inclusion of aggregation state in ecosystem models, suchas ecopath is likely to influence ecosystem understanding.

5. Social Aggregation, Climate Change and

Ocean Management

The direct impacts of climate change on species and populationsinclude changes in distribution, abundance, phenology and physiology.The pelagic ocean is well buffered from some of the impacts of climate

Figure 4.16 Simplified food web of the Great Australian Bight, Australia. In this ecosys-

tem representation, southern bluefin tuna are represented in two separate boxes, lone tuna

and tuna in schools, illustrating the different trophic relationships related to school mem-

bership. With permission from Willis (2007b).


change; however, some secondary processes, including aggregation, willlikely be affected. Climate impacts will also be expressed indirectly viachanges in the habitats where aggregation can occur, as well as directly onthe aggregating species. For example, changes in CO2 concentration thatchange acidity of the ocean, or changes in ocean circulation that controloxygen concentrations, may limit the waters where aggregation canoccur. Given the many instances of change in distributions of marine spe-cies (Occhipinti-Ambrogi, 2007; Last et al., 2011), species composition ischanging regionally. A change in local water temperature could changethe competitive advantages of one species over another. If one specieslived in schools and the other did not, this could have a profound influ-ence not only on the food web but also potentially on commercialfisheries.

While climate change has already begun to impact waters around theworld (Harley et al., 2006; Cao and Caldeira, 2008), the variation infuture predictions and the spatial and temporal resolution of predictionsmake forecasting biological change difficult. The response to historicalclimate variability is a window to the future, and so we first describesome of the documented responses of aggregations to climate variabilityat a range of scales. We explore the implication of these changes forpelagic trophic linkages (e.g. changes in energy that pass from aggrega-tions of zooplankton to fish/seals/whales).

Disruption to physiological abilities, such as smell, has been shown toaffect homing, predator and conspecific detection in larval fishes (Mundayet al., 2009). If senses important to establishing aggregations are disrupted,then benefits to populations discussed in earlier sections will be reduced,with possible ecosystem impacts.

One outcome of global warming is that warmer seawater will havemarkedly decreased viscosity. For example, a rise in temperature from0�C to 5�C decreases viscosity of seawater by about 15%. This could pos-sibly lead to aggregation at a smaller size if, as Ritz (2000) suggests, cur-rents generated by the group serve to offset the tendency to sink inheavier than water animals, i.e. crustaceans. Antarctic krill only begin toaggregate when they reach late furcilia stage at a length of around 10 mm(Hamner et al., 1989). Ritz (2000) suggests that this is the size at whichthe cost of resisting sinking becomes too great to remain solitary. On theother hand, increasing acidity of oceanic waters could result in crustaceansand molluscs (e.g. pteropods) precipitating less calcium carbonate in theircuticles and shells and becoming less dense, which may offset the decreas-ing viscosity.

Polar marine ecosystems are particularly vulnerable to climate changebecause of the effects that small increases in temperature can have on thecritical interfacial habitat between ice and water (Smetacek and Nicol,2005). The atmosphere around the Antarctic Peninsula region has been


warming faster than any other part of the planet and a concomitantdecrease in sea ice extent and krill population has been documented(Atkinson et al., 2004). However, the effect of climate warming in thisregion could be more subtle and complex than described by these authorsgiven the influence of ocean/atmosphere interactions such as El NinoSouthern Oscillation (Loeb et al., 2009).

Overall, the impact of climate change on social aggregation is unknownbut may deliver ecosystem surprises. What effects might climate variabilityhave on social species? In the first place, warming might change species dis-tributions, but it is not clear whether warming per se would affect aggrega-tive behaviour directly. However, there is a range of indirect consequencesof warmer oceans. For example, it seems inconceivable that DVM wouldnot be affected. If DVM is traded for aggregation in certain circumstances,could climate change push it in a particular direction?

One serious consequence of climate change for social aggregatuonsmight be through an affect on seasonal migrations. Barbaro et al. (2009)have modelled the migration of the Icelandic Capelin stock, an importantcommercial resource. The most significant factor in determining theroute of migration was oceanic temperature and the way the fish schoolsresponded to it. In a warming ocean, the migration routes and aggrega-tion tendency of fishes may change.

The possible effects of both predator and climate-change-induced altera-tions in schooling behaviour of krill need to be considered for sustainablemanagement. Predation risk is commonly held to be an important stimulusfor aggregation in krill (Ritz, 1994; Kaartvedt et al., 1996; Folt and Burns,1999; De Robertis, 2002), but the urgency for aggregation can be overrid-den, e.g. if predation risk is low, if light levels are low enough to frustratevisual predators or possibly if energetic considerations dictate that verticalmigration is a more economical option. The urge to aggregate with similarindividuals is very strong (Bakun and Cury, 1999), but clearly there is greatflexibility in this behaviour (Bertrand et al., 2006) and possibilities for trade-offs are many. Willis (2007a) proposes that decimation of whales in theSouthern Ocean has led not to widely predicted large increases in krillstocks but to a change of vertical migratory behaviour that resulted in lowerkrill abundance. If the disappearance of a large proportion of the whalepopulation had been the direct result of climate change rather than maninduced, this would have profound implications for our strategies to preparefor the effects of a changing climate.

A second and more immediate threat for pelagic systems comes fromcommercial fisheries which are already inducing profound changes in fishpopulations (Heino and Dieckmann, 2009). Since the majority of com-mercially important fish are those living in schools (Pauly et al., 2005),evolutionary changes induced by fishing will also affect the food webdynamics. Fisheries management can benefit from including information


on social behaviour to better manage fisheries. The basin hypothesis, thatfish stocks collapse to a core area of habitat as population declines(MacCall, 1990), is in part due to social interactions between aggrega-tions. Mixed-species or mixed age-class aggregations provide informationon population status. Effective closed area management is dependent onidentifying the vulnerable stages of an animal’s life, and social theory canidentify these stages.

Introduction of exotic species into new regions could disrupt competi-tive interactions and predator�prey relationships. The net result of suchintroductions could be complicated by aggregative behaviour of the new-comer or the ousted competitor, making prey availability more problem-atic. This could extend to human fisheries activities which are highlydependent on schooling behaviour of prey species (Quinn and Deriso,1999; Cury et al., 2000).

6. Conclusion

6.1. Do reviews stimulate new work?

In his earlier review of social aggregation in pelagic invertebrates, Ritz(1994) suggested that several topics deserved particular attention, including:

1. General behavioural studies regarding individuals in aggregations.2. Determining genetic relatedness among individuals in aggregations.3. Evaluating whether particle capture is more successful by aggregated

than by solitary individuals. This was considered to be challenging, asrigorous assessment requires experimental reproduction of patchy fooddistribution instead of the more common homogeneous distributionused in laboratory containers.

4. Effectiveness of aggregations with regard to successful mate findingand reproduction.

5. Experimental studies of decision-making, e.g. trade-offs betweenaggregation or other behavioral choices.

Evaluating if this earlier review was successful in promoting researchin specific areas is worthwhile before suggesting additional areas for futurestudy, as such insight can help modify the way such suggestions can bemade. This type of analysis is rare in review papers (Roberts et al., 2006)but in our view is a worthy consideration. Using the search tools availablein ISI (http://apps.isiknowledge.com), publication trends for these topicswere analysed. Comparing pre-1995 and post-1995 is difficult as the elec-tronic indexing of publications was incomplete in the early period, sohere we consider the effort against the suggested areas in Ritz (1994). We


http://www.apps.isiknowledge.com

also do not claim a direct influence, only that these topics did receiveattention in the subsequent years, and although we concede that theresults are somewhat inconclusive, they do suggest that for some topicareas a different approach to the one suggested by Ritz (1994) might leadto breakthroughs. Since 1994 a total of 76 papers have been publishedwith a focus on ‘pelagic invertebrates’ (search term: pelagic SAME inver-tebrates); of these 8 (11%) had a behavioural component (topic 1), four(5%) a genetic component (topic 2), although none focused on the rela-tionship between individuals in an aggregation, and 16 (21%) consideredfeeding aspects, again few compared aggregated against solitary individuals(topic 3). Papers focusing on reproductive advantages as a consequence ofaggregation (topic 4) and trade-offs in decision-making (topic 5) have notreceived any experimental attention, according to our search approach(but see the example described below). Thus, some of these topics are stillworthy of attention some 15 years on.

