You are on page 1of 44

F I S H and F I S H E R I E S, 2003, 4, 147^190

Compensatory growth in fishes: a response to growth


depression
M Ali1,y, A Nicieza2 & R J Wootton1,
1

Institute of Biological Sciences, University of Wales Aberystwyth, Aberystwyth SY233DA,Wales, UK; 2Departamento de

Biolog| a de Organismos y Sistemas, Universidad de Oviedo, E-33071 Oviedo, Spain

Abstract
Compensatory growth (CG) is a phase of accelerated growth when favourable conditions are restored after a period of growth depression. CG reduces variance in size by
causing growth trajectories to converge and is important to sheries management,
aquaculture and life history analysis because it can oset the eects of growth arrests.
Compensatory growth has been demonstrated in both individually housed and
grouped sh, typically after growth depression has been induced by complete or partial
food deprivation. Partial, full and over-compensation have all been evoked in sh,
although over-compensation has only been demonstrated when cycles of deprivation
and satiation feeding have been imposed. Individually housed sh have shown that CG is
partlya response to hyperphagiawhen rates of food consumptionare signicantly higher
thanthose in shthat have notexperiencedgrowthdepression.The severityof thegrowth
depression increases the duration of the hyperphagic phase rather than maximum daily
feeding rate. In many studies, growth eciencies were higher during CG. Changes in
metabolic rate and swimming activity have not been demonstrated yet to playa role.
Periods of food deprivation induce changes in the storage reserves, particularly lipids,
of sh. Apart from the strong evidence for the restoration of somatic growth trajectories, CG is a response to restore lipid levels. Although several neuro-peptides, including neuropeptide-Y, are probably involved in the control of appetite, their role and the
role of hormones, such as growth hormone (GH) and insulin-like growth factor (IGF),
in the hyperphagia associated with CG are still unclear.
The advantages of CG probably relate to size dependencies of mortality, fecundityand
diet that are characteristic of teleosts. These size dependencies favour a recovery from
the eects of growth depression if environmental factors allow. High growth rates
may also impose costs, including adverse eects on future development, growth, reproduction and swimming performance. Hyperphagia may lead to riskier behaviour in
the presence of predators. CGs evolutionary consequences are largely unexplored. An
understanding of why animals grow at rates below their physiological capacity, an evaluation of the costs of rapid growth and the identication of the constraints on growth
trajectories represent major challenges for life-history theory.


Correspondence:
R J Wootton,
Institute of
Biological Sciences,
University of Wales
Aberystwyth,
Aberystwyth SY23
3DA,Wales, UK
Tel.:
44 1970 622346
Fax:
44 1970 622350
E-mail:
rjw@aber.ac.uk

Present address:
Institute of Pure &
Applied Biology,
Bahauddin
Zakariya
University, Multan,
Pakistan

Received15 July 2002


Accepted 26 Feb 2003

Keywords growth eciency, growth regulation, hyperphagia, storage levels, trade-os

Introduction

148

Characteristics of compensatory growth

149

Degree of compensation

149

Focus of compensation

150

# 2003 Blackwell Publishing Ltd

147

Compensatory growth in shes M Ali et al.

Taxonomic distribution of compensatory response in teleosts

150

Methodological problems in studies on compensatory growth

155

Experimental studies

155

Field studies

157

Factors affecting compensatory growth in fish

158

Eect of feeding protocols including length and intensity of deprivation

158

Inuence of social factors

162

Ontogenetic changes

162

Seasonal variation

163

Sexual maturation and reproduction

164

Compensatory growth of body components

165

Proximate causes of compensatory growth

165

Hyperphagia

166

Responses to free access to food after a period of growth depression

166

Eects of food quality on hyperphagic response

168

Physiological basis of hyperphagia: consumption, absorption and evacuation rates

168

Physiological basis of compensatory growth: metabolic rate

169

Growth eciency

170

Proximate composition, metabolites and endocrines during food deprivation


and recovery

171

Lipids

171

Protein and RNA:DNA ratios

172

Carbohydrates

173

Water content

173

Endocrines

173

Modelling compensatory growth

174

Structural and storage model

174

Model based on specic growth rate

175

Functional significance of compensatory growth

176

Size-dependent mortality

176

Size-dependent prey selection

176

Size-dependent ontogenetic changes

176

Size-dependent reproductive success

177

Costs of compensatory growth

177

Compensatory growth as a by-product of flexible growth and dynamic trade-offs

178

Future directions for research

179

Acknowledgements

181

References

181

Introduction
Many organisms exhibit faster growth during recovery from total or partial food deprivation than they
do during periods of continuous food availability
(Wilson and Osbourn 1960; Jobling1994). The consequence is that animals experiencing a period of
148

growth depression may achieve the same size-at-age


as conspecics experiencing environmental conditions that are more favourable. The response, which
tends to restore the original growth trajectory, is
called compensatory, or catch-up, growth. The phenomenon raises some important, but largely unanswered, questions in the elds of evolutionary and
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

applied ecology.Why do animals sometimes grow at a


rate below their physiological potential? What are
the implications for the development of life-history
theory? What are the implications for the management of exploited or endangered species?
Compensatory growth has mostly been studied in
domesticated endotherms, although it has been
observed in a range of invertebrates and vertebrates
(Wilson and Osbourn 1960; Sibly and Calow 1986).
Teleosts (referred to hereafter as sh, unless otherwise qualied) are ectotherms with patterns of indeterminate growth, which allow the examination of
the compensatory process at almost every stage of
their life cycles. Compensatory growth has been studied, at least supercially, in more sh species than
in any other vertebrate taxa. Research on sh has
also pioneered the analysis of the eects of growth
depressors other than starvation, including unseasonably low temperatures, exposure to hypoxia or
prophylactics, and reproductive eort. There are also
practical reasons for studies of growth compensation
in sh. The group includes many species of economic
and social importance because of exploitation in
commercial or recreational sheries, or use in aquaculture. This review provides a framework on which
further studies can build in a more integrated way
than has characterised studies on compensatory
growth so far.
Compensatory growth is identied by being significantly faster than the growth rate of control animals
that have not experienced growth depression, held
under comparable conditions. This accelerated
growth eventually declines to growth rates typical
of the control animals. Compensatory growth has
also been dened as any growth that reduces the variance in the system (Atchley 1984) or as the negative
correlation between the growth of a trait in successive time intervals (Ricker 1969, 1975; Riska et al.
1984). Growth may be controlled by feedback
mechanisms, which adjust growth rates to achieve a
target trajectory (Tanner 1963; Monteiro and Falconer 1966; Hubbell 1971). An important consequence of growth compensation is the convergence
of the growth trajectories of individuals, which
reects the target-seeking consequence of the
catch-up growth. It is a process that tends to canalise
ontogenetic changes in size, and to buer the eects
of environmental variability (Wootton1998).
In the literature on sh growth, the term compensatory growth has been used in several ways:
1 To describe the accelerated growth by an individual after a period of growth depression, or more
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

generally, the negative correlation between


growth rates in successive time periods (Ricker
1975 and references therein).
2 To describe the response of increased individual
growth rates in a population after a reduction in
population density has increased food supply
(density-dependent; Ferreri and Taylor1996). Such
an increase is predictable from the known relationship between growth and feeding rate in sh
(Weatherley and Gill 1987; Wootton 1998). Such
competitive release is often observed in predation
experiments (Wootton1998).
3 To describe the growth resulting from an extended
period of feeding prior to some size-dependent
life-cycle switch, such as smoltication in salmonids (Nicieza and Brana1993a,b).
Only the rst form of compensatory growth is the
primary subject of this review. It is still unclear what
role compensatory growth plays in natural populations, or whether the phenomenon can be used in the
context of aquaculture to optimise sh production.
Characteristics of compensatory growth
Accelerated growth in response to previous food
restriction provides evidence that growth rates are
regulated. It suggests that animals can evaluate
their achieved growth trajectory and adjust growth
rates to buer deviations from an ideal trajectory
(Hubbell 1971; but see Broekhuizen et al. 1994). The
expression of this growth regulation is dependent on
several factors (Wilson and Osbourn 1960; Ryan
1990). These include the nature, severity and duration of the under-nutrition, the stage of development
at the start of under-nutrition, the age at sexual
maturity and the pattern of re-alimentation.
Degree of compensation
In terms of a targeted growth trajectory (Monteiro
and Falconer1966; Riska et al.1984), the consequence
of a compensatory growth response is the attainment
of a size status relative to the size achieved by an
organism that has not experienced any phase of
growth impairment. Implicitly, the size of the latter
organism is assumed to be optimum at every time, so
the ratio between the size of compensating and control animals when the compensatory response has
abated provides a measure of the eectiveness of the
compensation. In full compensation, the deprived
animals eventually achieve the same size at the same
age as continuously fed contemporaries (Fig. 1). In
149

Compensatory growth in shes M Ali et al.

Figure 1 Idealised patterns of


growth compensation based on
Jobling (1994).

partial compensation, the deprived animals fail to


achieve the same size at the same age as nonrestricted contemporaries, but do show relatively rapid
growth rates, and may have better food conversion
ratios during the re-alimentation period (Fig. 1). Overcompensation occurs when the animals that had
experienced a restricted ration achieve a greater size
at the same age than nonrestricted animals (Fig. 1).
The compensation is so strong that animals subject
to variable food supplies exhibit a higher growth
rate than those for whom food was continuously
available. This was observed in Lepomis hybrids by
Hayward et al. (1997), but seems to be a rare outcome.
If the re-alimented sh resume growing at a rate characteristic of the size reached at the end of the deprivationperiod, no compensation has occurred (Fig. 1).

primary body components. The eects of nutritional


conditions on body composition and physiological
state of teleosts are well understood (e.g. McMillan
and Houlihan 1992; Jobling 1993; Mommsen 1998).
The inuence of undernutrition on the growth
dynamics of length and mass can dier substantially
(for instance, mass loss is often associated with no or
negligible decrease in length). Therefore, the
dynamics of growth recovery may also be dierent,
both in the degree and in the time of compensation
(e.g. Nicieza and Metcalfe 1997). In particular, there
is evidence that recovery of lipid or energy levels
can operate through a pathway separate from that
leading to compensation for structural growth (Bull
et al. 1996; Nicieza and Metcalfe 1997; Metcalfe et al.
2002).

Focus of compensation

Taxonomic distribution of compensatory


response in teleosts

In sh, the term compensatory growth is used


usually to describe increases in growth rates in
whole body length or mass. However, the relative
growth of body components can be altered by manipulation of environmental conditions, thus producing
shape modication (Emerson 1986). The organism
can experience a reduced growth of some body components without signicant alteration in its whole
size. It remains unknown whether there can be compensation for such uncoupling in the growth of dierent body parts with the normal shape restored.
Compensatory responses can be examined in relation to the growth of dierent tissues, organs or
150

The term compensatory growth was rst used in


relation to mammals and has subsequently been
demonstrated in a range of domesticated endotherms
(Wilson and Osbourn 1960). Earlier studies on sh
had reported negative correlations between growth
rates in successive periods (review in Zivkov 1982).
Despite these early observations, there were few studies on compensatory growth in sh until the early
1990s when the subject attracted attention, especially in relation to aquaculture. The occurrence of
compensatory growth has been reported for a
restricted number of sh taxa (Table 1). Even so, sh
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Table 1 Taxonomic distribution of studies on growth compensation in teleosts and incidence of full compensation (FC), partial compensation (PC), over-compensation (OvC) and absence of
compensation (AC), where information is available.

Methods/experimental protocol

Manipulated factor

Effect

Response variable

Reference

Clupeidae
Clupea harengus (Herring)
C. harengus
C. harengus

Group-housed, posthatch larvae


Group-housed, posthatch larvae
Natural population, back-calculation

Food supply
Food supply
None

FC
AC
PC

Length, soluble protein


Length
Length

Pedersen et al. (1990)


Pedersen (1993)
Moores and Winters (1982)

Salmonidae
Salmo salar (Atlantic salmon)
S. salar
S. salar
S. salar
S. salar
S. salar
S. salar
S. salar
S. salar

Group-housed, mature vs. immature parr


Group-housed
Group-housed, juveniles
Group-housed, preadults
Group-housed
Group-housed
Natural population, back-calculation
Group-housed, juvenile
Group-housed

None
Food supply
Food supply
Food supply
Temperature
Salinity
None
None
Food supply,
temperature
Food supply
Temperature
Food supply
Food supply

FC
PC
FC
PC
PC
FC
PC
PC
FC, PC

Length
Mass
Lipid content
Mass
Mass
Mass
Length
Length
Length, mass

Skilbrei (1990)
Thorpe et al. (1990)
Metcalfe and Thorpe (1992)
Reimers et al. (1993)
Mortensen and Damsgard (1993)
Damsgard and Arnesen (1998)
Nicieza and Brana (1993a,b)
Nicieza et al. (1994a)
Nicieza and Metcalfe (1997)

FC
FC
PC, FC
PC, FC

Lipid content
Length
Length, lipid content

Bull and Metcalfe (1997)


Maclean and Metcalfe (2001)
Morgan and Metcalfe (2001)
Johansen et al. (2001)

Food
Food
Food
Food

PC, FC, OvC


FC
FC
FC

Body components
Mass
Mass
Mass

Weatherley and Gill (1981)


Dobson and Holmes (1984)
Smith (1987)
Kindschi (1988)

PC
FC,PC
FC

Mass, body components


Mass
Mass
Mass

Quinton and Blake (1990)


Teskeredzic et al. (1995)
Jobling and Koskela (1996)
Speare and Arsenault (1997)

S. salar
S. salar
S. salar
S. salar

Group-housed,
Group-housed,
Group-housed,
Group-housed,

juveniles
juveniles
juveniles
postsmolts

Oncorhynchus mykiss (rainbow trout) Group-housed


O. mykiss
Group-housed
O. mykiss
Group-housed
O. mykiss
Group-housed, cycles of
deprivation and refeeding
O. mykiss
Group-housed
O. mykiss
Group-housed
O. mykiss
Group-housed
O. mykiss
Group-housed

151

Oncorhynchus kisutch
(Coho salmon)
Oncorhynchus nerka
(Sockye salmon)

supply
supply
supply
supply

Food supply
Food supply
Food supply
H2O2 prophylactic
treatment

Group-housed

Food supply

FC

Mass

Damsgard and Dill (1998)

Group-housed

Food supply

PC, FC

Mass

Bilton and Robins (1973)

Compensatory growth in shes M Ali et al.

Species (by family)

Species (by family)

Salvelinus alpinus (Arctic charr)


S. alpinus
S. alpinus
First feeding Coregonid larva
Characiformes
Piractus brachypomus
Cyprinidae
Cyprinus carpio (carp), Leuciscus
cephalus, Alburnus alburnus,
Rutilus rutilus (roach) Carassius
auratus (goldfish)
Cyprinus carpio, Leuciscus cephalus,
Rutilus rutilus, Abramis brama
Cyprinus carpio
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Tor putitora
Labeo rohita,
Cirrhina mrigala, Catla catla
Cyprinus carpio
Cyprinus carpio
Scardinus erythropthalmus (rudd),
Leuciscus cephalus (chub)
Chalcalburnus chalcoides mento
(Danubian bleak)
Phoxinus phoxinus
(European minnow)

Methods/experimental protocol

Manipulated factor

Effect

Response variable

Reference

Individually housed
Group-housed, cycles of deprivation and refeeding
Group-housed

Food supply
Food supply
Temperature

PC
PC
PC

Mass
Mass
Mass

Miglavs and Jobling (1989a,b)


Jobling et al. (1993)
Mortensen and Damsga rd (1993)

Group-housed

Food supply

Mass

Dabrowski et al. (1986)

Group-housed

Food supply

FC

Mass

Koppe et al. (1993)

PC, FC

Length

Zivkov (1982)

Length

Zivkov (1996)

Mass

Schwarz et al. (1985)

Natural populations, back-calculation

Natural populations, back-calculation


Group-housed:
protein or energy deprived
Group-housed
Group-housed

Food supply

NC

Tandon and Johal (1983a)


Tandon and Johal (1983b)

Group-housed
Group-housed
Group-housed

Food supply
Food supply

Mass

Johal and Tandon (1986)


Bastrop et al. (1991).
Wieser et al. (1992)

Individually housed

Food supply

NC, FC

Mass

Russell and Wootton (1992)

Phoxinus phoxinus
Rutilus rutilus

Individually housed
Group-housed

Food supply
Food supply

FC

Zhu et al. (2001)


Mendez and Wieser (1993)

Group-housed

Food supply

FC

Mass
Mass, enzyme
concentrations
Mass

Carassius auratus gibelio


(Gibel carp)
C. auratus gibelio
Abramis brama

Individually housed
Group-housed

Food supply

FC

Mass

Xie et al. (2001)


Raikova-Petrova and Zivkov (1998)

Qian et al. (2000)

Compensatory growth in shes M Ali et al.

152

Table 1 continued

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Ictaluridae
Ictalurus puntatus, Channel catfish
I. punctatus
I. punctatus
Heterobranchus longifilis
(African catfish)

Group-housed
Group-housed, cycles of deprivation
Group-housed, cycles of deprivation
Group-housed, single deprivation and cyclic
deprivation

Cyprinodontidae
Cypronodon nevadensis

Group-housed

Gadidae
Gadus morhua (Atlantic cod)
G. morhua

Food
Food
Food
Food

supply
supply
supply
supply

FC
FC
NC
PC, FC

Mass
Mass
Mass
Mass

Kim and Lovell (1995)


Gaylord and Gatlin (2001)
Gaylord et al. (2001)
Luquet et al. (1995)

Gerking and Raush (1979)

Food supply

NC, FC

Mass

Beacham (1981)
Jobling et al. (1994)

Full-compensation

Food supply

FC

Mass

Zhu et al. (2001)

Cichlidae
Oreochromis mossambicus
O. mossambicus
O. mossambicus
O. mossambicus
O. niloticus
O. niloticus
O. mossambicus  O. niloticus hybrids

Group-housed
Group-housed
Group-housed
Group-housed
Group-held, juveniles
Group-housed
Group-housed in sea water

Temperature

Gonad size

Sparidae
Pagrus pagrus

Group-held, demand feeders

Carcharhinidae
Trichiurus japonicus

Natural population

Gasterosteidae
Gasterosteus aculeatus
(three-spined stickleback)

Food supply
Density
Density
Food supply

FC
FC
PC, FC

Mass
Mass
Mass

Chmilevskii (1994)
Chmilevskii (1998)
Chmilevskii (1996)
Christensen and McLean (1998)
Vera-Cruz and Mair (1994)
Basiao et al. (1996)
Wang et al. (2002)

FC

Mass

Rueda et al. (1998)

Length

Haweet et al. (1996)

Centropomidae
Lates calcarifer

Aranyakananda et al. (1996)

Anarhicadidae
Anarhicas minor

Group-housed

Oxygen concentration

Mass

Foss and Imsland (2002)

Percichthyidae
Golden perch, Macquaria ambigua

Group-housed

Food supply

Lipid, protein

Collins and Anderson 1995

153

Compensatory growth in shes M Ali et al.

Natural population
Group-housed, cycles of deprivation and
single deprivation

Species (by family)

Methods/experimental protocol

Manipulated factor

Effect

Response variable

Reference

Individually housed; cycles of deprivation


and refeeding
Group-housed

Food supply

PC, FC

Mass

Hayward and Wang (2001)

Centrarchidae
Lepomis cyanellus  L. macrochirus
(hybrid sunfish)
L. cyanellus  L. macrochirus

Individually housed; cycles of deprivation


and refeeding
Group-housed; cycles of deprivation and feeding

Food supply

OvC, FC, PC

Mass

Food supply

PC

Mass

Pomatomidae
Pomatomus saltatrix

Group-housed

Percidae
Perca flavescens Yellow perch

Stizostedion lucioperca (pike-perch)

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Pleuronectidae
Pleuronectes americanus
(Winter flounder)
P. americanus
Hippoglossus hipoglossus
Pleuronectus asper
(Alaska yellowfin sole)

Zivkov and Raikova-Petrova (1991)

Group-housed, pre- and postmetamorphic


Group-housed
Group-housed

Buckel et al. (1998)

PC

Group-housed

Food supply
Food supply

PC

Hayward et al. (1997) Whitledge


et al. (1998)
Hayward et al. (2000)

Length

Chambers et al. (1988)

Length

Bertram et al. (1993).


Bjo rnsson et al. (1992)
Paul et al. (1995)

Mass

Compensatory growth in shes M Ali et al.

154

Table 1 continued

Compensatory growth in shes M Ali et al.

studies constitute the best source of information on


compensatory growth in ectotherms.
A disproportionate number of studies of compensatory growth have focused on six salmonid species.
The number of studies describing compensatory
growth in cyprinids is lower, but there are data on
about 13 species. To our knowledge, there are only a
few studies on compensatory growth for some taxonomic groups whose biology is well researched (e.g.
Poeciliids, Cichlids). The lack of information is even
more striking for strictly marine species. Moreover,
there is evidence of interspecic variation in patterns
of growth compensation, so species having similar
geographical ranges, similar social behaviour, and
similar diets can have distinct compensatory capacities and dierent mechanisms of compensation
(e.g. Sogard and Olla 2002).This accentuates the need
for a wider taxonomic coverage in the study of compensatory growth. A lack of comparative studies
using the same protocols means that it is unclear
whether these reports are describing the same phenomenon or whether several dierent causal pathways can lead to the observation of unusually high
growth rates.
Methodological problems in studies on
compensatory growth
Experimental studies
An evaluation of the experimental evidence for compensatory growth is hampered by the prevalence of
poor experimental design and inappropriate statistical analyses, which obscure the valid identication
of a compensatory response. A common aw is the
misidentication of the appropriate level of replication in relation to the detection of experimental
eects, leading to the problem of pseudoreplication
(see Underwood 1997 for an excellent discussion).
The use of grouped sh in experiments can be justied on two grounds. First, some species characteristically live in shoals and, if kept in isolation, show
pathological behaviour with reduced appetite and
poor growth (e.g. Stirling 1976). Second, many
experiments on compensatory growth have been in
relation to aquaculture, a context in which sh are
maintained at high densities. However, if grouped
sh are used, it has to be recognised that the unit of
replication for testing an experimental treatment is
the tank, not the individual sh in the tank, even if
individuals are identiable. The misuse of individual
sh as the unit of replication leads to an ination of
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

the degrees of freedom used for testing the experimental eect (Ruohonen et al. 2001).
The use of grouped sh also introduces two more
factors that are usually ignored in the analysis and
interpretation: density and numerical abundance.
Any conclusions from a study will strictly be valid
only for the density of sh used unless density is
deliberately manipulated as an experimental factor.
If density is manipulated, care needs to be taken to
avoid changing numerical abundance simultaneously. Equally, an experimental manipulation of
numerical abundance per group as a factor must
avoid changing density. In some studies of compensatory growth, the numerical size of groups in tanks
has been reduced during an experiment to provide
intermediate samples. This manipulation introduces
another confounding factor, changing both density
and numerical abundance simultaneously. The use
of individually housed sh, if they do not show pathological behaviours, avoids all these problems.
Growth and consumption rates have an allometric
relationship to body mass in sh (Jobling 1994;
Wootton 1998). Size-specic rates tend to decline
with size. Fish that have experienced growth depression will be smaller than control sh, and so may
exhibit higher specic growth and consumption
rates simply because of allometry. Consequently,
comparisons for these variables between control and
experimental sh should be made for sh of approximately the same size, either by experimental manipulation or by covariance adjustment.
A third problem, particularly with earlier studies,
is the failure to use appropriate statistical analyses.
A common error was the use of serial t-tests to make
multiple comparisons where analysis of variance
with preplanned contrasts was required (Underwood
1997). In some studies, a series of analyses of variance
have been used, where repeated-measures anova
should have been used to analyse changes in growth
and consumption over time.
These problems mean that some highly cited studies may not have demonstrated real growth compensation, but simply the plasticity of sh growth in
relation to environmental variables, including food
availability, sh density and abundance, social interactions and even unique tank eects. Table 2 provides a critique of methodologies of some studies
considered in further detail below. Despite these diculties, which are not unique to studies of compensatory growth, the body of experimental evidence
suggests that compensatory growth in sh is a real
phenomenon.
155

Species studied

Experimental design

Statistical analysis

Reference

Oncorhynchus nerka

No replication of treatments. Group-housed, but


individuals within tanks treated as valid replicates.
Densities changed during treatment.
No replication of treatments. Group-housed, but
individuals within tanks treated as valid replicates.
Densities changed during experiment.
Inadequate replication with unbalanced design.
Group-housed, but individuals within tanks
treated as valid replicates.
Individually housed fish as valid replicates.

Invalid use of serial t-tests to compare treatments

Bilton and Robins (1973)

ANOVA and multiple range tests. No adjustment


for size differences.

Weatherley and Gill (1981)

Invalid use of t-test. No adjustment for size effects.

Dobson and Holmes (1984)

Invalid use of serial t-tests. No adjustment for


size differences
Invalid use of KruskalWallis and t-tests.
No adjustment for size effects.

Miglavs and Jobling(1989a,b)

O. mykiss

O. mykiss

Salvelinus alpinus
O. mykiss

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Salmo salar

Phoxinus phoxinus

Chalcalburnis chalcoides, Leuciscus


cephlaus, Scardinius erythrothalamus
S. salar
S. salar and Salvelinus alpinus
S. alpinus

Ictalurus punctatus
Pleuronectes platessa

No replication of treatments. Group-housed, but


individuals within tanks treated as valid replicates.
Densities changed during experiment.
No replication of treatments. Group-housed, but
individuals within tanks treated as valid
replicates. Indirect estimate of rate of
food consumption
Individually housed fish as valid replicates.

No replication of treatments. Group-housed,


but individuals within tanks treated as valid replicates.
No replication of treatments.
No replication of treatments. Group-housed, but
individuals within tanks treated as valid replicates
Group-housed, with three tanks per treatment, tank
means used as valid replicates. Densities
changed during experiment.
No replication of treatments. Group-housed,
but individuals within tanks treated as valid replicates.
No replication of treatments. Group-housed, but
individuals within tanks treated as valid replicates.

Quinton and Blake (1990)

Invalid use of serial t-tests.

Metcalfe and Thorpe (1992)

Size adjustment by covariance, preplanned


comparisons, but invalid use of serial rather
than repeated measures ANOVA.
Invalid use of ANOVA

Russell and Wootton (1992)

Wieser et al. (1992)

Invalid use of ANOVA with multiple comparisons.

Reimers et al. (1993)


Mortensen and Damsgard (1993)

Valid use of KruskalWallis, but invalid use of


serial MannWhitney U-tests.

Jobling et al. (1993)

Invalid use of ANOVA and multiple range tests

Kim and Lovell (1995)

Invalid use of serial t-tests

Paul et al. (1995)

Compensatory growth in shes M Ali et al.

156

Table 2 Survey of experimental designs and analyses in a selection of studies on compensatory growth in sh.

Compensatory growth in shes M Ali et al.

Zhu et al. (2003)

Studies are presented in historical order.

Gasterosteus aculeatus

S. salar

No replication of treatments. Group-housed, but


individuals within tanks treated as valid replicates
Individually housed as valid replicates

Repeated measures with ANCOVA adjustment for


size. Pre-planned orthogonal contrasts. Failure to
apply Bonferroni adjustment for multiple comparisons.

Metcalfe et al. (2002)

Zhu et al. (2001)

Gasterosteus aculeatus,
P. phoxinus

O. mykiss

Group-housed with 3 tanks per treatment in used


refeeding phase, with tank means as valid replicates
Individually housed fish as valid replicates.

Repeated measures, with ANCOVA adjustment


for size, preplanned, but nonorthogonal contrasts.
Failure to apply Bonferroni adjustment for multiple
comparisons
Invalid use of ANOVA and t-tests.

Boujard et al. (2000)

Saether and Jobling (1999)

Repeated measures ANOVA, but contrasts not


preplanned. Failure to apply Bonferroni
adjustment for multiple comparisons
ANOVA and size-adjustment by ancova

Scophthalmus maximus

Salmo salar

No replication of treatments. Group-housed, with all


treatments combined during re-alimentation phase,
but individuals treated as valid replicates
Group-housed, with two tanks per treatment, tank
means used as valid replicates

Repeated measures, with ANCOVA


adjustment for size.

Nicieza and Metcalfe (1997)

Field studies

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

There are at least three possible approaches to the


study of compensatory growth under natural conditions. Observational studies compare the responses
of animals of equivalent reproductive and developmental status, whose growth history was aected
dierentially by some identied or unknown environmental factor. Experimental eld studies address
compensatory responses that result from the manipulation of the environment either to induce growth
depression in some individuals or to relax an existing
factor constraining growth rates such as a high risk
of predation. A third approach is to analyse growth
rates under natural conditions after exposure of the
experimental animals to controlled conditions in the
laboratory.
The demonstration of compensatory growth in the
wild requires estimates of the size of individuals or
groups of individuals at dierent times. Recapture
rates of individually marked sh must be suciently
high to generate adequate sample sizes, which
requires tagging large numbers of sh. The eld site
used must not impose conditions such as low temperatures or food limitation that preclude the detection of compensatory responses. An additional
diculty in manipulative eld studies is to obtain
individuals of equivalent developmental status, but
diering in their recent growth history. Indirect
methods, including the back-calculation of size at
known ages, can avoid the costs and diculties associated with tracking marked individuals (Nicieza
and Brana 1993a,b), but have limitations. They are
only applicable to some species, do not allow the precise denition of the time intervals, and provide no
information about the environmental factors inuencing growth dynamics. These diculties associated with eld studies raise doubts about the
relevance of compensatory responses in natural
populations.
Most of the empirical evidence for compensatory
growth comes from manipulative experiments, in
which a reduction of food intake is imposed either
directly or indirectly. To assess the role that compensatory growth can play as an important process of
growth regulation in natural populations, a rst step
is to dene the constraints in the habitat and ecology
of the population that might evoke a compensatory
response. The spatio-temporal variation in the environment must be sucient for a variable fraction of
the population to experience a potentially reversible
reduction in growth rates. If the entire population
157

Compensatory growth in shes M Ali et al.

experiences an unpredictable reduction in growth


rates, the demonstration of compensatory growth
will be dicult because of the problem of identifying
an undisturbed growth trajectory. In this situation,
an alternative method of detecting a process of compensation would be to use an intercohort analysis of
growth patterns over a time series for which data on
relevant environmental variables are available.
Another factor that indirectly induces growth
depression is density. In natural conditions, the compensatory adjustment would require the existence of
some behavioural mechanism (e.g. condition-dependent dispersal).
Factors affecting compensatory
growth in fish
Most studies in sh have provoked compensatory
growth bya period of total or partial food deprivation.
The consequences of changing food qualityon growth
are also susceptible to compensation (Bjornsson et al.
1992). It has also been induced after a period of
unseasonably low temperatures, which reduced the
rate of food consumption in juvenile Atlantic salmon,
Salmo salar (Salmonidae; Mortensen and DamsgQrd
1993; Nicieza and Metcalfe 1997; Maclean and
Metcalfe 2001) or Atlantic cod, Gadus morhua (Gadidae; Purchase and Brown 2001) and during treatment of diseased rainbow trout, Oncorhynchus
mykiss (Salmonidae) with hydrogen peroxide (Speare
and Arsenault 1997). Atlantic salmon smolts showed
growth compensation after a short period of growth
depression associated with a move from fresh to sea
water (DamsgQrd and Arnesen 1998). Spotted wolfsh, Anarhichas minor (Anarhichadidae), displayed
compensatory growth after being returned to normal oxygen conditions following 75 days in hypoxic
conditions (Foss and Imsland 2002). In other studies,
the origin of the reduced growth and the subsequent
compensation has not been clear (Zivkov 1982, 1996;
Zivkov and Raikova-Petrova 1991; Nicieza and Brana
1993a,b; Nicieza et al.1994a).
Eect of feeding protocols including length
and intensity of deprivation
Experimental studies on compensatory growth
using food deprivation to induce growth depression
can be classied along several dimensions. The
experiments have used either individually or grouphoused sh as the unit of replication. Deprivation
has either been total or partial. There has been a
158

Figure 2 Full growth compensation demonstrated by


individually housed, juvenile gibel carp, Carassius auratus
gibelio, after periods of total food deprivation. , controls
(fed to satiation daily); &, deprivation for1 week; ~,
deprivation for 2 weeks prior to re-alimentation. Means for
9^10 sh. (a) Growth in mass; (b) mean daily specic growth
rate, Gw (% day1).

single period of deprivation, or periods of deprivation


and refeeding have alternated in cycles. Not surprisingly, this diversity of protocols and in some studies
poor experimental design and analysis have led to a
diversity of reported results.
Clear evidence of the pattern of compensatory
growth, with either full or partial compensation
(Fig. 2), has emerged from experiments with
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

individually housed sh exposed to a single period of


deprivation. Full compensation following 1 or
2 weeks of total deprivation was achieved by two
size-classes of juvenile European minnow, Phoxinus
phoxinus (Cyprinidae; Russell and Wootton 1992;
Zhu et al. 2001), juvenile gibel carp, Carassius auratus
gibelio (Cyprinidae; Xie et al. 2001), and was achieved
or approached by the three-spined stickleback, Gasterosteus aculeatus (Gasterosteidae; Zhu et al. 2001).
In these experiments, the deprived sh reached the
mass of control sh after 2^4 weeks of refeeding.
When two groups of three-spined sticklebacks were
reduced by dierent deprivation protocols to a mean
mass of 80% of control sh fed ad libitum, they
showed full compensation in mass and lipid concentration (Zhu et al. 2003). The trajectories of recovery
were similar for both groups. In the minnows, 4 days
of deprivation were not sucient to evoke a detectable compensatory response (Russell and Wootton
1992), and after a week of deprivation, the response
was weak but still sucient to restore the mass trajectory (Zhu et al. 2001).
In the pioneering study using individually housed
sh, Miglavs and Jobling (1989a,b) imposed 8 weeks
of restricted feeding at a level close to maintenance
ration on 1-year-old Arctic charr, Salvelinus alpinus
(Salmonidae). On re-alimentation, the deprived charr
showed a compensatory response, but failed to reach
the mass of control sh before the compensatory
response abated. Minnows fed a maintenance ration
for 16 days did achieve full compensation on being
returned to satiation feeding, but the compensatory
response declined earlier than that shown by minnows experiencing total deprivation for 16 days
(Russell and Wootton1992).
The eects of a single period of deprivation on
group-housed sh have been more variable and not
always consistent within taxa. Sockeye salmon fry,
Oncorhynchus nerka (Salmonidae), showed compensatory growth after 1^3 weeks of starvation and
reached similar masses to controls within 8 weeks
of ad libitum refeeding. After 4 weeks of starvation,
mortality rates were high and survivors were not able
to catch up with the mass of control sh within the
8 weeks of recovery period (Bilton and Robins 1973).
Rainbow trout, with an initial mass of about10 g, still
showed compensatory growth following 13 weeks of
starvation (Weatherley and Gill 1981). Dobson and
Holmes (1984) also reported compensatory growth
in rainbow trout subjected to total deprivation for
3 weeks followed by 3 weeks of feeding. Nicieza and
Metcalfe (1997) evoked compensatory growth in
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

group-housed Atlantic salmon juveniles in two ways.