By way of example, one suggestion made by Ritz (1994) has acted as acatalyst for further research. Ritz (1994) suggested that because schoolingbehaviour sub-served predator protection and facilitated foraging and feed-ing, social species might find vertical migration redundant since it too servesthe same functions. Although this work was not experimental in nature, itcan be regarded as an example under topic 5 (trade-offs). Observationalresearch by De Robertis (2002) and De Robertis et al. (2003) suggestedthat Euphausia pacifica did not form subsurface social aggregations in certainenvironments, instead relying on DVM for protection from fish predators.The fact that E. pacifica forms schools in other environments (Mauchline,1980; Hanamura et al., 1984) highlights the flexibility of the behaviouralrepertoire, and indicates that flexibility in behavior can result from environ-mental differences. On a related theme, Kaartvedt et al. (1996) suggestedthe significance of ‘anti-predation windows’ that may encourage verticalmigration of the predators (Norway Pout) at times of low visibility, e.g.dawn and dusk, when it is less risky to chase prey (krill) closer to the sur-face. At other times, vertical migration is suppressed and fish remain in dee-per water. If social aggregation is a cost, e.g. if schools are more likely toattract the attention of piscivorous predators, then it might be suggestedthat schooling would be abandoned while migrating vertically. This argu-ment could be negated if the energy saved by migrating in schools is morethan offset by the increased risk of predation.

6.2. Future needs and synthesis

Benefits of aggregation in a wide range of systems are well known, and it isno surprise that aggregation is also commonly found amongst pelagic spe-cies. There are many aspects of aggregations in the ocean still in need ofresearch, e.g. energy saving: larger swarms expended less energy than smaller


ones, which, in turn, saved more energy than un-aggregated individuals(Ritz, 2000). The conversion of these energy savings into enhanced fitnessis assumed to follow, but empirical evidence is lacking to date.

Network analysis is a relatively new field that should lead to morebreakthroughs in future, particularly for understanding social structureand membership persistence. Social networks have been demonstrated inaggregations of fish, birds and mammals. There is ample scope for extend-ing this kind of analysis to other aggregations including invertebrates.

Understanding sensory processes, and their role in forming, maintain-ing and dispersal of groups is an area ripe for future research. Conflictsbetween sensory inputs and processes (e.g. camouflage and communica-tion) can occur, and the decision-making processes of individuals couldbe studied empirically using the latest optical or acoustic technology. Orit might lend itself to analysis by simulation or modelling.

Technologies for the study of aggregation have progressed markedly inrecent decades, and offer insights in the future, using tools such as simul-taneous acoustic imaging of individuals (Kaartvedt et al., 2009) and elec-tronic communicating tags (Holland et al., 2009). We expect three-dimensional analysis to extend to thousands of individuals within marineaggregations to occur soon, as has been reported for terrestrial birds(Cavagna et al., 2008a,b).

Membership of aggregations is often claimed to be an effective strategyin successful mate finding and reproduction (topic 4 in Section 6.1). Wecould find no evidence of progress since Ritz (1994) suggested that moreexperimental evidence was needed. This is perhaps an area that might beadvanced through modelling studies, to examine the benefit in terms offitness, and help refine experimental design. Ultimately, tests on real spe-cies would be desirable.

Modelling studies (topic 5 in Section 6.1) are the basis of a suggestionthat predatory risk-induced changes in behaviour could lead to majorchanges in population abundance (Alonzo and Mangel, 2001; Alonzoet al., 2003; Willis, 2007a). Using a dynamic state-variable model, Alonzoand Mangel (2001) predict that, in the face of extreme temperatures and/or predation risk, Antarctic krill will shrink in size and spatial distributionmay change. In Alonzo et al. (2003) they use the same approach to try tounderstand the relationship between krill fisheries and penguin foragingsuccess in the Antarctic. The model suggests that a change in krill behav-iour is likely to cause stronger effects of the fishery on penguins than canbe explained solely by the percentage of biomass removed. This isbecause, as offshore krill are depleted by fishing, the deeper location ofinshore krill near penguins, which are land based for reproduction, willmake them less accessible to diving penguins. Note, though, that theabsence of the influence of schooling behaviour in these models mightalter the predator�prey dynamics (Ritz, 2002).


Experimental studies of decision-making in the face of conflicting sig-nals also seem rare in recent decades, and perhaps modelling or simulationstudies might be timely to explore such trade-offs.

Impacts of climate change on aggregations are difficult to predictthough we do not anticipate direct effects. However, there are many pos-sibilities for indirect consequences. These include the consequence oflikely changes in water viscosity with increasing temperature; decreasedacidity with increasing dissolved CO2; and distribution changes both geo-graphical and vertical. With unprecedented rates of environmentalchange, the ability of pelagic species to adapt is questionable, and so weexpect more attention to this area in the coming years.

Aggregation in the pelagic zone is common, and likely important inthe survival and well-being of many species, including humans who, viafisheries, rely heavily on aggregation to efficiently catch food. In fact, theaverage encounter rate of non-aggregated prey would lead to starvationin many predators (e.g. whales). With many of the world’s commercialtarget species already overfished, the importance of aggregation shouldnot be underestimated in a functioning pelagic zone.

ACKNOWLEDGEMENTS

The encouragement of the former editor for Advances in Marine Biology, David Sims, inseeking this review and the comments of the editorial board in focusing the scope areappreciated. We thank all of our colleagues who allowed us to reproduce their publishedand unpublished figures.

REFERENCESAdioui, M., Treuil, J. P. and Arino, O. (2003). Alignment in a fish school: A mixed

Lagrangian�Eulerian approach. Ecological Modelling 167, 19�32.Allan, J. R. (1986). The influence of species composition on behaviour in mixed species

cyprinid shoals. Journal of Fish Biology 29(Suppl. A):97�106.Alldredge, A. L. and Hamner, W. M. (1980). Recurring aggregation of zooplankton by a

tidal current. Estuarine and Coastal Marine Science 10, 31�37.Alldredge, A. L., Robison, B. H., Fleminger, A., Torres, J. J., King, M. and Hamner,

W. M. (1984). Direct sampling and in situ observation of a persistent copepodaggregation in the mesopelagic zone of the Santa Barbara Basin. Marine Biology 80,75�81.

Allen, S. and Demer, D. A. (2003). Detection and characterization of yellowfin and blue-fin tuna using passive-acoustical techniques. Fisheries Research 63, 393�403.

Alonzo, S. H. and Mangel, M. (2001). Survival strategies and growth of krill: Avoidingpredators in space and time. Marine Ecology Progress Series 209, 203�217.

Alonzo, S. H., Switzer, P. V. and Mangel, M. (2003). An ecosystem-based approach tomanagement: Using individual behaviour to predict the indirect effects of Antarctickrill fisheries on penguin foraging. Journal of Applied Ecology 40, 692�697.

Angel, M. (1993). Biodiversity of the pelagic ocean. Conservation Biology 7, 760�772.


Atkinson, A., Siegel, V., Pakhomov, E. and Rothery, P. (2004). Long-term decline in krillstock and increase in salps within the Southern Ocean. Nature 432, 100�103.

Atkinson, A., Shreeve, R. S., Hirst, A. G., Rothery, P., Tarling, G. A., Pond, D. W.,Korb, R. E., Murphy, E. J. and Watkins, J. L. (2006). Natural growth rates inAntarctic krill (Euphausia superba): II. Predictive models based on food, temperature,body length, sex, and maturity stage. Limnology and Oceanography 51, 973�987.

Axelsen, B. E., Anker-Nilssen, T., Fossum, P. and Nøttestad, L. (2001). Pretty patternsbut a simple strategy: Predator�prey interactions between juvenile herring andAtlantic puffins observed with multibeam sonar. Canadian Journal of Zoology 79,1586�1596.

Baker, C. F. and Montgomery, J. C. (2001). Species-specific attraction of migratorybanded kokopu juveniles to adult pheromones. Journal of Fish Biology 58, 1221�1229.

Bakun, A. and Cury, P. (1999). The ‘school trap’: A mechanism promoting large-ampli-tude out-of-phase population oscillations of small pelagic fish species. Ecology Letters 2,349�351.

Ballerini, M., Cabibbo, R., Candelier, A., Cavagna, E, Cisbani, I., Giardina, V., Lecomte,A., Orlandi, G., Parisi, A., Procaccini, M., Viale, M. and Zdravkovic, V. (2008).Interaction ruling animal collective behavior depends on topological rather than metricdistance: Evidence from a field study. Proceedings of the National Academy of Sciences ofthe United States of America 105, 1232�1237.

Banas, N. S., Wang, D.-P. and Yen, J. (2004). Experimental validation of an individual-based model for zooplankton swarming. In Scales in aquatic ecology: Measurements, analy-sis, modeling (L. J. Seuront and P. G. Strutton, eds), pp. 161�180. CRC Press, BocaRaton, FL.