One group had a reduced food availability (a ration
that was expected to give about 10% of maximum
growth). The second group experienced a period of
unseasonably low temperatures, which reduced the
rate of food consumption of the fry. At the re-imposition of ad libtum rations and restoration of control
temperatures, the low-temperature sh were smaller
than the food-restricted juveniles. The subsequent
pattern of compensatory growth diered between
the groups. The restricted food group showed full
compensation within 31 weeks. The low-temperature group commenced a phase of compensatory
growth a week later than the food-restricted group
and showed only partial compensationafter31 weeks.
It is not clear whether the dierences were an eect
of the dierent protocols used to achieve a phase of
slow growth or of the intensity of the growth retardation. There is growing evidence that periods of food
shortage or low temperatures can provoke dierent
patterns of compensation. In juvenile Atlantic salmon, a 3-week period of unseasonably low temperatures in early summer induced full compensation,
but the compensatory growth was not displayed until
the following autumn (Maclean and Metcalfe 2001).
Taxonomically close species may have dierent
propensities to show compensatory responses. In
contrast to Atlantic salmon, two studies failed to
demonstrate compensatory growth in brown trout,
Salmo trutta (Salmonidae). One-year-old brown trout
that had been partially deprived of food between June
and November failed to show compensatory growth
when fed to satiation from November to March (Pirhonen and Forsman 1998). In this case, the low temperatures in the winter period and the associated
low appetite may have restricted the scope for detectable compensation. Age 0 juvenile brown trout
maintained at reduced rations for 1 month did not
show growth compensation after release in their
natal stream in two consecutive years (Alvarez 2002).
Growth rates were apparently not constrained after
release. Further studies on the capacity of brown
trout to show compensatory growth are currently in
progress (Hayward, personal communication).
In cyprinids, Wieser et al. (1992) noted an inverse
relationship between the length of starvation period
and the subsequent compensatory growth of grouphoused, early juveniles of three species of cyprinid
(Scardinius erythrophthalmus, Leuciscus cephalus and
Chalcalburnus chalcoides mento). Juvenile roach, Rutilus rutilus (Cyprinidae), housed at temperatures that
ranged from 4 to 33 8C and starved for 3 weeks,
159

Compensatory growth in shes M Ali et al.

showed compensatory growth in the following


3 weeks of ad libitum feeding (van Dijk et al. 2002).
Even at 4 and 12 8C, the roach showed a compensatory growth response, although before the period of
starvation, they had not grown at these low temperatures. One example in which no growth compensation was observed is a study on carp, Cyprinus carpio
(Cyprinidae). Schwarz et al. (1985) studied the eects
of protein or energy restriction on growth performance of carp. Both restricted groups grew more
slowly than the control, but both showed positive
growth rates. The duration of the restriction period
was determined by the time taken to reach a target
mass set in the experiment so that treatment groups
and controls started the re-alimentation period at
approximately the same mass. At the end of re-alimentation period, there were no dierences in carcass composition between protein restricted and
control groups, but the energy-restricted group was
low in dry matter, fat and energy content. During realimentation, previously restricted groups had the
same growth rate as the control group, with no evidence of a compensatory growth response. The lack
of a compensatory growth response may have
reected inadequate rates of feeding in the re-alimentation period, or the eects of the deprivations may
not have been sucient to trigger a compensatory
response.
Marine species also have the capacity for compensatory growth. Atlantic cod showed full compensation when fed to satiation for 10 weeks after 8 weeks
of deprivation (Jobling et al. 1994). Cod that had been
lightest at a given length showed the highest growth
rates during the refeeding period. In the Pleuronectiformes, both full and partial compensation have been
evoked. During early spring, four groups of subadult
Alaska yellown sole, Pleuronectes asper (Pleuronectidae), were fasted for either 0, 2, 4 or 6 weeks at the
beginning of a 12-week experiment and then fed to
satiation. The groups of yellown sole fed continuously or fasted for 2 weeks gained the maximum
mass, 25 and 24%, respectively. Fishes fasted for
either 4 or 6 weeks exhibited signicantly lower
mass gains of 16 and 15%, respectively, over the
12-week period, suggesting that P. asper had a limited
capacity for compensatory growth (Paul et al. 1995).
When 2-year-old turbot, Scophthalmus maximus
(Pleuronectidae), were partially deprived but not
starved of food for 41 days, they showed full compensation after being placed on a high ration for a further
34 days (Saether and Jobling 1999). Spotted wolsh
showed compensatory growth when returned to
160

normoxic conditions after 75 days under hypoxic


conditions (Foss and Imsland 2002). The study was
terminated after 21 days of the normoxic conditions,
and it was unclear whether the compensatory response would have persisted for longer.
Studies in which individually housed sh have
been exposed to cycles of deprivation and refeeding
have provided some of the most compelling evidence
for compensatory growth in teleosts. Most studies of
compensatory growth have used xed periods of
deprivation and refeeding. However, a study of 0
Lepomis cyanellus  L. macrochirus (Centrarchidae)
hybrids used cycles of deprivation and refeeding
(Hayward et al.1997) in which the deprivation periods
were xed at 2, 4, 6, 10 or 14 days, but the length of
the post deprivation period was determined by the
duration of the hyperphagic response that followed
each period of deprivation.With this protocol, sh in
some of the groups experiencing cyclical phases of
feeding and deprivation grew more over a 105-day
period than control Lepomis fed ad libitum daily.
This over-compensation was particularly high in the
2-day deprivation group (Fig. 3). This is the rst
demonstration of such a strong over-compensatory
response. Similar protocols applied to juvenile yellow
perch, Perca avescens (Percidae), failed to evoke
over-compensation (Hayward and Wang 2001), although full compensation was achieved by males

Figure 3 Growth over-compensation demonstrated by


individually housed Lepomis hybrids. , controls (fed to
satiation daily); &, sh experiencing cycles of 2 days of
deprivation followed by a phase of hyperphagia of varying
length. Means for nine sh (redrawn from Hayward et al.
1997).
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

Figure 4 Eect of four cycles of1-week deprivation and


2 weeks of satiation refeeding on growth compensation. ,
controls; &, deprived. (a) Full compensation illustrated by
the three-spined stickleback, Gasterosteus aculeatus. Means
for six sh. (b) Partial compensation illustrated bygibel carp,
Carassius auratus gibelo. Means for15 sh.

experiencing12 days of deprivation per cycle. Growth


over-compensation could not be evoked in theLepomis
hybrids when they were housed in groups of 10 sh
percontainer (Haywardetal.2000).Incontrast,groups
of juvenile channel catsh did showgrowth over-compensation when subjected to cycles of deprivation for
1, 2, or 3 days and then to refeeding in the hyperphagic phase (Chatakondi andYant 2001).
In a comparative study, four cycles of1 week of total
deprivation and 2 weeks of satiation feeding were
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

imposed on individually housed minnows, threespined sticklebacks, gibel carp and Chinese longsnout catsh Leiocassis longirostris (Bagridae; Xie
et al., unpublished data; Wu et al. 2003). At the end of
the experiment, the sticklebacks had achieved full
compensation (Fig. 4a), but the gibel carp, the catsh
and the minnows had shown only partial compensation (Fig. 4b). Such experimental protocols can evaluate the capacity of sh to buer the eects of
alternating periods of high and low food availability.
When three-spined sticklebacks were exposed to 1, 2
or 6 days of deprivation followed by 2 days of satiation feeding for 56 days, the sh showed a partial
compensatory response that increased in expression
over the length of the experiment. The sticklebacks
on the 2-day deprivation regime almost matched the
growth of the control, although fed at half the frequency (Ali and Wootton 2001).
Partial compensatory growth has been reported
for group-housed sh subjected to cycles of deprivation and refeeding. In Arctic charr deprived for either
1, 1.5 or 3 weeks and then fed for the same length of
time in an experiment lasting for 24 weeks, the
deprived sh showed partial compensatory growth.
After 6 weeks of unrestricted feeding at the end of
the experiment, there was no signicant dierence
in the mean masses of the deprived sh, irrespective
of the previous feeding pattern, although they were
smaller than the control sh fed daily throughout
the experiment (Jobling et al. 1993). In contrast,
groups of cod experiencing cycles of 3-week feeding
and 3-week deprivation did not show a compensatory growth response (Jobling et al.1994). Frequently,
symmetrical cycles of deprivation and refeeding do
produce partial compensation (e.g. Aranyakananda
et al.1996; Christensen and McLean1998). The inconsistency of results when using grouped sh is illustrated by laboratory studies on the channel catsh.
When groups were exposed to three cycles, each consisting of 3 days of deprivation and 11 days of satiation feeding, the deprived sh showed full
compensation by the end of the 6 weeks (Gaylord
and Gatlin 2001). But in a comparable experiment in
which catsh experienced 3, 5 or 7 days of deprivation in each 14-day cycle, even the sh experiencing
the lowest level of deprivation failed to show compensation (Gaylord et al. 2001).
The frequency of food provision (or inversely, the
frequency and duration of food shortage) can have a
decisive inuence on sh growth. If the periods of
food deprivation are short and sucient food is available between the starvation episodes, a hyperphagic
161

Compensatory growth in shes M Ali et al.

response when food is present can prevent measurable growth depression; so, the growth patterns of
continuously fed and temporarily deprived sh are
virtually indistinguishable. In this case, behavioural
compensation occurs before feeding restriction can
result in detectable growth depression. This has been
demonstrated in immature three-spined sticklebacks
for deprivation periods with a mean length of 3 days
(Ali and Wootton 1998; Ali et al. 1998). No dierences
in growth, food conversion eciency or body composition were detected between rainbow trout held in
groups fed daily and three times per week. Trout
experiencing two lower rations, being fed twice or
once per week, were signicantly smaller than those
fed daily (Teskeredzic et al. 1995). Compensatory
growth may not be evoked if the food restriction
exceeds a certain severity (Wilson and Osbourn
1960; Ryan 1990). Small, juvenile sockeye salmon
starved for 1^3 weeks caught up with control sh
over 7 weeks of refeeding, but sh starved for longer
did not (Bilton and Robins1973).
Some aspects of the factors controlling the magnitude of the compensatory response remain obscure.
There is a limited understanding of how the compensatory process depends on the factor evoking growth
depression. Growth reductions associated with temperature change can elicit compensatory growth
(Mortensen and DamsgQrd 1993; Chmilevskii 1994;
Nicieza and Metcalfe 1997; Maclean and Metcalfe
2001; Purchase and Brown 2001). However, some
studies have reported important dierences between
the growth patterns observed after a period of low
temperatures and a period of food shortage. The time
lag between cessation of the cause of growth depression and the start of the compensatory response
seems to be greater for temperature than for foodinduced growth arrest (Nicieza and Metcalfe 1997;
Maclean and Metcalfe 2001). Dierent forms of
growth reduction can aect other processes (e.g.
reproductive investment, storage, hormonal control)
dierentially, which suggests that there may be dierences in both the temporal pattern of compensation
and the nal outcome in terms of body size.
Inuence of social factors
Density and social interactions also have consequences for compensatory growth. In juvenile tilapia,
Oreochromis niloticus (Cichlidae), relative growth
losses caused by crowding conditions were rapidly
eliminated when the sh were returned to lower, control densities (Vera-Cruz and Mair 1994; Basiao et al.
162

1996). The pattern of compensation can vary


between individuals depending on their behavioural
status and body condition (Jobling and Koskela
1996). In groups of rainbow trout on restricted
rations, there was a high interindividual variability
in food intake. A dominance hierarchy probably generated this variability by regulating access to food.
The consequence was highly heterogeneous growth
amongst the restricted sh.When the trout were provided with unrestricted rations, they became hyperphagic. Growth compensation was highest in trout
that had shown the poorest growth under restricted
rations. This suggests that with unrestricted rations,
the dominance hierarchy relaxed and the individuals
whose growth had been most suppressed could display high rates of consumption (Jobling and Koskela
1996).
Ontogenetic changes
Comparisons between studies are dicult because of
dierences in experimental protocols, species and
ontogenetic stages. Overall, the studies suggest that
there are levels of deprivation that are too small to
evoke a compensatory response, levels that will provoke a full compensatory response and levels so
severe that full recovery cannot be achieved. These
levels probably change with ontogentic stage and
state of sexual maturation. There have been few studies that have examined ontogenetic changes in the
capacity for compensatory growth within a species.
Because of their small size, rst-feeding teleosts
are usually at a higher risk of starvation than when
theyare older and bigger (Miller et al.1988).The distribution of planktonic prey of larval sh is often patchy
in space and time (Mackas et al. 1985; Owen 1989), so
young sh may encounter alternating periods of
abundant and scarce food (Letcher et al. 1996). If the
period of food deprivation is suciently long, 4^
25 days, depending on species and sh size (Miller
et al. 1988), young sh die as a result of starvation.
Letcher et al. (1996) investigated the eect of the order
of feeding days that preceded starvation on times to
death from starvation and whether there is a size
dependence in time to 50% mortalityand to mass loss
before starvation occurs. The time to starvation in
young yellow perch was insensitive to short-term
(4 days) uctuations in food density. The proportional mass loss up to starvation was independent of
previous maximum sh mass, although sh
exposed to intermittent food availability starved at
relatively heavier masses than sh in constant prey
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

environments. Letcher et al. (1996) argued that the


existence of growth compensation suggested a
strong selection pressure for strategies to overcome
short periods of starvation.
Compensatory growth could be relatively common
during larval development because environmental
uncertainty can cause a delay in the time of rst feeding that is susceptible to further compensation. Dabrowski et al. (1986) starved groups of coregonid larva
from the time of hatching, thus delaying the time of
rst feeding by 3^18 days. Delayed rst feeding for
13 days resulted in 50% mortality.With more moderate starvation (3^11 days), there was a positive correlation between the length of delay (starvation) and
subsequent growth rate during ad libitum refeeding.
Compensatory growth in early feeding stages
could oset dierences in size at hatching that reect
dierences in egg size (e.g. Atlantic salmon; Thorpe
et al. 1984). Catch-up growth was observed in Clyde
herring larvae, Clupea harengus (Clupeidae), exposed
to a low ration (15 copepods larva1 day1) for
10 days after yolk absorption followed by a high
ration (80 copepods larva1 day1) for 3 weeks. After
31 days of feeding, these larvae caught up with highration larvae in standard length and content of
water-soluble protein, without an increase in mortality. However, trypsin and trypsinogen content halved
in the re-alimented larvae compared to that of highration sh (Pedersen et al. 1990). In contrast, Baltic
herring exposed to two sequences of food restriction
for 5 days and re-alimentation for 10 days failed to
show compensatory growth and experienced higher
mortality than controls (Pedersen 1993). Larger eggs
in herring confer a higher survival potential on the
resulting larvae (Blaxter and Hempel 1963). The ability to exhibit catch-up growth found in the Clyde larvae (Pedersen et al.1990) may be linked to the higher
resources in the form of organic matter in these large
larvae compared to larvae from Baltic stock (Pedersen 1993). The protocols for restriction and refeeding
were dierent, so the lack of agreement may reect
methodology. Some species do seem to lack compensatory abilities at early stages of development. For
instance, delayed rst feeding, low temperature and
darkness reduced the intake of food and slowed down
somatic growth in Dover sole, Solea solea (Soleidae),
but there was no evidence of subsequent compensatory growth (Lagardere1989).
Major ontogenetic changes in physiology, morphology and behaviour can facilitate self-regulation
of growth patterns. Individuals that are poor performers during one phase of their life cycle can become
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

equal or superior performers in other phases. In the


winter ounder Pleuronectes americanus (Pleuronectidae) slower growing larvae grew faster as juveniles
(Chambers et al. 1988; Bertram et al. 1993). These
negative correlations between larval and juvenile
growth rates are noteworthy because rearing conditions that aect growth (temperature, density and
food) had no inuence on this response. The compensatory growth reported in these studies was not
induced directly by changes in physical conditions.
Bertram et al. (1993) speculated that the convergence
of growth trajectories was related to size-dependent
mortality and developmental processes. Fish with
longer developmental periods might have more ecient organs, which would result in more rapid
growth.
Ontogenetic niche shifts provide an obvious potential for compensatory growth, as these are often
associated with an improvement in feeding conditions. In brown trout, a shift from an invertebrate to
a piscivorous diet can result in a re-acceleration of
growth (Toledo 1996). Juvenile bluesh, Pomatomus
saltatrix (Pomatomidae), that delay the shift from
invertebrate feeding to piscivory grow initially at a
lower rate than piscivorous bluesh.Yet, they apparently experience a phase of increased consumption
and growth rates (leading to partial compensation
of their comparatively reduced growth) after the
change to a sh diet (Buckel et al. 1998).The seaward
migration of anadromous shes is often assumed to
be an escape from a highly food-limited breeding
habitat to a more favourable growing habitat (Gross
1987; Gross et al. 1988; McDowall 2001). Those sh
that experienced the poorest growth in freshwater
could exhibit fast growth at the sea. Although observations of negative correlations between smolt length
and length increment of Atlantic salmon during the
rst growing season at sea support this idea (Skilbrei
1990; Nicieza 1993; Nicieza and Brana 1993a), assessing the incidence of such interstage compensation
requires further research including manipulative
eld experiments in dierent taxonomic groups.
Seasonal variation
An experiment on grouped juvenile Atlantic salmon
suggested that the nature of the compensatory
response may change seasonally (Metcalfe et al.
2002). The sh were deprived of food in summer and
autumn for periods that were sucient to reduce
the fat reserves of the sh to similar levels. Because
of lower temperatures in autumn, the deprivation
163

Compensatory growth in shes M Ali et al.

period required was longer. On refeeding, the summer group showed compensatory growth both in
length and in fat when compared with control sh.
In contrast, the autumn sh recovered their fat content, but their growth in length did not dier from
control sh.
Sexual maturation and reproduction
Environmental conditions and nutritional status
may have strong modifying eects on maturation
and fecundity in sh (Wootton 1982, 1998). Conversely, the allocation of energy to gonad production
and increased activity associated with reproduction
(e.g. migration, defense, access to mates) results in
growth arrest. However, there have been no systematic experimental studies on compensatory growth in
relation to reproduction, and no coherent picture
has emerged.
In some species, like Oreochromis mossambicus
(Cichlidae), low temperatures reduce growth rates
and delay gonad development. Transfer to optimal
temperatures induces growth compensation, but
maturation is still delayed in comparison to sh that
did not experience the lowered temperatures (Chmilevskii1994). In salmonids, the physiological decision
to mature may be dependent on some measures of
the rate of storage or turnover of surplus energy
exceeding a threshold value during a limited season
(Thorpe 1986). The value for this threshold is not
known, nor is the precise season or its duration,
although there are some clues to both. For example,
lipid levels in winter and spring may inuence the
initiation and progression of maturation of salmon
parr (Rowe et al.1991). A series of studies investigated
the eect of food restriction on the precocial maturation of male Atlantic salmon (Rowe and Thorpe
1990; Thorpe et al. 1990; Rowe et al. 1991). These
experiments found evidence of growth compensation. However, some schedules of deprivation and
refeeding did reduce the proportion of maturing
males. Eects on the proportion of maturing females
were more pronounced. Similarly, Reimers et al.
(1993) starved second sea winter farmed Atlantic salmon through February and March. Starved sh
showed partial compensation in growth, but the incidence of maturation was reduced by 48% among
females and 32% among males. In Arctic charr,
cycles of alternating periods of deprivation and feeding over a 24-week period followed by 6 weeks of
feeding produced no signicant dierences in the
proportion of mature males at the end of the experi164

ment (Jobling et al. 1993). Irrespective of feeding


regime, maturing males were signicantly smaller
than immatures when the experiment terminated.
The existing information indicates that, among
marine species, the eects of reduced food availability on growth and reproduction are similar to those
reported for freshwater or anadromous species.
One-year-old farmed Atlantic cod were subjected to
periods of short-term starvation that lasted 3, 6, or
9 weeks with alternating periods of 3-week starvation and 1-week feeding showed partial compensation (Karlsen et al. 1995. In the most restrictive
treatment (9-week starvation over a period of
11 weeks), body mass at maturity was only about
60% of control sh. Although controls had signicantly higher fecundities than sh starved for
9 weeks, these dierences were related to body size
and relative fecundities were similar for females in
both groups. However, reduced growth and liver sizes
in feed-restricted groups did not result in lower proportions of maturing sh. In a similar experiment,
Jobling et al. (1994) observed full compensation in
male and female cod after 12 weeks of re-alimentation and reported an increased investment in reproduction in male cod previously subjected to a
low-energy diet.
Apart from an obvious trade-o between somatic
growth and reproduction, the interpretation of these
results is complex. Compensatory growth may use
energy and nutrients that otherwise would be available for reproduction, thus causing a delay in age at
maturity. If food restriction does result in a subsequent lower rate of maturation despite a compensatory restoration of lipid levels or body mass, this may
reect an eect of priority in allocation to growth
and storage. However, if there are body size or condition thresholds for maturation (Thorpe 1986), compensatory growth would favour an earlier
attainment of the required threshold and so a lower
age at maturity (Fig. 5). Experiments that compare
maturation rates of sh that have dierent reactions
to a period of growth depression would be useful in
dening the relationships between growth compensation and reproduction. A potential problem for
such experiments is having adequate controls
because of the diculty in obtaining compensating
and noncompensating individuals from a group of
animals that have shared the same growth history.
There are two ways by which sh can mitigate the
eects of growth depression on reproductive output.
Apparently, the simplest form is through growth
compensation. For example, female three-spined
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

sticklebacks experiencing regular or irregular intervals between feeding did not dier in mean date of
rst spawning and fecundity from females fed daily
(Ali andWootton1999); in this case, females compensated for days without food by showing hyperphagia
on days they were fed. The second scenario is more
puzzling and refers to a direct compensation of the
potential reduction in reproductive output associated with a smaller body size. An example is the
study by Siems and Sikes (1998) on the eects of predictable and unpredictable food availability on male
and female fathead minnows, Pimephales promelas
(Cyprinidae). Males that were switched from an
unpredictable to a predictable food supply at an age
of 75 days showed compensatory growth after the
switch. Females on the unpredictable food supply
were smaller than females on a predictable supply
(i.e. they had not exhibited full compensation), but
even so the former started spawning at the same age
and had the same mean production of eggs, which
indicates a direct compensation of the reproductive
output by increasing reproductive investment per
unit of somatic mass.
Compensatory growth of body components
Figure 5 Potential problems in analysing the
consequences of compensatory growth on reproductive
parameters. For clarity, we assumed linear growth and timeconstrained reproduction, and used the rate of maturation
as response variable. T1 andT2 dene the start and
termination of a period of growth depression, andT3
indicates the end of the phase of compensation. R and R0
represent alternative time-windows for maturation. (a)
Compensating (broken line) and noncompensating
(thick line) individuals have experienced dierent
conditions which could have direct or indirect (e.g. if
maturation rate is size dependent) eects on maturation
rate. The eect of the compensatory process on the rate of
maturation can only be evaluated if decision to mature
occurs afterT3 (case R0); (b) after a period of growth
depression, some individuals show compensatory growth
(broken line) whereas others do not compensate
(thick line) and the control growth rate (dotted line).
Comparison of compensating and noncompensating
individuals to detect direct inuence of compensatory rates
on the rate of maturation is justied at both R and
R0. If the decision to start maturation takes place at any time
betweenT2 andT3, allocation and size (only if size
dependence is assumed) eects can be evaluated. AfterT3,
dierences can be attributed to size eects or
impairment of the reproductive/hormonal function
associated with a previous episode of very fast
growth.
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

The pattern of growth compensation may vary


between body components. In rainbow trout, refeeding following a period of restriction was associated
with rapid restoration of the relative sizes of organs
(Weatherley and Gill1981,1987). Both short- (3 weeks)
and long-term (13 weeks) starvation entirely utilised
visceral fat. Long-term starvation also reduced gut
dry mass signicantly. Despite the inuence of starvation on the proportions of body components, during re-alimentation, rainbow trout attained normal
proportions or even over-compensation of heart,
liver, gonad, gut and visceral fat. Skin and carcase
dry masses were lower in controls and rainbow trout
previously exposed to short-term starvation than in
those previously subjected to long-term starvation.
In O. niloticus,Takagi (2001) reported compensatory
growth in acellular bone when sh were refed for
15 days after 15 days of starvation, but did not detect
compensatory growth in total body mass.
Proximate causes of compensatory growth
Several factors could contribute to the compensatory
growth observed during re-alimentation after a period of relative or total food deprivation, or other
causes of growth depression. These include a rate of
165

Compensatory growth in shes M Ali et al.

food consumption greater than that of continuously


fed controls (hyperphagia), enhanced growth eciency, reduced metabolic costs and reduced expenditure on locomotion. Most experimental evidence
suggests that growth compensation operates by
adjustments in food intake of growth-depressed animals. In part, this evidence is biased by the disproportionate number of studies investigating the eects of
growth depression on feeding rates compared to
other responses that might be involved in the compensatory process. However, while almost all studies
measuring feeding rates have found that hyperphagia makes an important contribution to the acceleration of growth, a substantial proportion of studies
investigating other variables have failed to show a
signicant contribution of these in growth compensation and, when detected, their contribution is
minor compared to the hyperphagic eect. The available evidence suggests that hyperphagia is the main
mechanism involved in the compensatory response,
although increased conversion eciencies or behavioural adjustments might play a role.
Hyperphagia
Hyperphagia is, by far, the most common mechanism
of growth compensation, and its occurrence is not
limited to the existence of previous episodes of starvation or food shortage (e.g. Nicieza and Metcalfe
1997). Hyperphagia is a rate of food consumption signicantly higher than that shown by sh that have
been continuously on an ad libitum ration. Because
the rate of consumption is allometrically size dependent (Wootton 1998), comparisons have to be made
for sh of the same size or, alternatively, these must
be based on size-corrected measurements. Hyperphagia may allow an animal to achieve the same or
even a higher cumulative food intake than an individual that has had continuous access to food, and thus
to achieve the same size. It may result from increased
meal frequency, increased size of meals or a combination of both (Russell and Wootton1993).
Responses to free access to food after a period
of growth depression
Several studies on sh noted that food consumption
is elevated following starvation or deprivation (Tyler
and Dunn 1976; Jobling 1982; Dobson and Holmes
1984). Studies on individually housed sh suggested
a hyperphagic response analagous to that observed
in mammals and birds (Miglavs and Jobling 1989a,b;
166

Figure 6 Hyperphagia shown by gibel carp in


re-alimentation phase following1or 2 weeks of total food
deprivation (symbols as in Fig. 2). Means for nine sh.

Russell and Wootton 1992; Hayward et al. 1997; Ali


and Wootton 1998a). Hyperphagic Arctic charr had a
food intake of approximately1.5^2.0% live body mass
day1, whereas in the same period, controls consumed approximately 1.0^1.4% body mass day1
(Miglavs and Jobling 1989a,b). A hyperphagic
response was demonstrated in European minnows
(Russell and Wootton 1992) and in gibel carp (Xie
et al. 2001; Fig. 6). In Atlantic salmon, hyperphagia
followed periods of food restriction (Bull et al. 1996)
or unseasonably low temperatures (Nicieza and
Metcalfe1997). A raised food intake was also reported
in groups of re-alimented juvenile Piaractus brachypomus L., a South American characoid (Koppe et al.
1993), and channel catsh (Kim and Lovell1995).
An exception is a complex experiment on grouped
rainbow trout, which failed to detect hyperphagia
during compensatory growth following 21 days of
deprivation (Boujard et al. 2000). This study suggested that, in rainbow trout, hyperphagia was a
response to food restriction whereas total deprivation generated an increase in growth eciency
during refeeding. The experiment included a predeprivation period of 34 days, during which separate
groups of trout experienced three dierent ration
levels factored with two dierent diets, one high
energy and one low energy.
There is evidence that sh can maximise their
intake rates during recovery after a period of
depressed growth. The main eect of the intensity
of the growth depression on the subsequent
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

hyperphagia is on the duration of the hyperphagic


phase, rather than a higher maximum rate of consumption (Russell and Wootton 1992; Ali and Wootton 2001; Xie et al. 2001; Zhu et al. 2001; Fig. 6). This
suggests that the maximum daily rates observed during hyperphagia represent a rate close to the maximum possible rate at which the gut can process food.
The exibility of feeding rates seems to be large
enough to allow for a rapid acceleration of growth
rates. For example, three-spined sticklebacks fed at a
mean interval of 3 days showed hyperphagia, consuming as much as 21% of their body mass on the
day when food became available (Ali and Wootton
1998a); this gure contrasts with a consumption rate
of 9^11% in sticklebacks fed ad libitum (Wootton et al.
1980). Similarly, the size-adjusted consumption of
hybrid sunsh (Lepomis cyanellus  L. macrochirus)
subjected to dierent schedules of food restriction
and re-alimentation was more than double that of
controls (Hayward et al. 1997). Although absorption
eciencies decrease with increasing feeding rates,
the acceleration of growth associated with an
increase in intake rates of 50^90% cannot be
achieved through other mecahnisms of compensation. The amount of energy that can be saved by reducing nonfeeding activity would account for a tiny
proportion of the total energy that can be allocated
to the process of growth compensation. Individual
ranges of variation in digestive eciency are relatively narrow (e.g. compared to passage time or feeding rate; Nicieza et al. 1994b), which suggests that
the scope for growth rate acceleration associated
with the maximisation of absorption eciencies
might be negligible in comparison with the potential
eects of increasing feeding rates.
Although the occurrence of hyperphagic
responses after a period of growth depression is a
generalised phenomenon, observations on the daily
rate of consumption during the period of compensatory growth reveal that there is considerable interspecic variation in the temporal pattern of the
hyperphagia (Hayward et al. 1997; Ali et al. 2001;
Hayward andYang 2001;Wu et al. 2002; Fig. 7). These
patterns suggest important interspecic dierences
in the regulation of the hyperphagic response, and

Figure 7 Inter-specic dierences in patterns of daily food


compensation during period of hyperphagia after1 week of
deprivation; , controls; &, deprived sh. (a) Three-spined
stickleback (means for six sh); (b) gibel carp (means for15
sh); (c) Chinese longsnout catsh (means for15 sh).
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

167

Compensatory growth in shes M Ali et al.