Barbaro, A., Einarsson, B., Birnir, B., Sigurðsson, S., Valdimarsson, H., Palsson, O. K.,Sveinbjornsson, S. and Sigurðsson, P. (2009). Modelling and simulations of the migra-tion of pelagic fish. ICES Journal of Marine Science 66, 826�838.

Belcher, E., Hanot, W, and Burch, J. (2002). Dual-Frequency Identification Sonar(DIDSON). Proceedings of the International Symposium on Underwater Technology, April16�19, 187�192.

Benoit-Bird, K. J. and Whitlow, W. L. (2009). Phonation behavior of cooperatively forag-ing spinner dolphins. Journal of the Acoustical Society of America 125, 539�546.

Bertrand, A., Barbieri, M. A., Gerlotto, F., Leiva, F. and Cordova, J. (2006). Determinismand plasticity of fish schooling behaviour as exemplified by the South Pacific jackmackerel Trachurus murphyi. Marine Ecology Progress Series 311, 145�156.

Block, B. A., Booth, D. T. and Carey, F. G. (1992). Depth and temperature of theblue marlin, Makaira nigricans, observed by acoustic telemetry. Marine Biology 114,175�183.

Block, B. A., Keen, J. E., Castillo, B., Dewar, H., Freund, E. V., Marcinek, D. J., Brill,R. W. and Farwell, C. (1997). Environmental preferences of yellowfin tuna (Thunnusalbacares) at the northern extent of its range. Marine Biology 130, 119�132.

Block, B. A., Dewar, H., Blackwell, S. B., Williams, T. D., Prince, E. D., Farwell, C. J.,Boustany, A., Teo, S. L. H., Seitz, A., Walli, A. and Fudge, D. (2001). MigratoryMovements, Depth Preferences, and Thermal Biology of Atlantic Bluefin Tuna. Science293, 1310�1314.

Bode, N. W. F., Faria, J. J., Franks, D. W. and Krause, J. (2010). How perceived threatincreases synchronization in collectively moving animal groups. Proceedings of the RoyalSociety Series B � Biological Sciences 277, 3065�3070.

Boehlert, G. W. and Genin, A. (1987). A review of the effects of seamounts on biologicalprocesses. In Seamount, Islands and Atolls (Geophysical Monograph) (B. H. Keating, P.Fryer, R. Batiza and G. W. Boehlert, eds), pp. 319�334. American GeophysicalUnion, Washington D.C.


Burrows, M. T. and Tarling, G. (2004). Effect of density dependence on diel verticalmigration of populations of northern krill: A genetic algorithm model. Marine EcologyProgress Series 277, 209�220.

Buskey, E. (1998). Components of mating behavior in planktonic copepods. Journal ofMarine Systems 15, 13�21.

Buskey, E. J., Lenz, P. H. and Hartline, D. K. (2002). Escape behavior of planktonic cope-pods in response to hydrodynamic disturbances: High speed video analysis. MarineEcology Progress Series 235, 135�146.

Butler, M. J., Macdermid, A. B. and Booth, J. D. (1999). The cause and consequence ofontogenetic changes in social aggregation in New Zealand spiny lobsters. MarineEcology Progress Series 188, 179�191.

Cadavid, L. F., Powell, A. E., Nicotra, M. L., Moreno, M. and Buss, L. W. (2004). Aninvertebrate histocompatibility complex. Genetics 167, 357�365.

Cao, L. and Caldeira, K. (2008). Atmospheric CO2 stabilization and ocean acidification.Geophysical Research Letters 35, L19609.

Cass-Calay, S. L. (2003). The feeding ecology of larval Pacific hake (Merluccius productus)in the California Current region: An updated approach using a combined OPC/MOCNESS to estimate prey biovolume. Fisheries Oceanography 12, 34�48.

Cavagna, A. and Giardina, I. (2008). The seventh starling. Significance 5, 62�66.Cavagna, E., Giardina, I., Orlandi, G., Parisi, A., Procaccini, A., Viale, M. and

Zdravkovic, V. (2008). The STARFLAG handbook on collective animal behaviour.Part I: Empirical methods. Animal Behaviour 76, 217�236.

Cavagna, E., Giardina, I., Orlandi, G., Parisi, A. and Procaccini, A. (2008). TheSTARFLAG handbook on collective animal behaviour. 2: Three-dimensional analysis.Animal Behaviour 76, 237�248.

Cavagna, A., Cimarelli, A., Giardina, I., Parisi, G., Santagati, R., Stefanini, F. andTavarone, R. (2010). From empirical data to inter-individual interactions: Unveilingthe rules of collective animal behavior. Mathematical Models and Methods in AppliedScience 20, 1491�1510.

Checkley, D. M. J., Ortner, P. B., Settle, L. R. and Cummings, S. R. (1997). A continu-ous, underway fish egg sampler. Fisheries Oceanography 6, 58�73.

Chilvers, B. L. and Corkeron, P. J. (2002). Association patterns of bottlenose dolphins(Tursiops aduncus) off Point Lookout, Queensland, Australia. Canadian Journal of Zoology80, 973�979.

Chivers, D. P., Brown, G. E. and Smith, R. J. F. (1995). Familiarity and shoal cohesion infathead minnows (Pimephales promelas) � implications for antipredator behavior.Canadian Journal of Zoology 73, 955�960.

Clapham, P. J. and Baker, C. S. (2003). How many whales were killed in the SouthernHemisphere in the 20th century? CAMLR � SC/53/O 14, 1�3.

Clutter, R. I. (1969). The microdistribution and social behaviour of some pelagic mysidshrimps. Journal of Experimental Marine Biology and Ecology 3, 125�155.

Cohen, P. J. and Ritz, D. A. (2003). Role of kairomones in feeding interactionsbetween seahorses and mysids. Journal of the Marine Biological Association of the UK 83,633�638.

Conradt, L. and Roper, T. J. (2000). Activity synchrony and social cohesion: Afission�fusion model. Proceedings of the Royal Society of London Series B � BiologicalSciences 267, 2213�2218.

Couzin, I. D. (2006). Behavioral ecology: Social organization in fission�fusion societies.Current Biology 16, R169�R171.

Couzin, I. D., James, R., Croft, D. P. and Krause, J. (2006). Social organization and infor-mation transfer in schooling. In Fish Cognition and Behaviour (C. Brown, K. N. Lalandand J. Krause, eds), Blackwell Publishing, Oxford.


Cox, M. J., Warren, J. D., Cutter, G. R. and Brierley, A. S. (2009). Multibeam echosoun-der observations reveal interactions between Antarctic krill and air-breathing predators.Marine Ecology Progress Series 378, 199�209.

Croft, D., James, R., Ward, A. J. W., Botham, M. S. and Mawdsley, D. (2005). Assortativeinteractions and social networks in fish. Oecologia 143, 211�219.

Croft, D. P., James, R. and Krause, J. (2008). Exploring Animal Social Networks. PrincetonUniversity Press, Princeton.

Cross, P. C., Lloyd-Smith, J. O., Bowers, J. A., Hay, C. T., Hofmeyr, M. and Getz,W. M. (2004). Integrating association data and disease dynamics in a social ungulate:Bovine tuberculosis in African buffalo in the Kruger National Park. Annales ZoologiciFennici 41, 879�892.

Cury, P., Bakun, A., Jarre, A., Quinones, R. A., Shannon, L. J. and Verheye, H. M.(2000). Small pelagics in upwelling systems: Patterns of interaction and structuralchanges in ‘wasp-waist’ ecosystems. ICES Journal of Marine Science 57, 603�618.

Dagorn, L. and Freon, P. (1999). Tropical tuna associated with floating objects: A simulationstudy of the meeting point hypothesis. Canadian Journal of Fisheries and Aquatic Sciences 56,984�993.

Daly, K. L. and Macaulay, M. C. (1991). Influence of physical and biological mesoscaledynamics on the seasonal distribution and behavior of Euphausia superba in theAntarctic marginal ice zone. Marine Ecology Progress Series 79, 37�66.

Dare, J.E. (2008). Remaining Unseen in the Pelagic World: The Conflict betweenCamouflage and Feeding in Trachurus novaezelandiae. MSc Thesis, University ofAuckland.

Darwin, C. (1953). Note on hydrodynamics. Proceedings of the Cambridge PhilosophicalSociety � Biological Science 49, 342�354.

Davis, T. L. O. and Stanley, C. A. (2002). Vertical and horizontal movements of southernbluefin tuna (Thunnus maccoyii) in the Great Australian Bight observed with ultrasonictelemetry. Fishery Bulletin 100, 448�465.

De Robertis, A. (2002). Small-scale spatial distribution of the euphausiid Euphausiapacifica and the overlap with planktivorous fishes. Journal of Plankton Research 24,1207�1220.

DeBlois, E. M. and Rose, G. A. (1996). Cross-shoal variability in the feeding habits ofmigrating Atlantic cod (Gadus morhua). Oecologia 108, 192�196.

Denton, E. J. (1970). Review lecture: On the organization of reflecting surfaces in somemarine animals. Philosophical Transactions of the Royal Society of London Series B �Biological Sciences 258, 285�313.

Denton, E. J. (1980). Reflectors in fishes. Scientific American 242, 65�72.Denton, E. J. and Gray, J. A. B. (1993). Stimulation of the acoustico-lateralis system of

clupeid fish by external sources and their own movements. Philosophical Transactions ofthe Royal Society of London Series B � Biological Sciences 341, 113�127.

Denton, E. J. and Rowe, D. M. (1994). Reflective communication between fish, withspecial reference to the greater sand eel, Hyperoplus lanceolatus. Philosophical Transactionsof the Royal Society of London Series B � Biological Sciences 344, 221�237.

Denton, E. J. and Rowe, D. M. (1998). Bands against stripes on the backs of mackerel,Scomber scombrus. Proceedings of the Royal Society of London Series B � Biological Sciences265, 1051�1058.

Dewar, W. K., Bingham, R. J., Iverson, R. L., Nowacek, D. P., St. Laurent, L. C. andWiebe, P. H. (2006). Does the marine biosphere mix the ocean? Journal of MarineResearch 64, 541�561.

Doall, M. H. (1998). Locating a mate in 3D: The case of Temora longicornis.Philosophical Transactions of the Royal Society of London Series B � Biological Sciences 353,681�689.


Doall, M., Strickler, J., Fields, D. and Yen, J. (2002). Mapping the free-swimming attackvolume of a planktonic copepod, Euchaeta rimana. Marine Biology 140, 881�882.

Dudzinski, K. M., Douaze, E. and Thomas, J. (2002). Communication. In Encyclopedia ofMarine Mammals (W. F. Perrin, B. Wursig and H. C. M. Thewissen, eds), AcademicPress, Inc., Burlington, MA, USA.

Eloffson, R. (1966). The nauplius eye and frontal organs of the non-malacostracans(Crustacea). Sarsia 25, 1�128.

Faucher, K., Parmentier, E., Becco, C., Vandewalle, N. and Vandewalle, P. (2010). Fishlateral system is required for accurate control of shoaling behaviour. Animal Behaviour79, 679�687.

Fewell, J. H. (2003). Social insect networks. Science 301, 1867�1870.Field, C. B., Behrenfeld, M. J., Randerson, J. T. and Falkowski, P. (1998). Primary

Production of the Biosphere. Integrating Terrestrial and Oceanic Components Science 281,237�240.

Flierl, G., Grunbaum, D., Levin, S. and Olson, D. (1999). From individuals to aggregations:The interplay between behavior and physics. Journal of Theoretical Biology 196, 397�454.

Flynn, A. J. and Ritz, D. A. (1999). Effect of habitat complexity and predatory style oncapture success of fish feeding on aggregated prey. Journal of the Marine BiologicalAssociation of the U.K. 79, 487�494.

Folt, C. L. and Burns, C. W. (1999). Biological drivers of zooplankton patchiness. Trendsin Ecology and Evolution 14, 300�305.

Forrester, G. E. (1991). Social rank, individual size and group composition as determi-nants of food-consumption by humbug damselfish, Dascyllus aruanus. Animal Behaviour42, 701�711.

Frank, S. A. (2007). All of life is social. Current Biology 17, 648�650.Freon, P. and Dagorn, L. (2000). Review of fish associative behaviour: Towards a generaliza-

tion of the meeting point hypothesis. Reviews in Fish Biology and Fisheries 10, 183�207.Fuiman, L. A. and Magurran, A. E. (1994). Development of predator defenses in fishes.

Reviews in Fish Biology and Fisheries 4, 145�183.Gaertner, D., Menard, F., Develter, C., Ariz, J. and Delagardo de Molina, A. (2002).

Bycatch of billfishes by the European tuna purse-seine fishery in the Atlantic Ocean.Fishery Bulletin 100, 683�689.

Genin, A. (2004). Bio-physical coupling in the formation of zooplankton and fish aggre-gations over abrupt topographies. Journal of Marine Systems 50, 3�20.

Genin, A., Greene, C., Haury, L., Wiebe, P., Gal, G., Kaartvedt, S., Meir, E., Fey, C. andDawson, J. (1994). Zooplankton patch dynamics: Daily gap formation over abrupttopography. Deep-Sea Research 41, 941�951.

Genin, A., Haury, L. R. and Greenblatt, P. (1988). Interactions of migrating zooplanktonwith shallow topography: Predation by rockfish and intensification of patchiness. Deep-Sea Research 35, 151�175.

Genin, A., Jaffe, J. S., Reef, R., Richter, C. and Franks, P. J. S. (2005). Swimming againstthe flow: A mechanism of zooplankton aggregation. Science 308, 860�862.

Gerasimov, V. V. (1962). Feeding behavior of Murmansk herring in school and out of schoolin aquarium conditions. Trudy Murmanskogo Morskogo Biologicheskogo Instituta 2, 254�259.

Gerlotto, F. and Paramo, J. (2003). The three dimensional morphology and internal struc-ture of Clupeids schools as observed using vertical scanning multibeam sonar. AquaticLiving Resources 16, 113�122.

Gerlotto, F., Castillo, J., Saavedra, A., Barbieri, M. A., Espejo, M. and Cotel, P. (2004).Three-dimensional structure and avoidance behaviour of anchovy and common sardineschools in central southern Chile. ICES Journal of Marine Science 61, 1120�1126.

Giardina, I. (2008). Collective behavior in animal groups: Theoretical models and empiri-cal studies. Human Frontier Science Project Journal 2, 205�219.


Gilly, W. F., Markaida, U., Baxter, C. H., Block, B. A., Boustany, A., Zeidberg, L.,Reisenbichler, K., Robison, B., Bazzino, G. and Salinas, C. (2006). Vertical and hori-zontal migrations by the jumbo squid Dosidicus gigas revealed by electronic tagging.Marine Ecology Progress Series 324, 1�17.

Giraldeau, L.-A. and Caraco, T. (2000). Social Foraging Theory. Princeton University Press,Princeton.

Giraldeau, L.-A. and Gillis, D. (1985). Optimal group size can be stable: A reply to Sibly.Animal Behaviour 33, 666�667.

Giske, J., Huse, G. and Fiksen, O. (1998). Modelling spatial dynamics of fish. Reviews inFish Biology and Fisheries 8, 57�91.

Glance, N. S. and Huberman, B. A. (1994). The dynamics of social dilemmas. ScientificAmerican 270, 58�63.

Goni, N., Arregui, I., Lezama, A., Arrizabalaga, H. and Moreno, G. (2009).Small scale vertical behaviour of juvenile albacore in relation to their bioticenvironment in the Bay of Biscay. In Tagging and Tracking of Marine Animals withElectronic Devices (J. Nielsen, H. Arrizabalaga, N. Fragosa, A. Hobday, M. E. Lutcavageand J. Sibert, eds), Vol. 2, Springer Academic Publishers, Dordrecht, The Netherlands.

Goodwin, N. B., Grant, A., Perry, A. L., Dulvy, N. K. and Reynolds, J. D. (2006). Lifehistory correlates of density-dependent recruitment in marine fishes. Canadian Journalof Fisheries and Aquatic Sciences 63, 494�509.

Gowans, S., Whitehead, H. and Hooker, S. K. (2001). Social organization in northernbottlenose whales, Hyperoodon ampullatus: Not driven by deep-water foraging? AnimalBehaviour 62, 369�377.

Gray, J. A. B. and Denton, E. J. (1991). Fast pressure pulses and communication betweenfish. Journal of the Marine Biological Association of United Kingdom 71, 83�106.

Grothues, T.M. (2009). A Review of Acoustic Telemetry Technology and a Perspectiveon its Diversification Relative to Coastal Tracking Arrays. In Tagging and Tracking ofMarine Animals with Electronic Devices Reviews: Methods and Technologies in FishBiology and Fisheries, Vol. 9, 77-90. Springer Science, Netherlands.