thereby in the process of growth self-regulation. The


analysis of interspecic variation in compensatory
patterns is important because it can improve our
understanding of the evolutionary signicance of
compensatory growth. At present, the scarcity of
comparative studies precludes a deeper insight.
Eects of food quality on hyperphagic response
Like other vertebrates, sh regulate, at least partly,
food intake to meet energy needs (Rozin and Mayer
1961; Lee and Putnam 1973; Page and Andrews 1973;
Cho et al. 1976; Grove et al. 1978; Lovell 1979; Marais
and Kissil1979; Fletcher1984). Lovell (1979) fed channel catsh, Ictalurus punctatus, at three digestible
energy levels (230, 290, and 350 kcal 100 g1) and
recorded elevated food intakes amongst catsh fed
the low-energy diet. Rozin and Mayer (1961) and
Grove et al. (1978) manipulated the energy contents
of feed byaddition of kaolin to the diet of goldsh, Carassius auratus (Cyprindae), and rainbow trout,
respectively. In both studies, increased intake of lowenergy food compensated for nutrient dilution.
Furthermore, goldsh presented with large then
small pellets alternately adjusted the number of pellets consumed to achieve an energy balance (Rozin
and Mayer 1961). Self-selection of macronutrients
(protein, fat and carbohydrate) by sh, including
goldsh and rainbow trout, has been demonstrated
by the use of demand-feeders (Sanchez-Vazquez et al.
1998, 1999). A study of the eect of periods of food
deprivation on nutrient self-selection by sea bass,
Dicentrachus labrax, imposed periods of 6 and 15 days
of total food deprivation separated by a period of
8 days of demand feeding (Aranda et al. 2001). After
both periods of deprivation, the bass showed hyperphagia but the deprivation had only minor eects on
diet selection; the intake of protein was slightly elevated for 2 days when feeding resumed. Therefore,
although the intensity of hyperphagia can be somewhat reduced if high-energy food is available, there
is no evidence of compensation mediated by behavioural decisions inuencing diet selection.
If the regulation of intake rates maintains steady
energy inputs, growth compensation should be independent of energy content of the diet. Atlantic cod,
sh fed with either high-energy or low-energy food
after a period of starvation showed full growth compensation by 12^18 weeks (Jobling et al. 1994). Cod
fed high-energy diets tended to have higher levels of
lipids in the liver at the end of the experiment (Jobling
et al. 1994). The interpretation of these results is
168

ambiguous. It could be an indication that the compensation of structural growth (sensu Broekhuizen
et al. 1994) takes priority over allocation to storage,
but may be a direct by-product of dierences in diet
composition. In rainbow trout, the energy content of
the diet supplied in both the pre- and postdeprivation
period had no eect on daily energy intake during
the compensatory growth phase (Boujard et al.
2000).
Physiological basis of hyperphagia: consumption,
absorption, and evacuation rates
Rate of food consumption in sh is related to the rate
at which the gut is lled and evacuated (Brett 1971;
Elliott1975a,b; Grove et al.1985; Singh and Srivastava
1985; Russell and Wootton 1993). Under some conditions of temperature and food availability, intake
rates will be limited by the maximal rate at which
food can be digested. An elevation of the maximal
rate (i.e. a reduction of the passage time of food
through the digestive tract) would contribute to a
greater hyperphagic response. Whether the rate of
gut evacuation can change directly in response to a
period of growth depression is not known. The diculty rests on disentangling the potential direct eect
and the decrease in passage time associated with
increased ingestion (Elliott 1972; Nicieza et al.
1994b). In the European minnows, there was no signicant dierence between the rates of foregut evacuation of the rst satiation meal fed after 4 or
16 days of starvation (Russell and Wootton 1993).
There was no evidence that the increase in food consumption that followed a deprivation period (Russell
and Wootton 1992) was associated with changes in
the rate of gut evacuation. A greater absorption eciency during compensatory growth (Russell 1991)
did not result from a dierence in the rate of processing of ingested food in the foregut, although there
was some evidence of a longer retention time in the
hindgut. Studies incorporating a short (48 h), preprandial starvation had no detectable eect on subsequent evacuation rates in Atlantic salmon (Talbot
et al. 1984). In brown trout, a starvation period of up
to 7 days prior to feeding did not aect evacuation
rates, although periods of 10 days or more did reduce
the rate (Elliott 1972). Similarly, a decreased rate of
evacuation occurred in Pleuronectes platessa (Pleuronectidae) after 30 days of starvation (Goddard1974).
In European minnows, the decline in gut contents
after a meal is correlated with return of appetite
(Russell and Wootton 1993), a correlation also found
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

in other teleosts (Fletcher 1984; Grove et al. 1985;


Singh and Srivastava 1985). Because of this relationship, a reduction in evacuation rates would be
expected to result in a corresponding delay in the
return of appetite. However, a period of food restriction usually evokes a subsequent compensatory rise
in consumption (Tyler and Dunn 1976; Godin 1981;
Jobling 1982; Miglavs and Jobling 1989b; Ali and
Wootton 1998). A more rapid return of appetite may
cause the hyperphagia following food deprivation in
O. kisutch (Dunbrack 1988). An increase in the rate
of gut evacuation and more rapid return of appetite
have been implicated in the increased consumption
of low-energy diets by sh (Grove et al.1978).
There is evidence for structural changes in the gut,
which increase its capacity, when sh are exposed to
intermittent feeding (Carter et al. 2001). Although
these long-term changes probably cannot account
for the immediate hyperphagia seen in compensatory growth, they may account for the longer term
changes in hyperphagia shown when sh are
exposed to cycles of feeding and deprivation (Ali and
Wootton 2001).
Physiological basis of compensatory growth:
metabolic rate
Metabolic rates of shes may decrease during food
restriction (Beamish 1964; Love 1970, 1980; Blaxter
and Ehrlich 1974; Jobling 1980; Johnston 1981; Du
Preez et al.1986a,b; Du Preez 1987;Wieser et al.1992).
Brown (1946, 1957) reported that when the quantity
of food given to brown trout was reduced towards
the maintenance level, the sh rst lost mass but
soon became acclimatised to the lower ration levels
and then gained mass. Similar results were obtained
by Hepher et al. (1983) for red tilapia, Oreochromis sp.
(Cichlidae), and suggest an adjustment in the metabolic rate. After an acclimatisation period, the metabolic rate of tilapia was regulated, as loss of body
mass was higher during the initial week of fasting
but mass declined more slowly in subsequent weeks.
In these studies, metabolic rate during the initial period of starvation was higher than during the subsequent period, indicating an acclimation response to
fasting. However, Wieser et al. (1992) suggested four
phases in response to deprivation and subsequent
refeeding: (i) stress: characterised by a state of hyperactivity (i.e. search for food); (ii) transition: with continued deprivation, reduction in routine rate of
respiration associated with reduction in locomotion
and decline in the activity of some major glycolytic
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Figure 8 Hypothetical responses of standard metabolic


rate (SMR) to starvation. (a) Metabolic rate decreases during
starvation towards a physiological stable minimum. (b) The
adjustment of SMR is not immediate, but there is a time-lag
between deprivation and SMR decline. (c) SMR initially
increases as a result of hyperactivity and then declines
towards the stable level.

and glycogenolytic enzymes in swimming muscles;


(iii) adaptation: stabilisation of whole body metabolic
rate around reduced level. If deprivation persists,
then there is increasing replacement of lipid by protein as the major metabolic fuel; (iv) recovery: rapid
increase in rates of oxygen consumption and growth,
and compensatory growth spurts correlate positively
with the length of the starvation period. Therefore, it
is crucial in studies on the eects of starvation on
standard metabolic rate to control sh activity; otherwise, it would be dicult to distinguish between a
reduction in metabolic rate associated with an acclimatisation eect and a re-adjustment after the initial
increase of activity during the stress phase (Wieser
1991; Fig. 8). Nevertheless, a metabolic adjustment to
conditions of nutritional stress has been clearly
demonstrated. In juvenile cyprinids, both routine
and standard rates of metabolism can decrease by
30^40% after 2 days of deprivation in relation to
more recently fed individuals (Wieser et al.1992).
If, during restriction, there is an acclimatisation
period before metabolic rate declines, there could
also be a comparable time lag upon refeeding before
metabolic rate reaches levels normal for feeding sh.
This would increase the scope for growth during the
initial stages of re-alimentation of sh. The monitoring of oxygen consumption during starvation and
refeeding of juveniles of three cyprinid species, Leuciscus cephalus, Chalcalburnus chalcoides mento and
Scardinius erythrophthalmus, revealed only brief
169

Compensatory growth in shes M Ali et al.

suppression of metabolic rate (compared to controls)


during the initial stages of re-alimentation, although
compensatory growth was observed (Wieser et al.
1992). In this study, mass-specic maintenance and
routine rates of metabolism rapidly increased to
levels normal for feeding sh, suggesting that their
suppression could not account for observed growth
compensation. Food restriction in juvenile carp also
led to biochemical changes that probably reduced
metabolic rate (Bastrop et al. 1991). Activities of
NADP-specic dehydrogenases (G6PDH, 6PGDH,
IDH, ME, G1DH) implicated in lipogenesis and in
amino acid metabolism were depressed after food
deprivation. Temporal priority of enzyme recovery
on re-alimentation was observed with the level of
enzymes recovering at dierent rates.
Energy expenditure during starvation of sh can
be reduced by decreased locomotor activity (Beamish
1964; Blaxter and Ehrlich 1974; Wieser et al. 1992).
Reduced activity during refeeding could contribute
to compensatory gain by increasing the proportion
of the energy available to growth. However, hyperphagia is usually associated with a higher level of
foraging activity; so, from a cost-benet perspective,
only a reduction of nonfeeding activity could be
regarded as a benecial eect. Although swimming
activity was not measured directly, Wieser et al.
(1992) found that both standard and routine metabolic (i.e. the metabolism of the sh performing normal, spontaneous movements in the absence of
other stimuli) rates increased upon refeeding, suggesting that activity also had increased. Juvenile
roach showed a reduction in swimming activity during 3 weeks of starvation, but activity returned to
control levels as soon as refeeding started (van Dijk
et al. 2002). In the sablesh (Anoplopoma mbria; Anaplopomatidae), compensatory growth seems to be
mediated by a shift in resource allocation, which
results in a reduced physiological capacity for forced
swimming (Sogard and Olla 2002). The compensatory ability of this species is reduced (Sogard and Olla
2002), which has been interpreted as a result of the
sh having high routine consumption rates that cannot be exceeded. In the same study, walleye pollocks
(Theragra chalcogramma; Gadidae) exhibited a strong
compensatory response by greatly increasing their
consumption rates compared to control sh (Sogard
and Olla 2002).
In ectotherms, standard metabolic rates are a measure of the resting metabolism and represent the
minimum cost of maintenance that a healthy, inactive, unstressed, unfed and nonreproductive indivi170

dual can incur. An important, yet unresolved,


question is how standard metabolic rates react during a phase of compensatory growth? This issue is
relevant because it can provide useful information
about the costs and risks associated with rapid
growth, and an understanding of the relationships
between resting metabolism, maximum rate of metabolism and metabolic scope (e.g. Priede 1985; Cutts
et al. 2002).
Growth eciency
Dobson and Holmes (1984) suggested that compensatory growth in sh could be attributed to an
increased eciency of food utilisation. Studies on
individually housed sh demonstrated improved
conversion eciency, based on growth in mass and
mass of food consumed on re-alimentation in Arctic
charr, European minnows and the three-spined
sticklebacks (Miglavs and Jobling 1989a,b; Russell
and Wootton 1992; Jobling et al.1994; Zhu et al. 2001).
Koppe et al. (1993) reported that food conversion eciency, and protein and energy utilisation of realimented sh were superior to controls fed at satiation levels in groups of Piractus brachypomus. In rainbow trout, one study found that compensatory
growth was entirely caused by an increase in growth
eciency, with no hyperphagic response (Boujard
et al.2000). In contrast, Hayward et al. (1997) detected
no overall dierences in growth eciencies between
controls and groups on cycles of starvation and
refeeding in Lepomis hybrids. Lepomis experiencing
cycles of 14 days of deprivation did have high growth
eciencies during the phases of compensatory
growth that occurred on refeeding. In channel catsh, there was no dierence between control and
starved sh during 18-week trials (Kim and Lovell
1995). Growth eciency did not dier between control and starved group after refeeding for 7 weeks in
rainbow trout (Speare and Arsenault1997) or in carp
that had been fed diet reduced in protein or energy
content during the deprivation period (Plank et al.
1984).
If growth eciency was higher during compensatory growth, this can be demonstrated by comparing
the growth of re-alimented sh with controls when
both groups are fed identical rations. Again, because
of the potential inverse correlation between consumption rate and conversion eciency (Kinne
1960; Pandian 1967; Solomon and Braeld 1972;
Elliott1976;Talbot 1985), evaluating the contribution
of changes in food conversion eciency to growth
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

compensation requires adequate control of variation


in intake rates.
Proximate composition, metabolites and
endocrines during food deprivation
and recovery
If compensatory growth is to be initiated on refeeding, changes in the physiological state of sh during
the deprivation period must signal a discrepancy
between the current state and the state of sh that
have not experienced that deprivation. It is unclear
which, if any, of the changes observed during deprivation provide the error signals for compensatory
growth when food becomes freely available.
In mammals, there are distinctive neural systems,
with their associated peptides, that regulate appetite
for fat, protein and carbohydrate (Hoebel1997).Appetite in sh is probably under comparable multifactorial control (Fletcher 1984; Le Bail and Boeuf 1997;
DePedro and Bjornsson 2001), although the details
probably dier in some respects from endotherms.
The control of appetite after a period of food deprivation has both short-term and longer term components (Carter et al. 2001). The short-term eects
partly reect the immediate eects of re-lling and
emptying the alimentary canal, including sensory
and hormonal feedbacks, together with short-term
changes in circulating nutrients. The longer term
component would reect the error signalled by physiological state. Changes in proximate composition
of the sh during deprivation and refeeding may oer
clues as to the control of appetite during the hyperphagic phase, although such changes have rarely
been studied in the context of compensatory growth.
Lipids
Fish, whose metabolism is largely based on lipids and
proteins (Jobling1994), store lipids in the liver, viscera
and muscle. The detailed distribution between these
body components does vary between species (Love
1970). Lipids are broken down early in starvation
(e.g. Anguilla anguilla (Anguillidae), Larsson and
Lewander (1973); Dicentrarchus labrax, Stirling
(1976); Esox lucius (Esocidae), Ince and Thorpe 1976;
rainbow trout, Jezierska et al. 1982; Oreochromis niloticus, Satoh et al. 1984). Lipids often constitute the
main energy source for maintenance of sh during
over-wintering fasts (Weatherley and Gill 1987; Bull
and Metcalfe 1997). Zamal and Ollevier (1995)
reported that the relative composition of fatty acids
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

was inuenced by starvation in juvenile African catsh, Clarias gariepinus (Clariidae).


Changes in lipid levels take place before and during
compensatory growth. Decreases in liver and viscera
as proportions of body mass occurred in Arctic charr
in response to food restriction for 8 weeks and starvation for 8, 12 and 16 weeks (Miglavs and Jobling
1989a). Starvation (for 16 weeks) resulted in signicant decreases in the lipid contents of the carcase
and viscera. Charr restricted for 8 weeks showed a
marked decrease in the proportion of body energy
present as lipid. The lipid energy:protein energy ratio
decreased from 1.16 in the initial sample to 0.80 after
8 weeks of restriction. Restoration of satiation feeding was followed by signicant increases in liver and
viscera as per cent of body mass. After 8 weeks of realimentation, there were higher percentages of lipid
and lower percentages of water in the liver, compared
to continuously fed controls. By the end of re-alimentation, whole body ratios of lipid energy:protein
energy had risen to1.33, similar to the1.30 of controls.
However, in terms of total body mass, only partial
compensation had been achieved.
Rainbow trout starved for 3 weeks showed a signicant change in carcase composition (Quinton
and Blake 1990); a reduction in fat and an increase
in moisture and protein concentration had occurred
compared to controls. No dierences were evident
after a subsequent re-alimentation period lasting for
3 weeks. In gibel carp, the ratio of fat to lean body
mass was signicantly reduced by 2 weeks of starvation, but restored within 2 weeks of satiation feeding
(Xie et al. 2001).
Golden perch, Macquaria ambigua (Percichthidae),
a carnivorous Australian freshwater sh, experiences marked seasonal hydrological changes that
aect food availability in its natural habitat. It
showed declines in hepatosomatic index (HIS) and
depletion of liver lipid, protein and glycogen over
60 days of starvation. Then, levels were maintained
over a longer period of starvation. Following depletion of lipid stores in liver, lipid contained in perivisceral adipose tissue was utilised along with epaxial
muscle glycogen. Muscle was not mobilised until
after depletion of hepatic and perivisceral reserves.
Hepatic energy reserves were rapidly restored on
refeeding. Levels of protein and lipid in the musculature did not change in response to food deprivation.
No signicant increase in water content of the muscle
of starved sh was observed over 210 days of the
experiment. The water content in the tissue of fed sh
did decrease as lipid accumulated. Liver was used as
171

Compensatory growth in shes M Ali et al.