Grunbaum, D. (2006). Align in the sand. Science 312, 1320�1322.Gunn, J. and Block, B. (2001). Advances in acoustic, archival, and satellite tagging of

tunas. In Tuna: Physiology, Ecology and Evolution (B. A. Block and E. D. Stevens, eds),pp. 167�224. Academic Press, San Diego, CA.

Hallier, J.-P. and Gaertner, D. (2008). Drifting fish aggregation devices could act asan ecological trap for tropical tuna species. Marine Ecology Progress Series 353,255�264.

Hamilton, W. D. (1964). The genetical evolution of social behaviour. Journal of TheoreticalBiology 7, 1�52.

Hamner, W. M. (1985). The importance of ethology for investigations of marinezooplankton. Bulletin of Marine Science 37, 414�424.

Hamner, W. M. (1988). Behaviour of plankton and patch formation in pelagic ecosystems.Bulletin of Marine Science 43, 752�757.

Hamner, W. M. and Parrish, J. K. (eds), (1997). Animal Groups in Three Dimensions: HowSpecies Aggregate Cambridge University Press, Cambridge.

Hamner, W. M., Hamner, P. P., Obst, B. S. and Carleton, J. H. (1989). Field observationson the ontogeny of schooling of Euphausia superba furciliae and its relationship to icein Antarctic waters. Limnology and Oceanography 34, 451�456.

Hanamura, Y., Endo, Y. and Taniguchi, A. (1984). Underwater observations on thesurface swarm of a euphausiid, Euphausia pacifica in Sendai Bay, northeastern Japan. LaMer 22, 63�68.

Hardy, A. C. and Gunther, E. R. (1935). The plankton of the South Georgia whalinggrounds and adjacent waters 1926�27. Discovery Reports 11, 1�456.


Harley, C. D. G., Hughes, A. R., Hultgren, K., Miner, B. G., Sorte, C. J. B., Thornber,C. S., Rodriguez, L. F., Tomanek, L. and Williams, S. L. (2006). The impacts of cli-mate change in coastal marine systems. Ecology Letters 9, 228�241.

Hartman, K. L., Visser, F. and Hendricks, A. J. E. (2008). Social structure of Risso’s dol-phins (Grampus griseus) at the Azores: A stratified community based on highly associ-ated social units. Canadian Journal of Zoology 86, 294�306.

Haury, L. R., McGowan, J. A. and Wiebe, P. H. (1978). Patterns and processes in thetime-space scales of plankton distributions. In Spatial Patterns in Plankton Communities(J. H. Steele, ed), pp. 277�327. Plenum Press, New York.

Hay, D. E. and McKinnell, S. M. (2002). Tagging along: Association among individualPacific herring (Clupea pallasii) revealed by tagging. Canadian Journal of Fisheries andAquatic Sciences 59, 1960�1968.

Heino, M. and Dieckmann, U. (2009). Fisheries-induced evolution in Encyclopedia of LifeSciences. John Wiley & Sons, Chichester, UK.

Helfman, G. (1984). School fidelity in fishes: The yellow perch pattern. Animal Behaviour32, 673�689.

Hensor, E., Couzin, I. D., James, R. and Krause, J. (2005). Modelling density-dependentfish shoal distributions in the laboratory and field. Oikos 110, 344�352.

Heppner, F. (1997). Three-dimensional structure and dynamics of bird flocks. In AnimalGroups in Three Dimensions: How Species Aggregate (J. K. Parrish and W. M. Hamner,eds), pp. 68�87. Cambridge University Press, Cambridge.

Heupel, M. R., Semmens, J. M. and Hobday, A. J. (2006). Automated acoustic trackingof aquatic animals: Scales, design and deployment of listening station arrays. Marine andFreshwater Research 57, 1�13.

Higgs, D. M. and Fuiman, L. A. (1996). Light intensity and schooling behaviour in larvalgulf menhaden. Journal of Fish Biology 48, 979�991.

Higham, J., Bejder, L. and Lusseau, D. (2009). An integrated and adaptive managementmodel to address the long-term sustainability of tourist interactions with cetaceans.Environmental Conservation 35, 294�302.

Hilborn, R. (1991). Modeling the stability of fish schools: Exchange of individual fishbetween schools of skipjack tuna (Katsuwonus pelamis). Canadian Journal of Fisheries andAquatic Science 48, 1081�1091.

Hoare, D. and Krause, J. (2003). Social organisation, shoal structure and informationtransfer. Fish and Fisheries 4, 269�279.

Hoare, D. J., Ruxton, G. D., Godin, J.-G. J. and Krause, J. (2000). The social organizationof free-ranging fish shoals. Oikos 89, 546�554.

Hobday, A. J. and Campbell, G. (2009). Topographic preferences and habitat partitioningby pelagic fishes in southern Western Australia. Fisheries Research 95, 332�340.

Hobday, A. J., Kawabe, R., Takao, Y., Miyashita, K. and Itoh, T. (2009). Correction of anabundance index using acoustic tag data for juvenile southern bluefin tuna in southernWestern Australia. In Tagging and Tracking of Marine Animals with Electronic Devices II.Reviews: Methods and Technologies in Fish Biology and Fisheries (J. Nielsen, J. R. Sibert, A. J.Hobday, M. E. Lutcavage, H. Arrizabalaga and N. Fragosa, eds), pp. 405�422. Springer,The Netherlands.

Holland, K. N., Meyer, C. G. and Dagorn, L. C. (2009). Inter-animal telemetry: Results fromfirst deployment of acoustic ‘business card’ tags. Endangered Species Research 10, 287�293.

Huntley, M. E. and Zhou, M. (2004). Influence of animals on turbulence in the sea.Marine Ecology Progress Series 273, 65�79.

Hurley, A. C. (1977). Mating behaviour of the squid Loligo opalescens. Marine Behaviour andPhysiology 4, 195�203.

IGBP Science 5 (2003). Marine Ecosystems and Global Change. (M. Barange and R. Harris,eds). Stockholm: IGBP, 32 pp.


Jaffe, J. S., Ohman, M. D. and Roberts, A. D. (1998). OASIS in the sea: Measurement ofthe acoustic reflectivity of zooplankton with concurrent optical imaging. Deep-SeaResearch II 45, 1239�1253.

Janssen, J. (1981). Searching for zooplankton just outside Snell’s window. Limnology andOceanography 26, 1168�1171.

Jiang, H., Osborn, T. R. and Meneveau, C. (2002). Hydrodynamic interaction betweentwo copepods: A numerical study. Journal of Plankton Research 24, 235�253.

Johnsen, S. and Sosik, H. M. (2003). Cryptic coloration and mirrored sides as camouflagestrategies in near-surface pelagic habitats: Implications for foraging and predator avoid-ance. Limnology and Oceanography 48, 1277�1288.

Johnston, N. and Ritz, D. A. (2001). Synchronous development and release of broods bythe swarming mysids Anisomysis mixta australis, Paramesopodopsis rufa and Tenagomysis tas-maniae (Mysidacea: Crustacea). Marine Ecology Progress Series 223, 225�233.

Kaartvedt, S., Melle, W., Knutsen, T. and Skjoldal, H. R. M (1996). Vertical distributionof fish and krill beneath water of varying optical properties. Marine Ecology ProgressSeries 136, 51�58.

Kaartvedt, S., Røstad, A., Klevjer, T. A. and Staby, A. (2009). Use of bottom-mountedecho sounders in exploring behavior of mesopelagic fishes. Marine Ecology ProgressSeries 395, 109�118.

Kaiser, M. J., Attrill, M. J., Jennings, S., Thomas, D. N., Barnes, D. K. A., Brierley, A. S.,Polunin, N., Raffaelli, D. G., Williams, P. J. and le, B (2005). Marine Ecology Processes,Systems and Impacts. Oxford University Press, Oxford.

Katija, K. and Dabiri, J. O. (2009). A viscosity-enhanced mechanism for biogenic oceanmixing. Nature 460, 624�627.

Kawaguchi, S., King, R., Meijers, R., Osborn, J. E., Swadling, K. M., Ritz, D. A. andNicol, S. (2010). An experimental aquarium for observing the schooling behaviour ofAntarctic krill (Euphausia superba). Deep-Sea Research II 57, 683�692.

Kils, U. (1992). The ecoSCOPE and dynIMAGE: Microscale Tools for in situ Studies ofPredator Prey Interactions. Archiv fur Hydrobiologie Beih 36, 83�96.

Klimley, A. P. and Holloway, C. F. (1999). School fidelity and homing synchronicity ofyellowfin tuna. Thunnus albacares. Marine Biology 133, 307�317.

Klimley, A. P., Jorgensen, S. J., Muhlia-Melo, A. and Beavers, S. C. (2003). The occur-rence of yellowfin tuna (Thunnus albacares) at Espiritu Seamount in the Gulf ofCalifornia. Fishery Bulletin 101, 686�692.

Krause, J. and Ruxton, G. D. (2002). Living in groups. Oxford University Press, Oxford240 pp.

Krause, J., Godin, J. G. J. and Brown, D. (1996). Phenotypic variability within andbetween fish shoals. Ecology 77, 1586�1591.

Krause, J., Hoare, D. J., Croft, D., Lawrence, J., Ward, A. J. W., Ruxton, G. D., Godin,J. G. J. and James, R. (2000). Fish shoal composition: Mechanisms and constraints.Proceedings of the Royal Society of London Series B � Biological Sciences 267,2011�2017.

Krause, J., Ward, A. J. W., Jackson, A. L., Ruxton, G. D., James, R. and Currie, S.(2005). The influence of differential swimming speeds on composition of multi-speciesfish shoals. Journal of Fish Biology 67, 866�872.

Land, M. (1988). The optics of animal eyes. Contemporary Physics 29, 435�455.Last, P. R., White, W. T., Gledhill, D. C., Hobday, A. J., Brown, R., Edgar, G. J. and

Pecl, G. T. (2011). Long-term shifts in abundance and distribution of a temperate fishfauna: A response to climate change and fishing practices. Global Ecology andBiogeography 20, 58�72.

Latora, V. and Marchiori, M. (2001). Efficient behavior of small-world networks. PhysicalReview Letters 87


Leising, A. W. (2001). Copepod foraging in patchy habitats and thin layers using a 2-Dindividual-based model. Marine Ecology Progress Series 216, 167�179.

Leising, A. W. and Franks, P. J. S. (2000). Copepod vertical distribution within a spatiallyvariable food source: A simple foraging-strategy model. Journal of Plankton Research 22,999�1024.

Leising, A. W. and Yen, J. (1997). Spacing mechanisms within light-induced copepodswarms. Marine Ecology Progress Series 155, 127�135.

Levin, S. A. (1992). The problem of pattern and scale in ecology. Ecology 73,1943�1967.

Levin, S. A. (1997). Conceptual and methodological issues in the modeling of biologicalaggregations. In Animal Groups in Three Dimensions (J. K. Parrish and W. M. Hamner,eds), pp. 247�256. Cambridge University Press, Cambridge.

Liao, J. C. (2007). A review of fish swimming mechanics and behaviour in altered flows.Philosophical Transactions of the Royal Society Series B 362, 1973�1993.

Liao, J. C., Beal, D. N., Lauder, G. V. and Triantafyllou, M. S. (2003). The Karman gait:Novel kinematics of rainbow trout swimming in a vortex street. Journal of ExperimentalBiology 206, 1059�1073.

Liao, J. C., Beal, D. N., Lauder, G. V. and Triantafyllou, M. S. (2003). Fish exploitingvortices decrease muscle activity. Science 302, 1566�1569.

Loeb, V. J., Hofmann, E. E., Klinck, J. M., Holm-Hansen, O. and White, W. B. (2009).ENSO and variability of the Antarctic Peninsula pelagic marine ecosystem. AntarcticScience 21, 135�148.

Lusseau, D. (2003). The emergent properties of a dolphin social network. Proceedings of theRoyal Society of London Series B � Biological Sciences (Suppl.) 270, S186�S188.

Lusseau, D. and Newman, M. E. J. (2004). Identifying the role that animals play in theirsocial networks. Proceedings of the Royal Society of London Series B � Biological Sciences(Suppl.) 271, S477�S481.

MacCall, A.D. (1990). Dynamic geography of marine fish populations. Washington SeaGrant Program (Seattle) 153pp.

Mackas, D. L., Denham, K. L. and Abbott, M. R. (1985). Plankton patchiness: Biologyin the physical vernacular. Bulletin of Marine Science 37, 652�674.

Major, P. F. (1977). Predator�prey interactions in schooling fishes during periods of twi-light: A study of the silverside Pranesus insularum in Hawaii. Fisheries Bulletin US 75,415�426.

Makris, N., Ratilal, P., Symonds, D., Jagannathan, S., Lee, S. and Nero, R. (2006). Fishpopulation and behavior revealed by instantaneous continental shelf-scale imaging.Science 311, 660�663.

Malkiel, E., Alquaddoomi, O. and Katz, J. (1999). Measurements of plankton distributionin the ocean using submersible holography. Measurement Science and Technology 10,1142�1152.

Masuda, R. and Tsukamoto, K. (1999). School formation and concurrent developmentalchanges in carangid fish with reference to dietary conditions. Environmental Biology ofFishes 56, 243�252.

Mauchline, J. (1980). The Biology of Mysids and Euphausiids. Advances in Marine Biology18, 1�681.

McFarland, W. N. and Kotchian, N. M. (1982). Interaction between schools of fish andmysids. Behavioural Ecology and Sociobiology 11, 71�76.

McKinnell, S., Pella, J. J. and Dahlberg, M. L. (1997). Population-specific aggregations ofsteelhead trout (Oncorhynchus mykiss) in the North Pacific Ocean. Canadian Journal ofFisheries and Aquatic Sciences 54, 2368�2376.

Milinski, M. (1987). Tit-for-tat in sticklebacks and the evolution of cooperation. Nature325, 433�437.


Milinski, M., Pfluger, D., Kulling, D. and Kettler, R. (1990). Do sticklebacks cooperaterepeatedly in reciprocal pairs? Behavioural Ecology and Sociobiology 27, 17�21.

Miller, D. G. M. and Hampton, I. (1989). Biology and Ecology of the Antarctic krill. SCARand SCOR Scott Polar Institute, Cambridge.

Mochek, A. D. (1987). Ethological Organization of Coastal Fish Communities. Nauka,Moscow [in Russian]

Montgomery, J.C. and Carton, A.G. The senses: chemosensory, visual and octavolateralis.In: Fish Behaviour. (C. Magnhagen, V.A. Braithwaite, E. Forsgren and B.G. Kapooreds.) Enfield, New Hampshire.Science Publishers: pp. 3-32.

Montgomery, J. C., Coombs, S. and Halstead, M. B. D. (1995). Biology of the mechano-sensory lateral line in fishes. Reviews in Fish Biology and Fisheries 5, 399�416.

Montgomery, J. C., Baker, C. F. and Carton, A. G. (1997). The lateral line can mediaterheotaxis in fish. Nature 389, 960�963.

Montgomery, J. C., McDonald, F., Baker, C. F., Carton, A. G. and Ling, N. (2003).Sensory integration in the hydrodynamic world of rainbow trout. Proceedings of theRoyal Society of London Series B � Biological Sciences 270(Suppl. 2):195�197.

Munday, P. L., Dixson, D. L., Donelson, J. M., Jones, G. P., Pratchett, M. S., Devitsina,G. V. and Doving, K. B. (2009). Ocean acidification impairs olfactory discriminationand homing ability of a marine fish. Proceedings of the National Academy of Sciences of theUnited States of America 106, 1848�1852.

Nicol, S. (2006). Krill, Currents, and Sea Ice: Euphausia superba and its ChangingEnvironment. BioScience 56, 111�120.

Nielsen, J., Sibert, J. R., Hobday, A. J., Lutcavage, M. E., Arrizabalaga, H. and Fragosa,N. (2009). ‘Tagging and Tracking of Marine Animals with Electronic Devices’ Reviews:Methods and Technologies in Fish Biology and Fisheries. Springer, Netherlands.

O’Brien, D. P. (1987). Direct observations of the behaviour of Euphausia superba andEuphausia crystallorophias (Crustacea: Euphausiacea) under pack ice during the Antarcticspring of 1985. Journal of Crustacean Biology 7, 437�448.

O’Brien, D. P. (1988). Direct observations of clustering (schooling and swarming)behaviour in mysids (Crustacea: Mysidacea). Marine Ecology Progress Series 42,235�246.

O’Brien, D. P. (1989). Analysis of the internal arrangement of individuals within crusta-cean aggregations (Euphausiacea, Mysidacea). Journal of Experimental Marine Biology andEcology 128, 1�30.

O’Brien, D. P. and Ritz, D. A. (1988). The escape responses of gregarious mysids(Crustacea; Mysidacea): Towards a general classification of escape responses in aggregatedcrustaceans. Journal of Experimental Marine Biology and Ecology 116, 257�272.

O’Brien, D. P., Tay, D. and Zwart, P. R. (1986). Laboratory method of analysis ofswarming behaviour in macroplankton: Combination of a modified flume tank andstereophotographic techniques. Marine Biology 90, 517�527.