an initial source of endogenous energy in response to


food deprivation (Collins and Anderson1995).
Periods of food deprivation imposed on slow-growing, juvenile Atlantic salmon in winter signicantly
reduced fat levels in comparison to control sh. During the refeeding period, the food-restricted groups
restored their fat levels to levels similar to those of
control sh (Metcalfe and Thorpe 1992; Bull et al.
1996; Bull and Metcalfe 1997). Both slow- and fastgrowing salmon exhibited a rapid, full re-adjustment
of their mass:length ratios, regardless of the degree
of compensation achieved for absolute mass or length
(Nicieza and Metcalfe 1997); such rapid re-adjustments seem to be related to changes in lipid levels.
When deprived, the salmon responded by exhibiting
a hyperphagic response (Metcalfe and Thorpe 1992;
Bull et al. 1996; Bull and Metcalfe 1997; Nicieza and
Metcalfe 1997). This protective response was at least
in part responsible for the restoration of body fat lost
during the food deprivation. The extent of the estimated lipid decit incurred during a period of food
restriction primarily aected the duration of the
hyperphagic response that occurred when food was
once again available, rather than intensity of feeding
(Bull and Metcalfe1997).
Arctic charr showed an inverse relationship
between food intake (% body mass day1) and carcase lipid (% wet mass) and visceral lipid (% wet
mass; Jobling et al. 1993). Koppe et al. (1993) reported
a correlation between food intake and visceral fat
contents in juvenile Piaractus brachypomus. A comparison between temporal changes in food intake
and body composition suggested that the level of food
intake in re-alimented sh decreased, as visceral fat
contents approached the level of controls fed at satiation level. Peter (1979) indicated the possibility of a
similar mechanism in other species of sh (see also
Machiels and van Dam 1987). In channel catsh, sh
were manipulated by diet composition to reach dierent levels of fat content and then fed the same diet.
In the rst 2 weeks following a switch to the same
diet, the low-fat catsh ate more than the high-fat sh
(Silverstein and Plisetskaya 2000). These results are
in accord with the expectations of a theory of appetite
regulation based on maintenance of energy reserves
within narrowly dened set point limits (Jobling and
Johansen 1999). One result not compatible with the
model was the observation that, in rainbow trout
subjected to 1, 11 or 21 days of deprivation, growth
and feed intake in a refeeding period of 10 days were
not related to lipid levels at the end of deprivation
(Boujard et al. 2000). This study was also unusual in
172

that the growth compensation in the refeeding period was entirelycaused byan improvement in growth
eciency and not hyperphagia.
In mammals, the brain neuropeptides galanin (stimulatory) and leptin (inhibitory) have been implicated in the control of an appetite for lipid (Hoebel
1997). Their roles in teleosts are still to be claried
(Hoebel 1997; Le Bail and Boeuf 1997; DePedro
and Bjornsson 2001), although Silverstein and
Plisetskaya (2000) detected no eect of injection of a
fragment of rat leptin into the brain of channel catsh
on appetite.
Protein and RNA:DNA ratios
In sh, muscle RNA concentrations, because of the
role of RNA in protein synthesis, reect dierent
somatic growth rates induced by dierent levels of
feedings (Buckley 1979; Westerman and Holt 1988;
Bastrop et al.1991). In juvenile Atlantic salmon, RNA
concentrations and the RNA:DNA ratio responded
to three levels of reduced ration representing 0, 20
and 50% of the control level imposed for a maximum
of 8 days (Arndt et al. 1996). By day 4 of restriction,
the starved group had signicantly lower RNA concentrations than control sh. By day 8, all four ration
levels were reected in RNA concentrations. For salmon starved or fed a 20% ration for 4 days, the RNA
levels recovered close to control levels after 4 days of
refeeding, although for the starved sh, the
RNA:DNA ratio was a better indicator of refeeding.
Ornthine decarboxylase (ODC, the enzyme that plays
a rate-limiting role in polyamine synthesis, which
reects protein synthesis) activity also dropped sharply during food restriction. ODC activity increased
rapidly after feeding was returned to the level of the
control group and showed some evidence of overcompensation in the refeeding period (Arndt et al.
1996). Unfortunately, the length of this promising
experiment was too short for compensatory growth
to be clearly displayed. In Arctic charr subject to
regimes of food restriction and re-alimentation,
RNA:DNA ratios in muscle and liver correlated with
the specic growth rate (Miglavs and Jobling 1989b).
However, the growth rates during the compensatory
growth phase were higher than those that would have
been predicted from the RNA:DNA ratios observed.
In carp, food deprivation reduced cell size of liver
(estimated indirectly), but not of white muscle. This
reduction was associated with a decrease in soluble
and total protein concentrations (Bastrop et al.1991).
Starved carp had lower white muscle RNA and liver
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

RNA concentration than fed sh, indicating reduced


protein synthesis under food restriction. (Bastrop
et al.1991). Mendez and Wieser (1993) subjected juvenile roach, 280^460 mg in fresh mass, to 7-,14-, 22-,
29-, 36- and 49-day starvation regimes at 20 8C to
study the eects of deprivation and refeeding on
metabolic events. No eect on protein g1 dry mass
was observed. Enzyme activities were unaected in
the 7-day group. In the14-day group, enzyme activity
dropped during the rst week but remained constant
thereafter. In the 22-day group, enzyme activity
declined continuously over starvation. There was a
slow recovery of enzyme activity upon refeeding. In
the 29-day group, activity of one enzyme (glutamatepyruvate transaminase) increasedduring deprivation.
Activityof this enzyme declined on refeeding.
In mammals, growth hormone releasing factor
(GHRF) plays a role in regulating the appetite for protein (Hoebel1997). Clearly, it would be of value to estimate the levels of the equivalent hormone in teleost
during the periods of deprivation and compensatory
growth (Le Bail and Boeuf1997). A role for neuropeptideY (NPY) in stimulating appetite during the compensatory growth phase is suggested by the
observation that injection of NPY into the third ventricle of the brain of channel catsh stimulated appetite (Silverstein and Plisetskaya 2000; DePedro and
Bjornsson 2001).
Carbohydrates
Carbohydrates play onlya minor role as storage materials in sh (Mommsen1998). Carp had similar levels
of blood glucose and liver glycogen when starved or
fed for 22 days (Nagai and Ikeda 1971). When carp
were subjected to 2 months of fasting and 12 days of
refeeding, the rst phase (until day 8 of fasting) was
characterised by a reduction in the hepatosomatic
index mainly because of glycogen mobilisation. A
transitory increase in plasma glucose and lactate
suggested an initial increase in energy demand. No
changes were produced in the percentage of glycogen
and protein in muscles, but the musclosomatic index
and the total body muscle protein decreased. Stabilisation of liver glycogen content, and plasma glucose
and lactate levels decreased muscle protein levels,
and a reduction in the rate of body mass loss characterised the second phase (from day 8 of fasting). Total
protein content in whole muscle decreased by 22%,
similar to the rst phase. Twelve days of refeeding
produced recovery of plasma glucose levels. Liver glycogen completely recovered. In contrast, muscloso# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

matic index, and protein and lipid contents indicated


that the muscle did not completely recover from
2 months of fasting, although an overshoot of muscle
glycogen was observed (Blasco et al.1992a,b). In goldsh (Chavin andYoung1970) and Opsanus tau: Batrachoididae (Tashima and Cahill 1968), blood glucose
levels were also maintained for long periods during
food restriction.
The role of glucose in regulating appetite in teleosts is unclear. In comparison to mammals, the sh
have a poor ability to regulate plasma glucose
(Mommsen1998). Carter et al. (2001) suggest that this
makes it unlikely that glucose plays an important role
in the regulation of food intake in sh. The lowest
blood glucose levels in Rooseveltiella nattererri (Characidae) preceded the period of greatest food intake
(Bellamy 1968). In contrast, Katsuwanus (Scombridae) continued to feed when blood glucose levels
were rising (Magnusson 1969). However, one study
has implicated plasma glucose levels in compensatory growth. In rainbow trout deprived of food for 1,
11 or 21 days, the growth eciency during a 10-day
re-alimentation period was inversely related to
plasma glucose at the start of the re-alimentation
period (Boujard et al. 2000).
Water content
Starvation results in tissue hydration (Jobling 1980;
Weatherley and Gill 1987; Miglavs and Jobling 1989a;
Quinton and Blake1990). An increase in tissue hydration plays a role in the limitation of the loss or even
the maintenance of wet body mass during starvation
(Love 1970). In juvenile roach, Rutilus rutilus L., it
took 36 days of deprivation before tissue water content was higher than that of controls (Mendez and
Wieser 1993). Whether the level of tissue hydration
plays any direct role in appetite regulation in teleosts
is not known. In sh with swim bladders, changes in
the specic mass of the body would require adjustments in the volume of the bladder to attain neutral
buoyancy, which might signal changes in body composition.
Endocrines
Growth is under endocrine control (Mommsen1998);
however, the role of hormones in compensatory
growth is little understood. Temporal changes in
growth, plasma thyroid hormones (triiodothyronine,
T3, and tetraiodothyronine,T4), cortisol, growth
hormone (GH) and nonesteried fatty acid (NEFA)
173

Compensatory growth in shes M Ali et al.

concentrations, hepatic T3 content and hepatic 50 monodeiodinase activity were measured in rainbow
trout subjected to a sustained fast for up to 8 weeks,
and during an 8-week refeeding period. Dierences
in growth rate between fed and fasted groups were
evident after 2 weeks, but signicant mass loss by
fasted groups was not evident until between 4 and
6 weeks into the fast. Liver protein content was
depressed inthe fasted shwithin 2 days of food deprivation. Interrenal activity suggested a depressed pituitary^interrenal axis in fasted animals.Within1 day of
refeeding, plasma cortisol levels had increased to
levels characteristic of controls. The status of thyroid
hormones indicated by plasma thyroid hormone level,
liver T3 content, hepatic 50 -monodeiodinase (MD)
activity and thyroid epithelial cell height also indicated a depressed pituitary^thyroid axis in fasted
animals, with recovery to levels of the fed animals
within 1 week. Despite compensatory changes in
accumulation of reserves (as indicated bya compensatory increase in hepatosomatic index), there were no
apparent compensatory changes in any of the endocrine variables evident during the refeeding period
(Farbridge and Leatherland 1992a,b; Farbridge et al.
1992; Leatherland and Farbridge1992).
In channel catsh, Gaylord et al. (2001) recorded a
decline in plasma T3 and T4 with food deprivation
for as few as 3 days, with a subsequent rapid recovery
on refeeding. However, in this experiment, there was
no evidence of hyperphagia during the refeeding
period.
In carp fasted for 2 months followed by 12 days of
refeeding, plasma insulin levels decreased twofold
and plasma glucagon threefold over the rst 8 days
of fasting and remained low in the second phase
(from day 8 onwards of fasting).Twelve days of refeeding produced a higher daily growth rate than in controls and a recovery of plasma insulin and glucagon
levels (Blasco et al.1992a,b).
In salmonids, exogenous treatment with GH can
induce increased appetite and growth rates in excess
of those normally observed (McLean et al. 1992; Le
Bail et al. 1993; Johnsson and Bjornsson 1994;
Johnsson et al. 1996, 1999). This mimics compensatory growth. The eect of GH on growth rates is at
least partly mediated through the production of insulin-like growth factor (IGF) by the liver as a response
to circulating GH. This response depends on the presence of growth hormone receptors (GHR) in the liver
cells. At least in some species, low food levels inhibit
the stimulation of IGF by GH, while levels of GH in
the blood plama increase (Perez-Sanchez and Le Bail
174

1999). This suggests that compensatory changes in


appetite and growth may reect temporary, but unusually high, levels of stimulation of the GHRF^GH^
IGF axis. Once the time trajectories of the compensatory growth response have been adequately dened,
the concomitant changes in levels of GHRF, GH,
GHR and IGF can also be dened. However, the decits that evoke compensatory changes in this endocrine axis for growth control are obscure.
Modelling compensatory growth
There are numerous models of sh growth in the literature (Wootton 1998). These range from complex
bioenergetic models, which partition the total metabolic expenditure into many terms, corresponding
to expenditures associated with locomotion, tissue
synthesis, tissue maintenance and digestion (e.g.
Kitchell et al. 1977; From and Rasmussen 1984), to
the simple von Bertalany (1960) model. However,
testing such models often reveals signicant discrepancies between predicted and observed growth
rates made at low or uctuating rations (Cui and
Wootton 1989; Ruohonen and Makinen 1992).
Whitledge et al. (1998) found that a detailed bioenergetics model was not a good predictor of compensatory growth or consumption in Lepomis hybrids,
although the model performed well for control sh
fed on a daily basis. In contrast, a simple, empirical
regression model provided good predictions of the
growth of three-spined sticklebacks experiencing
cycles of feeding and deprivation (Ali and Wootton
2001).
Most of the models, including models that describe
energy ows into and out of the sh in detail, treat
all internally stored energy as interchangeable, with
the instantaneous state of sh characterised by a single variable, normally total dry mass or energy content. A xed mass^length relationship is usually
assumed. However, this assumption of a constant
mass^length relationship may not be valid if changes
in food availability over days or weeks cause an
alteration in the nutritional status of the sh. The
assumption that two conspecics of equal mass, one
short and fat and the other long and thin, will grow
at identical rates may also be invalid (Broekhuizen
et al.1994).
Structural and storage model
Broekhuizen et al. (1994) formulated a simple, physiologically based model to investigate the phenomenon
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

of compensatory growth. The model is based upon


two key assumptions. The rst is that an individual
sh partitions net assimilate between two tissue
types: those which can and those which cannot be
remobilised once laid down. The second is that the
individual modulates its behaviour and physiology
in response to the instantaneous ratio of mobilisable
and nonmobilisable tissues. The model indicated that
the temporal pattern of growth trajectories observed
during recovery need not indicate the presence of a
complex memory of environmental conditions. The
trajectories can be produced by an instantaneous
response to a simple measure of nutritional wellbeing, the ratio of the mass of mobilisable reserve tissue to that of structural tissue. The model assumes
that, as this ratio falls, the rst response is that of
hunger or increased appetite. When reserves have
fallen far enough to indicate a signicant risk of starvation, the animal reduces its basal metabolic rate,
thus lowering the risk, but at the cost of a decrease
in foraging potential. This model was developed specically to address the question of how compensatory growth might be regulated, and the other
details were not incorporated. Although the model
illustrates some general principles and a subset of
the specic responses in one species group (salmonids), it is not a complete model of sh growth and tissue allocation pattern.
A related model is the lipostat of Jobling and Johansen (1999). This argues that appetite is regulated in
relation to lipid levels, so hyperphagia and the consequent compensatory growth cease as lipid levels
reach the optimal levels. In a study of grouped postsmolt Atlantic salmon, sh that had been deprived of
food showed hyperphagia and higher growth rates.
However, their compensatory response abated once
the deprived salmon had attained the same body
composition as that of the control sh, although the
deprived sh were smaller than controls and so had
only shown partial compensation in mass (Johansen
et al. 2001). Further evidence compatible with a lipostat model of appetite control was obtained by manipulating the lipid content of juvenile Atlantic salmon
by providing high rations of either high-fat or low-fat
feed (Johansen et al. 2002). This avoided the use of a
period of food deprivation to reduce lipid levels, which
confounds the eects of deprivation per se with the
eects of body lipid levels. It is unclear whether a lipostat model is relevant when growth depression is
induced by unseasonably low temperatures, rather
than food deprivation (Nicieza and Metcalfe 1997;
Maclean and Metcalfe 2001).
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Nicieza and Metcalfe (1997) found that compensatory growth continued after the deprived sh had
attained the same length^mass relationship as control sh. This suggests that compensatory growth is
more than just the readjustment of the ratio of
reserve to structural tissue. In cod, the early phase
of compensatory growth was characterised by
growth of swimming muscle rather than more
obviously reserve tissue such as liver or visceral fat
(Jobling et al. 1994). In the three-spined sticklebacks,
preliminary estimates of the characteristics of compensatory growth did not suggest that the growth of
reserve material (lipids) had precedence over protein
growth (Zhu et al. 2001). The time resolution of the
studies on cod and sticklebacks may have been too
coarse to detect short-term shifts in the rate of
growth of reserve and structural material. Strikingly,
when compensatory growth in rainbow trout was
entirely a consequence of increased growth eciencyand not hyperphagia, no eect of lipid reserves
on appetite was detected (Boujard et al. 2000).
Model based on specic growth rate
Hubbell (1971) developed a model of growth regulation in animals based on control theory. It assumed
that individuals had an optimal growth trajectory. If
the growth rate falls so that the individual grows
slower than the optimal trajectory, compensatory
changes are induced that reduce the error between
the optimum and achieved growth rate. The model
does not seek to identify the mechanisms by which
the control is achieved, but it does provide a theoretical framework within which such mechanisms could
be sought. This model was used to interpret the compensatory growth patterns of the European minnows
(Russell and Wootton1992).
The multifactorial control of appetite (Hoebel1997;
Le Bail and Boeuf 1997; DePedro and Bjornsson
2001) suggests that the control of compensatory
growth may be equally complex and may depend on
the eect that the period of deprivation has had on
body composition. There may be dierent trajectories
of recovery for the carbohydrate, lipid and protein
components of the body. The roles of hyperphagia
and increased growth eciency may also depend on
whether the preceding food deprivation was total or
partial (Boujard et al. 2000). If so, this complexity
needs to be reected in the models. There is a need to
integrate the general control theory framework
developed by Hubbell (1971) with models of the control of appetite such as provided by DePedro and
175

Compensatory growth in shes M Ali et al.