Occhipinti-Ambrogi, A. (2007). Global change and marine communities: Alien speciesand climate change. Marine Pollution Bulletin 55, 342�352.

Ohtsuka, S., Inagaki, H., Onbe, T., Gushima, K. and Yoon, Y. H. (1995). Direct observa-tions of groups of mysids in shallow coastal waters of western Japan and southernKorea. Marine Ecology Progress Series 123, 33�44.

Omori, M. and Hamner, W. M. (1982). Patchy distribution of zooplankton: behaviour,population assessment and sampling problems. Marine Biology 72, 193�200.

Osborn, J. (1997). Analytical and digital photogrammetry. In Animal Groups in ThreeDimensions (J. K. Parrish and W. M. Hamner, eds), pp. 36�60. Cambridge UniversityPress, Cambridge.

Ottensmayer, C. A. and Whitehead, H. (2003). Behavioural evidence for social units inlong-finned pilot whales. Canadian Journal of Zoology 81, 1327�1338.


Pakhomov, E. A., Perissinotto, R., Froneman, P. W. and Miller, D. G. M. (1997).Energetics and feeding dynamics of Euphausia superba in the South Georgia region dur-ing the summer of 1994. Journal of Plankton Research 19, 399�423.

Parrish, J. K. and Edelstein-Keshet, L. (1999). Complexity, pattern and evolutionarytrade-offs in animal aggregation. Science 284, 99�101.

Parrish, J. K., Viscido, S. V. and Grunbaum, D. (2002). Self-organized fish schools: Anexamination of emergent properties. Biological Bulletin 202, 296�305.

Partridge, B. L. and Pitcher, T. J. (1980). The sensory basis of fish schools: Relative rolesof lateral line and vision. Journal of Comparative Physiology 135, 315�325.

Patria, M. P. and Wiese, K. (2004). Swimming in formation in krill (Euphausiacea), ahypothesis: Dynamics of the flow field, properties of antennular sensor systems and asensory-motor link. Journal of Plankton Research 26, 1315�1325.

Patterson, T. A., Evans, K., Carter, T. I. and Gunn, J. S. (2008). Movement and behaviourof large southern bluefin tuna (Thunnus maccoyii) in the Australian region determinedusing pop-up satellite archival tags. Fisheries Oceanography 17, 352�367.

Pauly, D., Christensen, V., Guenette, S., Pitcher, T. J., Sumaila, U. R., Walters, C. J.,Watson, R. and Zeller, D. (2002). Towards sustainability in world fisheries. Nature 418,689�695 .

Pauly, D., Watson, R. and Alder, J. (2005). Global trends in world fisheries: Impacts onmarine ecosystems and food security. Philosophical Transactions of the Royal Society SeriesB 360, 5�12.

Pavlov, D. S. and Kasumyan, A. O. (2000). Patterns and mechanisms of schooling behaviorof fish: A review. Journal of Ichthyology 40, S163�S231.

Pearson, H. C. (2009). Influences on dusky dolphin (Lagenorhynchus obscurus)fission�fusion dynamics in Admiralty Bay, New Zealand. Behavioural Ecology andSociobiology 63, 1437�1446.

Pitcher, T. J. and Parrish, J. K. (1993). Functions of shoaling behaviour in teleosts. InBehaviour of Teleost Fishes (T. Pitcher, ed), pp. 363�439. Chapman & Hall, London.

Pitcher, T. J. (ed) (1993). Behaviour of Teleost Fishes 2nd edn. Chapman & Hall, London.Poulet, S. A. and Ouellet, G. (1982). The role of amino acids in the chemosensory

swarming and feeding of marine copepods. Journal of Plankton Research 4,341�359.

Probatov, S. N. (1953). The results of the air exploring of the Caspian mullet and thepossibilities of its catch on the routes of migration. Fisheries 8, 18�22.

Quinn, T. J. and Deriso, R. B. (1999). Quantitative Fish Dynamics. Oxford UniversityPress, Oxford.

Quinn, T. P. and Tolson, G. M. (1986). Evidence of chemically mediated populationrecognition in coho salmon (Oncorhynchus kisutch). Canadian Journal of Zoology 64,84�87.

Radakov, D. V. (1973). Schooling in the Ecology of Fish. John Wiley & Sons, New York.Ratchford, S. G. and Eggleston, D. B. (1998). Size- and scale-dependent chemical

attraction contribute to an ontogenetic shift in sociality. Animal Behaviour 56,1027�1034.

Ritz, D. A. (1994). Social aggregation in pelagic invertebrates. Advances in Marine Biology30, 155�216.

Ritz, D. A. (1997). Costs and benefits as a function of group size: Experiments on aswarming mysid, Paramesopodopsis rufa Fenton. In Animal Groups in Three dimensions:How Species Aggregate (W. M. Hamner and J. K. Parrish, eds), pp. 194�206.Cambridge University Press, Cambridge.

Ritz, D. A. (2000). Is social aggregation in aquatic crustaceans a strategy to conserveenergy? Canadian Journal Fisheries and Aquatic Science 57, 59�67.


Ritz, D. A. (2002). Comment on Alonzo & Mangel (2001) survival strategies and growthof krill: Avoiding predators in space and time. Marine Ecology Progress Series 244,307�308.

Ritz, D. A., Foster, E. G. and Swadling, K. M. (2001). Benefits of swarming: mysids inlarger swarms save energy. Journal of the Marine Biological Association of the UnitedKingdom 81, 543�544.

Ritz, D. A. and Swadling, K. M. (2006). Energy savings at school. JMBA Global MarineEnvironment 3, 10�11.

Ritz, D. A., Cromer, L., Swadling, K. M., Nicol, S. and Osborn, J. (2003). Heart rate asa measure of stress in Antarctic krill, Euphausia superba. Journal of the Marine BiologicalAssociation of the United Kingdom 83, 329�330.

Roberts, P. D., Stewart, G. B. and Pullin, A. S. (2006). Are review articles a reliablesource of evidence to support conservation and environmental management? A com-parison with medicine. Biological Conservation 132, 409�423.

Roman, J. and Palumbi, S. R. (2003). Whales Before Whaling in the North Atlantic.Science 301, 508�510.

Rowe, D. M. and Denton, E. J. (1997). The physical basis of reflective communicationbetween fish, with special reference to the horse mackerel, Trachurus trachurus.Philosophical Transactions of the Royal Society of London Series B � Biological Sciences 352,531�549.

Ryer, C. H. and Olla, B. L. (1992). Social mechanisms facilitating exploitation of spatiallyvariable ephemeral food patches in a pelagic marine fish. Animal Behavior 44, 69�74.

Sale, P. F. (1971). Extremely limited home range in a coral reef fish, Dascyllus aruanus(Pisces: Pomacentridae). Copeia 1971, 324�327.

Shane, S. H., Wells, R. S. and Wursig, B. (1986). Ecology, behavior and social organiza-tion of the bottlenose dolphin: A review. Marine Mammal Science 2, 34�63.

Shaw, E. (1978). Schooling fishes. American Scientist 66, 166�175.Sibert, J. R., Hampton, J., Kleiber, P. and Maunder, M. (2006). Biomass, Size, and

Trophic Status of Top Predators in the Pacific Ocean. Science 314, 1773�1776.Sibly, R. M. (1983). Optimal group size is unstable. Animal Behaviour 31, 947�948.Siegel, V. and Kalinowski, J. (1994). Krill demography and small-scale processes: A review.

In Southern Ocean Ecology the Biomass Perspective (S. Z. El-Sayed, ed), pp. 145�164.Cambridge University Press, Cambridge.

Slooten, E., Dawson, S. M. and Whitehead, H. (1993). Associations among photographi-cally identified Hector’s Dolphins. Canadian Journal of Zoology 71, 2311�2318.

Smetacek, V. and Nicol, S. (2005). Polar ocean ecosystems in a changing world. Nature437, 362�368.

Soria, M., Dagorn, L., Potin, G. and Freon, P. (2009). First field-based experiment sup-porting the meeting point hypothesis for schooling in pelagic fish. Animal Behavior 78,1441�1446.

Steele, J. H. (1980). Patterns in plankton. Oceanus 23, 2�8.Steele, J. H. and Henderson, E. W. (1981). A simple plankton model. American Naturalist

117, 676�691.Strand, S. W. and Hamner, W. M. (1990). Schooling behaviour of Antarctic krill

(Euphausia superba) in laboratory aquaria: Reactions to chemical and visual stimuli.Marine Biology 106, 355�360.

Strickler, J. R. (1998). Observing free-swimming copepods mating. PhilosophicalTransactions of the Royal Society of London Series B 353, 671�680.

Svendsen, J. C., Skov, J., Bildsoe, M. and Steffensen, J. F. (2003). Intra-school positionalpreference and reduced tail beat frequency in trailing positions in schooling roachunder experimental conditions. Journal of Fish Biology 62, 834�846.


Taquet, M., Dagorn, L. C., Gaertner, J.-C., Girard, C., Aumerruddy, R., Sancho, G. andItano, D. (2007). Behavior of dolphinfish (Coryphaena hippurus) around drifting FADsas observed from automated acoustic receivers. Aquatic Living Resources 20, 323�330.

Teo, S. L. H., Boustany, A., Blackwell, S., Walli, A., Weng, K. C. and Block, B. A.(2004). Validation of geolocation estimates based on light level and sea surface temper-ature from electronic tags. Marine Ecology Progress Series 283, 81�98.

Turchin, P. (1997). Quantitative Analysis of Movement. Sinauer Associates, Sunderland, MA.Twining, B. S., Gilbert, J. J. and Fisher, N. S. (2000). Evidence of homing behavior in the

coral reef mysid Mysidium gracile. Limnology and Oceanography 45, 1845�1849.Tyack, P. L. (2000). Animal behavior: dolphins whistle a signature tune. Science 289,

1310�1311.Verdy, A. and Flierl, G. R. (2009). Evolution and social behavior in krill. Deep-Sea

Research II 55, 472�484.Villinger, J. and Waldman, B. (2008). Self-referent MHC type matching in frog tadpoles.

Proceedings of the Royal Society of London Series B � Biological Sciences 275, 1225�1230.Viscido, S. V., Parrish, J. K. and Grunbaum, D. (2004). Individual behavior and emergent

properties of fish schools: A comparison of observation and theory. Marine EcologyProgress Series 273, 239�249.

Viscido, S. V., Parrish, J. K. and Grunbaum, D. (2007). Factors influencing the structureand maintenance of fish schools. Ecological Modelling 206, 153�165.

Ward, A. J. W. and Hart, P. J. B. (2003). The effects of kin and familiarity on interactionsbetween fish. Fish and Fisheries 4, 348�358.

Ward, A. J. W. and Hart, P. J. B. (2005). Foraging benefits of shoaling with familiars maybe exploited by outsiders. Animal Behaviour 69, 329�335.

Ward, A. J. W., Botham, M. S., Hoare, D. J., James, R., Broom, M., Godin, J.-G. J. andKrause, J. (2002). Association patterns and shoal fidelity in the three-spined stickle-back. Proceedings of the Royal Society of London Series B � Biological Sciences 269,2451�2455.

Warner, R. R. (1988). Traditionality of mating-site preferences in a coral-reef fish. Nature335, 719�721.

Warren, J. D., Stanton, T. K., Benfield, M. C., Wiebe, P. H., Chu, D. and Sutor, M.(2001). In situ measurements of acoustic target strengths of gas-bearing siphonophores.ICES Journal of Marine Science 58, 740�749.

Watkins, J. J. and Murray, A. W. A. (1998). Layers of Antarctic krill, Euphausia superba:Are they just long krill swarms? Marine Biology 131, 237�247.

Watkins, J. L., Buchholz, F., Priddle, J., Morris, D. J. and Ricketts, C. (1992). Variation inreproductive status of Antarctic krill swarms: Evidence for a size-related sorting mech-anism? Marine Ecology Progress Series 82, 163�174.

Watts, D. J. and Strogatz, S. H. (1998). Collective dynamics of ‘small-world’ networks.Nature 393, 440�442.

Weber, L., El-Sayed, S. Z. and Hampton, I. (1986). The variance spectra of phytoplankton,krill and water temperature in the Antarctic Ocean south of Africa. Deep-Sea Research33, 1327�1343.

Webster, S. J. and Fiorito, G. (2001). Socially guided behaviour in non-insect inverte-brates. Animal Cognition 4, 69�79.

Weimerskirch, H., Martin, J., Clerquin, Y., Alexandre, P. and Jiraskova, S. (2001). Energysaving in flight formation. Nature 413, 697�698.

Welch, D. W. and Eveson, J. P. (1999). An assessment of light-based geoposition estimatesfrom archival tags. Canadian Journal of Fisheries and Aquatic Sciences 56, 1317�1327.

Weng, K. C., Castilho, P. C., Morrissette, J. M., Landeira-Fernandez, A. M., Holts, D. B.,Schallert, R. J., Goldman, K. J. and Block, B. A. (2005). Satellite tagging and cardiacphysiology reveal niche expansion in salmon sharks. Science 310, 104�106.


Whiteman, E. A. and Cote, I. M. (2004). Dominance hierarchies in group-living cleaninggobies: Causes and foraging consequences. Animal Behaviour 67, 239�247.

Wiebe, P. H., Greene, C. H., Benfield, M. C., Sosik, H. M., Austin, T. C., Warren, J. Dand Hammar, T. (2002). BIOMAPER-II: An integrated instrument platform for cou-pled biological and physical measurements in coastal and oceanic regimes. IEEEJournal of Oceanic Engineering 27, 700�716.

Wiese, K. (1996). Sensory capacities of euphausiids in the context of schooling. Marineand Freshwater Behaviour and Physiology 28, 183�194.

Wiese, K. and Ebina, Y. (1995). The propulsion jet of Euphausia superba (Antarctic krill) asa potential communication signal among conspecifics. Journal of the Marine BiologicalAssociation of the UK 75, 43�54.

Williams, R. and Lusseau, D. (2006). A killer whale social network is vulnerable totargeted removals. Biology Letters 2, 497�500.

Williams, R., Lusseau, D. and Hammond, P. S. (2009). The role of social aggregationsand protected areas in killer whale conservation: The mixed blessing of critical habitat.Biological Conservation 142, 709�719.

Willis, J.K. (2007a). Could whales have maintained a high abundance of krill? EvolutionaryEcology Research 9, 1�12.

Willis, J.K. (2007b). Building Models of Pelagic Marine Ecosystems PhD thesis,University of Tasmania.

Willis, J. (2008). Simulation model of universal law of school size distribution applied tosouthern bluefin tuna (Thunnus maccoyii) in the Great Australian Bight. EcologicalModelling 213, 33�44.

Willis, J. and Hobday, A. J. (2007). Influence of upwelling on movement of southernbluefin tuna (Thunnus maccoyii) in the Great Australian Bight. Marine and FreshwaterResearch 58, 699�708.

Wilson, D. S. and Dugatkin, L. A. (1997). Group selection and assortative interactions.American Naturalist 149, 336�351.

Wilson, D. S. and Sober, E. (1994). Reintroducing group selection to the human behav-ioral sciences. Behavioral and Brain Sciences 17, 585�654.

Wilson, E. O. (1975). Sociobiology. Harvard University Press, Boston, MA.Wiszniewski, J., Allen, S. J. and Moller, L. M. (2009). Social cohesion in a hierarchically

structured embayment population of Indo-Pacific bottlenose dolphins. Animal Behavior77, 1449�1457.

Wolf, J. B. W., Mawdsley, D., Trillmich, F. and James, R. (2007). Social structure in acolonial mammal: Unravelling hidden structural layers and their foundations by net-work analysis. Animal Behaviour 74, 1293�1302.

Yen, J. and Bundock, E. A. (1997). Aggregative behavior in zooplankton: Phototacticswarming in four developmental stages of Coullana canadensis (Copepoda,Harpacticoida). In Animal Groups in Three Dimensions: How Species Aggregate (J. K.Parrish and W. M. Hamner, eds), pp. 143�162. Cambridge University Press,Cambridge.

Yen, J., Brown, J. and Webster, D. R. (2003). Analysis of the flow field of the krill.Euphausia pacifica. Marine and Freshwater Behavior and Physiology 36, 307�319.

Young, S., Watt, P. J., Grover, J. P. and Thomas, D. (1994). The unselfish swarm? Journalof Animal Ecology 63, 611�618.

Young, J. W., Bradford, R. W., Lamb, T. D. and Lyne, V. D. (1996). Biomass of zooplank-ton and micronekton in the southern bluefin tuna fishing grounds off easternTasmania, Australia. Marine Ecology Progress Series 138, 1�14.

Zhou, M. and Tande, K. (eds) (2002). Optical Plankton Counter Workshop. GLOBECReport 17, 1�67pp.


[advances in marine biology] advances in marine biology volume 60 volume 60 || social aggregation in...

Documents