Bjornsson (2001) for the eects of neuropeptides and


hormones and Carter et al. (2001) for short-, mediumand long-term physiological control.
Functional significance of compensatory
growth
Sibly and Calow (1986) argued that, on theoretical
grounds, compensatory growth is likely to be advantageous only under certain environmental conditions. These include seasonality associated with
changes in the suitability of the environment and
conditions in which individuals of reduced size
experienced higher risks.
Size-dependent mortality
Mortality rates in aquatic systems are inversely
related to size (McGurk 1986; Sogard 1997). In sh
populations, mortality over some seasons can be
inversely related to size, partly because smaller individuals accumulate relatively smaller energy
reserves (Schultz and Conover1999). Menidia menidia
display size-dependent over-wintering mortality
(Conover 1992). Over-wintering mortality of youngof-the-year yellow perch and white perch, Morone
americana (Moronidae), predominantly aects the
smallest individuals (Post and Evans 1989; Johnson
and Evans 1990, 1991). This pattern appears to be
common among teleosts (Hunt 1969; Oliver et al.
1979; Toneys and Coble 1979; Adams et al. 1982; Conover and Ross1982; Henderson et al.1988). Often susceptibility to predators is also higher for the smaller
sh, even when the larger ones are more prone to risk
exposure to predators (Johnsson 1993; but see Litvak
and Leggett 1992). In brook trout, Salvelinus fontinalis, postreproductive, over-wintering survival of
females increased with body size (Hutchings 1994).
When sh with a small body size do suer higher
rates of mortality, there is clearly an advantage if
compensatory growth can allow some recovery of
body size after a period of growth depression (Letcher
et al. 1996). In a eld experiment with juvenile Atlantic salmon, the juveniles were stocked at dierent
sizes in late spring, but no dierences in size were
detected in the following autumn (Letcher and
Terrick 2001). Estimates of the survival rates suggested that the result was not a consequence of
size-dependent mortality over the summer. Letcher
and Terrick (2001) proposed that the convergence
of size by autumn was a result of compensatory
growth.
176

Juvenile Atlantic salmon rely heavily on lipid


reserves to avoid starvation during their freshwater
phase, drawing upon reserves during winter
(Egglishaw and Shackley 1977; Gardiner and Geddes
1980) when food supply is inadequate and unpredictable. During this time, fat stores and appetite are
regulated in relationship to body condition. Sizedependent mortality suggests a function for compensatory growth at this stage (Metcalfe and Thorpe
1992; Bull et al. 1996). In juvenile Atlantic salmon on
a slow-growing trajectory, an accelerated depletion
of fat reserves in early winter led to a hyperphagic
response, which replenished losses (Metcalfe and
Thorpe 1992). Appetite fell to a low level once lipid
levels had been restored. This is an example of a compensatory response that seems to relate solely to the
storage component of the body.
Size-dependent prey selection
Fish are gape-limited predators, and feeding habits
are strongly inuenced by sh size (Dunbrack and
Dill 1983; Macchiusi and Baker 1991). As they grow,
the increase in size of the mouth allows them to take
larger size prey. This often leads to large ontogenetic
changes in diet (Gerking1994;Wootton1998). A shift
to a larger and more protable prey may result in an
acceleration in growth, for example when sh switch
from a diet of invertebrates to one of sh (e.g. Gorman
and Nielson 1982; Buckel et al.1998). A phase of compensatory growth may allow individuals to reach
the size at which the ontogentic diet switch can be
made. A failure to make that switch may result in a
low body size throughout life.
Size-dependent ontogenetic changes
During ontogeny, there may be critical transitions
that depend on an individual reaching a threshold
size. Compensatory growth would allow an individual to reach that threshold size, despite experiencing a prior period of growth depression. Juvenile,
anadromous salmonids undergo smoltication as
the precursor to their migration to the sea. In some
species, there is a narrow time window during spring
when this process occurs. Marine survival is maximised by the juveniles reaching a critical minimum
size prior to seaward migration (e.g. Parker 1971;
Bilton et al. 1982; Healey 1982; Henderson and Cass
1991; Lundqvist et al.1994). The capacity for compensatory growth shown by juvenile salmon can
increase the chance of an individual achieving
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

that critical minimum at an early age (Nicieza and


Metcalfe 1997). In Atlantic salmon, juveniles show
dierent growth trajectories that result in dierences
in age at smoltication. The faster growing group
showed a stronger compensatory response than the
slower growing group. Nicieza and Metcalfe (1997)
suggest that growth, including any compensatory
growth, of the faster-growing group would be targeted at reaching the critical size for smoltication,
whereas that of the slower-growing group would be
targeted at a level of storage.
Size-dependent reproductive success
Fecundity in female teleosts is a function of body size
(Wootton 1998). A disadvantage of small size is an
absolute lower fecundity. In rst-spawning cod,
9 weeks of starvation was not fully compensated for
in the subsequent period of refeeding. On spawning,
the smaller females had lower absolute fecundities,
but fecundity per unit of body mass did not reect
nutritional history (Karlsen et al.1995). Thus, fecundity showed neither over- nor under-compensation.
The eects of timing of periods of food deprivation
and any subsequent period of growth compensation
on fecundity are yet to be adequately dened, despite
their potential importance in determining the egg
production by the spawning stock of a population.
Although the relationship between fecundity and
size seems general in teleosts, in some species, large
size also confers a breeding advantage on males. This
advantage may accrue when the mating system
involves sperm competition or territorial defence
(Parker 1992). In such species, growth compensation
would be advantageous to males.
Costs of compensatory growth
Compensatory growth with its associated changes in
appetite presents an evolutionary puzzle. The advantages of large body size suggest that individual sh
should grow at the maximum rate physiologically
possible to obtain those advantages (Sibly and Calow
1986; Arendt and Wilson 1997; Arendt 1997). However, rates during compensatory growth exceed
those shown by continuously fed sh, indicating that
the latter are consuming food and growing at rates
less than the rate that is maximally possible. This
suggests that there are costs associated with compensatory growth that must be traded o against
its benets (Arendt 1997; Metcalfe and Monaghan
2001).
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

The nature of compensatory growth may dier


from long-term growth and be less metabolically
expensive. This would allow a transient period of
unusually high growth rates. A careful analysis of
the nature of the material laid down during compensatory growth, in association with a detailed energy
budget, would identify this type of growth. However,
neither juvenile three-spined sticklebacks nor rainbow trout found showed unusual patterns of lipid or
protein deposition during compensatory growth
(Boujard et al. 2000; Zhu et al. 2001).
Priede (1985) has used an analogy with machines
to argue that organisms operating at maximum rates
pay a cost in high mortalities. This suggests that
long-term growth rates represent an optimal solution
to the trade-o between growth rate and mortality
that maximises tness (Sibly and Calow 1986). More
rapid growth entails physiological adjustments that
would increase mortality under adverse conditions.
For example, sh growing faster would be expected
to have higher oxygen consumption (Brown and
Cameron 1991; Houlihan et al. 1993; Jobling 1993)
and thereby a higher risk of mortality when oxygen
concentration is lowered. There are no reports of
increased mortality rates in groups of sh undergoing compensatory growth. Curiously, a study on
channel catsh found that groups of sh fed intermittently and showing growth compensation were less
susceptible to infection by the pathogenic bacteria
Edwardsiella ictaluri than continuously fed controls
(Chatakondi and Yant 2001).
High rates of growth may be associated with a
greater risk of developmental abnormalities (Leamy
and Atchley 1985; Sibly and Calow 1986; Arendt
1997). This potential cost of compensatory growth is
likely to be most obvious in the early life-history
stages when both growth and developmental rates
are high. In a comparison of two populations of
pumpkinseed sunsh, Lepomis gibbosus (Centrachidae), sh from the population with the faster posthatching growth showed a delay in the start of the
ossication of the cranial bones (Arendt and Wilson
2000). Comparisons both within and between populations of pumpkinseed sunsh found an inverse
relationship between the strength of body scales and
rate of growth (Arendt et al. 2001). An experimental
study on immature three-spined sticklebacks in
which growth rate was manipulated by dierences
in ration and temperature assessed developmental
stability by estimating the uctuating asymmetry of
n and skeletal traits (Robinson and Wardop 2002).
Sticklebacks with the higher growth rate had higher
177

Compensatory growth in shes M Ali et al.

levels of uctuating asymmetry, suggesting a developmental cost of rapid growth. In Atlantic salmon,
the prevalence and severity of coronary arteriosclerosis has been associated with high growth rates,
especially those associated with the marine phase of
the salmon life cycle (Saunders et al. 1992). None of
these studies explicitly analysed the costs of compensatory growth.
Conditions leading to rapid growth can inuence
muscle cellularity and development, and consequently function. Christiansen et al. (1992) noted that
rapidly growing Arctic charr sometimes showed a
distinctive pattern of damage in their white swimming muscle. Unusually fast-growing coho salmon,
which had been genetically manipulated by the
insertion of copies of a gene for growth hormone,
showed a trade-o between growth and swimming
performance (Farrell et al. 1997). In Menidia menidia,
interpopulation comparisons and manipulations of
growth rates within populations indicated that there
was a trade-o between growth rate and swimming
performance (Billerbeck et al. 2001): faster growing
silversides had a poorer prolonged and burst swimming performance. A cost in locomotion performance has also been described in the sablesh,
whose reduced compensatory ability appears to be
mediated bya diversion of energy otherwise allocated
to forced swimming (Sogard and Olla 2002).
One study has explicitly linked compensatory
growth with delayed but measurable growth costs
(Morgan and Metcalfe 2001). Grouped juvenile
Atlantic salmon that had shown full compensation
in autumn after a period of food deprivation showed
signicantly lower growth rates and lipid reserves
the following spring than continuously fed controls,
although food was freely available. The study also
suggested a cost to future reproduction. The male salmon that had shown compensatory growth also had
lower subsequent rates of sexual maturation in the
following autumn. However, female three-spined
sticklebacks exposed to alternating short-term periods of deprivation and refeeding in the month before
the breeding season started spawning at the same
time, at the same size and had the same fecundity
at rst spawning as females fed ad libtum (Ali and
Wootton1999).
The hyperphagia usually associated with compensatory growth may also impose behavioural costs.
Increased foraging may expose the animal to a
greater risk of predation. Foraging coho salmon juveniles showing compensatory growth were bolder in
the presence of a potential predator than control sh
178

(DamsgQrd and Dill 1998). Rainbow trout exogenously treated with GH, which induced an increased
appetite, showed increased boldness in the presence
of a model heron (Jonsson et al. 1996), as did transgenic Atlantic salmon with enhanced growth rates
(Abrahams and Sutterlin 1999). Fast-growing Atlantic silversides were more vulnerable to predation by
piscivorous shes (Lankford et al. 2001). Hyperphagia
may also induce higher levels of intraspecifc aggression as individuals compete for food, with associated
costs in terms of higher rates of energy expenditure
and increased risk of mortality. Food-deprived Atlantic salmon juveniles did show increased levels of
aggression during their compensatory growth phase
(Nicieza and Metcalfe 1997). Fast-growing juvenile
Atlantic salmon were more aggressive than slowgrowing sh, but they were also more vulnerable to
being attacked (Nicieza and Metcalfe1999).
Compensatory growth as a by-product of
flexible growth and dynamic trade-offs
Compensatory growth can be interpreted as a consequence of a dynamic process of optimising the allocation of resources between structural, storage and
gonadal growth, maintenance and energy-consuming activities such as foraging and predator avoidance (Sibly and Calow 1986; Nicieza and Metcalfe
1997). The extent of the development of a compensatory growth response should reect the balance
between the costs associated with increased growth
rates and the benets of regaining a growth trajectory and an appropriate balance between structural
and reserve materials.
Compensatory growth shows a diversity of forms.
In some conditions, it simply restores lost reserves,
or it may restore the growth trajectory in length or
mass, or it may consist of a combination of reserve
and somatic growth. Several hypotheses address this
diversity (Metcalfe et al. 2002). First, the model of
Broekhuizen et al. (1994) is a nutritional state hypothesis, predicting a consistent compensatory response
to a given decit in the ratio of storage to structural
materials. Second, there may be physiological constraints on the compensatory response that can be
mounted, so dierent responses occur depending on
the environmental circumstances. An example
would be low temperatures that depress appetite to
such an extent that a compensatory growth response
cannot be mounted. A third hypothesis is that compensatory responses are adaptive and vary facultatively according to ontogenetic stage, nutritional
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

state of the individual and time of the year. Metcalfe


et al. (2002) interpret their study of seasonal dierences in the compensatory response of juvenile
Atlantic salmon as a possible example of this facultative response hypothesis.
Compensatory growth is often treated as a discrete
process, isolated from the theoretical framework
developed to explain the exibility of growth trajectories and how these are aected by a trade-o
between growth and mortality rates (e.g. Ludwig
and Rowe 1990; Houston et al. 1993; Werner and
Anholt 1993; Abrams et al. 1996; Houston and
McNamara 1999). While this narrow approach has
proved useful in identifying generalised characteristics of thegrowthtrajectories (e.g. submaximalgrowth
rates), a wider view is needed to understand the
evolutionary mechanisms that gave rise to and maintain growth compensation in natural populations.
It is frequently assumed that growth rates are
directly determined (as opposed to just inuenced)
by environmental factors including temperature, risk
of predation or food availability (Ro 1992). Compensatory growth provides evidence that growth resilience persists without the direct inuence of
environmental conditions. There is not a single
growth rate associated with a given temperature,
meal size, risk or time, but the rate also depends on
individual state (e.g. size, energy reserves). The balance of costs and benets of rapid growth can vary
with individual size. Individuals would benet by
being able to vary growth rates with some independence from their environment (Case 1978; Sibly and
Calow 1986; Metcalfe et al. 2002). This suggests that
compensatory growth itself might be adaptive.
Compensatory growth might just reect the existence of trade-os betweengrowthand mortalityrates
and the state-dependent exibility of growth. Some
theoretical studies have stressed the importance of
recognising endogenously controlled variation in
growth rates (Abrams et al. 1996), which adjusts
growth trajectories to modications in existing conditions (e.g. photoperiod, predation risk, competition).
The type of endogenous control promoting compensatory growth as discussed here is much more subtle. It
has not been explicitly considered in these theoretical studies, despite their incorporation of time constraints, which can be inherently associated with
individual state (for example, when certain sizes or
energylevels must be achieved bycertaintimes).
Some investigators have made a distinction
between nonreal and real growth compensation
(see Zivkov 1982, 1996; Zivkov et al. 1999). The rst
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

type occurs when initially smaller individuals do


eventually reach a certain length by showing, but
not exceeding, the growth rates of initially larger
individuals, whereas true compensation refers to a
higher growth rate. Only one study on sh presented
a comprehensive analysis of the problems in the
determination, essence and reasons of the growth
compensation (Zivkov 1982). He suggested that
growth compensation must not be regarded as something dierent from normal growth and noted that
understanding the nature of compensatory growth
requires understanding the more general process of
growth regulation. His view that growth compensation and growth depensation constitute the two ends
(convergence and divergence of individual growth
trajectories) of the process of growth regulation has
not been discussed, although these two processes
are commonly observed in shes.This view implicitly
considers that selection might be acting not on
growth compensation itself but on the continuous
adjustment of growth rates. A wider view of growth
compensation goes beyond the partial compensation^full compensation dichotomy by considering
negative compensation. Fish often grow at a lower
rate than that imposed by direct, existing environmental conditions (chiey, temperature and food
availability), which provides evidence of growth costs
and suggests that there is an optimal growth trajectory. Positive deviations from this growth curve
would be penalised, and under some circumstances,
growth rates lower than thenormal, i.e. submaximal
growth rates, should be expected. To our knowledge,
no studies have determined yet whether or not this
negative compensation can occur. Experimental
research in this line is currently being developed
(A.G. Nicieza and D. Alvarez, unpublished data).
Future directions for research
The study of compensatory growth as dened in this
review has been almost entirely based on laboratory
experiments. Little is known on the incidence, consequences and importance of compensatory growth in
natural populations. Only two studies have considered geographical variation in the compensatory
growth response (Purchase and Brown 2001; Schultz
et al. 2002), and these lead to dierent conclusions.
Populations do show compensatory responses in
growth rates, but these do not necessarily depend on
the form of growth compensation discussed here. In
many cases, they may simply represent a response to
improved food availability, and not involve rates
179

Compensatory growth in shes M Ali et al.

greater the usual long-term rates. Most of the


reported information is based on indirect methods
like the back-calculation of growth from scales or
otoliths, and there is a bias towards species relevant
to commercial sheries (Beacham 1981; Moores and
Winters 1982; Nicieza and Brana 1993a,b; Haweet
and Ozawa 1996). This does not provide conclusive
evidence for the prevalence of true compensatory
responses. Studies monitoring the growth of individually marked sh under manipulated and nonmanipulated conditions could provide considerable
information on the extent and costs of compensatory
growth in the wild (Letcher and Terrick 2001).
The consequences of ignoring compensatory
growth for general studies of sh growth have
received little attention. For example, the existence
of compensatory growth can put at risk the suitability of usual condition indices (Raikova-Petrova and
Zivkov 1998), growth back-calculation methods
(Zivkov1996) and usual growth parameters for comparing sh growth (Zivkov et al.1999).Vertical studies
comparing older and younger individuals make up a
body of evidence for optimising selection in natural
populations. Yet, the restrictive assumption that
growth is not compensatory must be met for such
methods to be valid (Travis 1989). The ecological
implications cannot be ignored. The indirect eects
of growth compensation are simply not known.
Reinhardt and Healey (1998) have speculated that
the presence of predators may serve to depress
growth rates of the largest individuals in a cohort
through risk-avoidance behaviour, which indirectly
would favour better feeding opportunities of the
smallest sh, less responsive to predator presence.
Most of the studies on compensatory growth in sh
are focused on overall size (generally length or mass),
whereas there is little understanding of how the compensatory growth processes can aect morphology
(Bell 1974; Emerson 1986). This is a major gap in our
knowledge because sh morphology can inuence
swimming performance (Hawkins and Quinn 1996;
Webb et al. 1996), feeding eciency (Wainwright
1987; Norton 1991; Wainwright and Richard 1995)
and escape from predators (Taylor and McPhail 1985;
Swain 1992). At least, two simple, but important,
questions still deserve investigation: (i) Do all body
dimensions maintain their relative growth over
the compensatory period (i.e. does compensatory
growth entail morphological costs)? (ii) Can environmentally induced deviations from the normalshape
be corrected by dierential, compensating growth of
the dierent body parts?
180

The taxonomic coverage of studies dealing with


compensatory growth is still limited. Comparative
studies involving closely related species diering in
ecological or life-historical characteristics can contribute to an understanding of the functional signicance of compensatory growth processes. Even
groups that have attracted maximum attention have
important taxonomic and ecological gaps. Among
salmonids, some widely distributed species like
brown trout have attracted little interest (for an
exception, see Pirhonen and Forsman 1998). In that
case, it is not possible to determine whether the failure to demonstrate compensatory growth reects a
true absence or simply lack of information.
There is a need to develop experimental protocols
that can be used to compare the capacity for compensatory growth for dierent ontogenetic stages within
species and between species. Imposing xed periods
of deprivation may lead to dierent degrees of growth
depression. The alternative is to set a target change
in body mass before re-alimentation is imposed
(Schwarz et al. 1985). The protocols should dene the
range of growth depression that evokes growth compensation and the degree of compensation that is
achieved. The need for comparative studies extends
to the dierent factors provoking growth losses that
can be reversed. Most studies on compensatory
growth in sh have examined responses to changes
in food availability, with only a few studies on the
eects of food quality, temperature or sh density.
No studies have tested whether sh compensate
growth losses imposed through predator avoidance,
although research is being conducted to test this idea
(D. Alvarez and A.G. Nicieza, unpublished data).
Whichever mechanisms operate, compensatory
growth represents a temporal, within-individual
level of phenotypic plasticityand a source of variation
that is often overlooked in comparative studies and
theoretical analysis of growth and developmental
strategies (Nicieza and Metcalfe 1997). If shown to be
a widespread response in natural populations, it
could have important implications for the dynamics
of populations and hence communities because of
its role in buering populations against the eects of
environmental variations (Wootton 1998). Targeted
growth also has important implications at the evolutionary level. First, it reduces or eliminates some of
the variance that would otherwise be available
for selection acting on nal size (Riska et al. 1984).
Second, it could reduce serial genetic autocorrelation, allowing dierent segments of development
and growth to respond more independently to
# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

age-specic selection pressures (Riska 1986). Even if


it proves to have only a minor role in natural populations, compensatory changes in appetite and growth
may be important in developing optimal patterns of
feeding in aquaculture facilities (Gaylord and Gatlin
2001).
Acknowledgements
M. Ali was supported by a scholarship from the Government of Pakistan. R. J.Wootton acknowledges the
support of the Royal Society of London ^ Chinese
Academy of Sciences Exchange award.We would like
to thank the late Prof. Yibo Cui for his comments on
an earlier draft of this review and the constructive
comments of two anonymous referees.
References
Abrahams, M.V. and Sutterlin, A. (1999) The foraging and
antipredator behaviour of growth-enhanced transgenic
Atlantic salmon. Animal Behaviour 58,933^942.
Abrams, P.A., Leimar, O., Nylin, S. and Wiklund, C. (1996)
The eect of exible growth rates on optimal sizes and
development times in a seasonal environment. American
Naturalist147,381^395.
Adams, S.M., McLean, R.B. and Human, M.M. (1982) Structuring of a predator population through temperaturemediated eects on prey availability. Canadian Journal of
Fisheries and Aquatic Sciences 39,1175^1184.
Ali, M. and Wootton, R.J. (1998) Do random uctuations in
the intervals between feeding aect growth rate in juvenile three-spined sticklebacks? Journal of Fish Biology
53,1006^1014.
Ali, M. and Wootton, R.J. (1999) Coping with resource variability: eect of constant and variable intervals between
feeding on reproductive performance at rst spawning
of female three-spined sticklebacks, Gasterosteus aculeatus L. Journal of Fish Biology 55, 211^220.
Ali, M. and Wootton, R.J. (2001) Capacity for growth compensation in juvenile three-spined sticklebacks experiencing food deprivation. Journal of Fish Biology 58,
1531^1544.
Ali, M., Przybylski, M. and Wootton, R.J. (1998) Do random
uctuations in daily ration aect the growth rate of juvenile three-spined sticklebacks? Journal of Fish Biology
52, 223^229.
Ali, M., Cui,Y., Zhu, X. and Wootton, R.J. (2001) Dynamics of
appetite in three sh species (Gasterosteus aculeatus,
Phoxinus phoxinus and Carassius auratus gibelio) after
feed deprivation. Aquaculture Research 32, 443^450.
Alvarez, D. (2002) Implicaciones del comportamiento y siolog| a individuales en el ciclo de vida y dinamica poblacional de
la trucha comun (Salmo trutta L.). Doctoral Thesis, University of Oviedo, Spain,109pages.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Aranda, A., Sanchez-Vazquez, F.J. and Madrid, J.A. (2001)


Eect of short-term fasting on macronutrient self-selection in sea bass. Physiology and Behaviour 73,105^109.
Aranyakananda, P., Moore, N. and Singhagraiwan,T. (1996)
Eect of feeding frequency on compensatory growth of
Asian sea bass, Lates calcarifer. Arri Newsletter 3,11^13.
Arendt, J.D. (1997) Adaptive intrinsic growth rates: An integration across taxa. Quarterly Review of Biology 72,
149^177.
Arendt, J.D. andWilson, D.S. (1997) Optimistic growth: competition and an ontogenetic niche-shift select for rapid
growth in pumpkinseed sunsh (Lepomis gibbosus).
Evolution 51,1946^1954.
Arendt, J.D. and Wilson, D.S. (2000) Population dierences
in the onset of cranial ossication in pumpkinseed (Lepomis gibbosus), a potential cost of growth. Canadian Journal
of Fisheries and Aquatic Sciences 57,351^356.
Arendt, J.D.,Wilson, D.S. and Starck, E. (2001) Scale strength
as a cost of rapid growth in sunsh. Oikos 93,95^100.
Arndt, S.K.A., Benfy, T.J. and Cunjak, R.A. (1996) Eect of
temporary reductions in feeding on protein synthesis
and energy storage of juvenile Atlantic salmon. Journal
of Fish Biology 49, 257^276.
Atchley, R.W. (1984) Ontogeny, timing of development, and
genetic variance-covariance structure. American Naturalist123,519^540.
Basiao, Z.U., Doyle, R.W. and Arago, A.L. (1996) A statistical
power analysis of the internal reference technique for
comparing growth and growth depensation of tilapia
strains. Journal of Fish Biology 49, 277^286.
Bastrop, R., Spangenberg, R. and Jurss, K. (1991) Biochemical adaptations of juvenile carp (Cyprinus carpio L.) to
food deprivation. Comparative Biochemistry and Physiology 98,143^149.
Beacham,T.D. (1981) Variability in growth during the rst 3
years of life of cod (Gadus morhua) in the southern Gulf
of St. Lawrence. CanadianJournal of Zoology 59,614^620.
Beamish, F.W.H. (1964) Inuence of starvation on standard
and routine oxygen consumption. Transactions of the
American Fisheries Society 50,153^164.
Bell, G. (1974) The reduction of morphological variation in
natural populations of smooth newt larvae. Journal of
Animal Ecology 43,115^128.
Bellamy, D. (1968) Metabolism of the red piranha (Rooseveltiella nattererri) in relation to feeding behaviour. Comparative Biochemistry and Physiology 25,343^347.
von Bertalany, L. (1960) Principles and theory of growth.
In: Fundamental Aspects of Normal and Malignant Growth
(ed W. Nowinski). Elsevier, Amsterdam, pp.137^252.
Bertram, D.F., Chambers, R.C. and Leggett,W.C. (1993) Negative correlations between larval and juvenile growthrates in winter ounder- implications of compensatory
growth for variation in size-at-age. Marine Ecology Progress in Series 96, 209^215.
Billerbeck, J.M., Lankford,T.E. Jr and Conover, D. (2001) Evolution of intrinsic growth and energy acquisition rates.

181

Compensatory growth in shes M Ali et al.

Part I. Trade-os with swimming performance in Menidia menidia. Evolution 55,1863^1872.


Bilton, H.T. and Robins, G.L. (1973) The eects of starvation
and subsequent feeding on survival and growth of Fulton
Channel Sockeye Salmon fry (Oncorhynchus nerka). Journal of the Fisheries Research Board of Canada 30,1^5.
Bilton, H.T., Alderdice, D.F. and Schnute, J.T. (1982) Inuence
of time and size at release of juvenile coho salmon
(Oncorhynchus kisutch) on returns at maturity. Canadian
Journal of Fisheries and Aquatic Sciences 39, 426^447.
Bjornsson, B., Sigurthorsson, G., Hemre, G.I. and Lie, O.
(1992) Growth rate and feed conversion factor on young
halibut (Hippoglossus hippoglossus L.) fed six dierent
diets. Fiskeridirektoratets Skrifter Ernaering 5, 25^35.
Blasco, J., Fernandez, J. and Gutierrez, J. (1992a) Variations in
tissue reserves, plasma metabolites and pancreatic hormones during fasting in immature carp (Cyprinus carpio).
Comparative Biochemistry and Physiology A-Physiology
103,357^363.
Blasco, J., Fernandez, J. and Gutierrez, J. (1992b) Fasting and
refeeding in carp, Cyprinus carpio L. ^ the mobilization of
reserves plasma metabolites and hormone variations.
Journal of Comparative Physiology B-Biochemical Systemic
and Environmental Physiology162,539^546.
Blaxter, J.H.S. and Ehrlich, K.F. (1974) Change in behaviour
during starvation of herring and plaice larvae. In: The
Early Life History of Fish (ed J.H.S. Blaxter), Springer
Verlag, Heidelberg, pp.575^588.
Blaxter, J.H.S. and Hempel, G. (1963) The inuence of egg
size on herring larvae (Clupea harengus L.). Journal du Conseil international pour lExploration de la Mer 28, 211^240.
Boujard, T., Burel, C., Medale, F., Haylor, G. and Moisan, A.
(2000) Eect of past nutritional history and fasting on
feed intake and growth in rainbow trout Oncorhynchus
mykiss. Aquatic Living Resources13,129^137.
Brett, J.M. (1971) Satiation time, appetite and maximum food
intake of sockeye salmon (Oncorhynchus nerka). Journal
of Fisheries Research Board of Canada 28, 409^415.
Broekhuizen, N., Gurney,W.S.C., Jones, A. and Bryant, A.D.
(1994) Modelling compensatory growth. Functional Ecology 8,770^782.
Brown, M.E. (1946) The growth of brown trout (Salmo trutta
Linn.) II The growth of two-year-old trout at constant
temperature of 11.5 8C. Journal of Experimental Biology
22,130^144.
Brown, M.E. (1957) Experimental studies on growth. In:The
Physiology of Fishes, Vol. I (ed. M.E. Brown), Academic
Press, NewYork, pp.361^400.
Brown, C.R. and Cameron, J.N. (1991) The relationship
between specic dynamic action (SDA) and protein
synthesis rates in the channel catsh. Physiological Zoology 64, 298^309.
Buckel, J.A., Letcher, B.H. and Conover, D.O. (1998) Eects
of a delayed onset of piscivory on the size of age-0 bluesh. Transactions of American Fisheries Society 127,
576^587.

182

Buckley, L.J. (1979) Relationships between RNA-DNA ratio,


prey density, and growth rate in Atlantic cod (Gadus morhua) larvae. Journal of Fisheries Research Board of Canada
36,1497^1502.
Bull, C.D. and Metcalfe, N.B. (1997) Regulation of hyperphagia in response to varying energy decits in over-wintering juvenile Atlantic salmon. Journal of Fish Biology 50,
498^510.
Bull, C.D., Metcalfe, N.B. and Mangel, M. (1996) Seasonal
matching of foraging eort to anticipated energy requirements in anorexic juvenile salmon. Proceedings of the
Royal Society of London B 263,13^18.
Carter, C., Houlihan, D., Kiessling, A., Medalo, F. and Jobling,
M. (2001) Physiological eects of feeding. In: Feed Intake
in Fish (eds D. Houlihan, T. Boujard, T. and M. Jobling),
Blackwell Scientic, Oxford, pp. 297^331.
Case,T.J. (1978) On the evolution and adaptive signicance of
postnatal growth rates in terrestrial vertebrates. Quarterly Review of Biology 53, 243^279.
Chambers, R.C. Leggett, W.C. and Brown, J.A. (1988) Variation in and among early life history traits of laboratoryreared winter ounder Pseudopleuronectes americanus.
Marine Ecology Progress in Series 47,1^15.
Chatakondi, N.G. and Yant, R.D. (2001) Application of compensatory growth to enhance production in channel catsh Ictalurus punctatus. Journal of World Aquaculture
Society of 32, 278^285.
Chavin, W. and Young, J.E. (1970) Factors in the determination of normal serum glucose levels of goldsh, Carasssius auratus L. Comparative Biochemistry and Physiology
33,629^653.
Chmilevskii, D.A. (1994) Vliyanie ponizhennoj temperatury
na on oogenez tilapii Oreochromis mossambicus 2. Vozdejstvie na ryb v vozraste dvadtsati dvukh sutok posle vylupleniya.Voprosy Ikhtiologii 34,675^680. [In Russian].
Chmilevskii, D.A. (1996) The inuence of a low temperature
on the oogenesis of tilapia Oreochromis mossambicus 4.
Eect on sh 106 days after their hatching. Journal of
Ichthyology 36,615^620.
Chmilevskii, D.A. (1998) The inuence of low temperature
on the growth of Oreochromis mossambicus. Journal of
Ichthyology 38,86^92.
Cho, C.Y. Slinger, S.J. and Bayley, H.S. (1976) Inuence of level
and type of dietary protein and of level of feeding on feed
utilization by rainbow trout. Journal of Nutrition 106,
1547^1556.
Christensen, S.M. and McLean, E. (1998) Compensatory
growth in Mozambique tilapia (Oreochromis mossambicus) fed a sub-optimal diet. Ribarstvo 56,3^19.
Christiansen, J.S. Martinez, I. Jobling, M. and Amin, A.B.
(1992) Rapid somatic growth and muscle damage in a
salmonid sh. Basic and Applied Myology 2, 235^239.
Collins, A.L. and Anderson, T.A. (1995) The regulation of
endogenous energy stores during starvation and refeeding in the somatic tissues of the golden perch. Journal of
Fish Biology 47,1004^1015.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

Conover, D.O. (1992) Seasonality and the scheduling of lifehistory at dierent latitudes. Journal of Fish Biology 41,
161^178.
Conover, D.O. and Ross, M.R. (1982) Patterns in seasonal
abundance, growth and biomass of the Atlantic silverside, Menidia menidia, in a New England estuary.
Estuaries 5, 275^286.
Cui, Y. and Wootton, R.J. (1989) Bioenergetics of growth of
a cyprinid, Phoxinus phoxinus (L.): development and
testing of growth model. Journal of Fish Biology 34,
47^64.
Cutts, C.J. Metcalfe, N.B. and Taylor, A.C. (2002) Juvenile
Atlantic salmon (Salmo salar) with relatively high standard metabolic rates have smal metabolic scopes. Functional Ecology16,74^78.
Dabrowski, K.Takashima, F. and Strussmann, C. (1986) Does
recovery growth occur in larval sh Bulletin of the of
Japanese Society of Science 52,1869.
DamsgQrd, B. and Arnesen, A.M. (1998) Feeding, growth
and social interactions during smolting and seawater
acclimation in Atlantic salmon, Salmo salar L. Aquaculture168,7^16.
DamsgQrd, B. and Dill, L.M. (1998) Risk-taking behaviour in
weight compensating coho salmon, Oncorhynchus
kisutch. Behavioural Ecology 9, 26^32.
DePedro, N. and Bjornsson, B.T. (2001) Regulation of food
intake by neuropeptides and hormones. In: Feed Intake
in Fish (eds D. Houlihan,T. Boujard and M. Jobling), Blackwell Scientic, Oxford, pp. 297^331.
van Dijk, P.L.M. Staaks, G. and Hardewig, I. (2002) The eect
of fasting and refeeding on temperature preference,
activity and growth of roach, Rutilus rutilus. Oecologia
130, 496^504.
Dobson, S.H. and Holmes, R.M. (1984) Compensatory
growth in the rainbow trout, Salmo gairdneri Richardson. Journal of Fish Biology 25,649^656.
Du Preez, H.H. (1987) Laboratory studies on the oxygen consumption of the marine teleost, Lichia amia (Linnaeus,
1758). Comparative Biochemistry and Physiology 88A,
523^532.
Du Preez, H.H. McLachlan, A. and Marais, J.K.F. (1986a) Oxygen consumption of a shallow water teleost, the spotted
grunter, Pomadasys commersonii (Lacepe'de, 1802). Comparative Biochemistry and Physiology 84A,61^70.
Du Preez, H.H. Strydom,W. and Winter, P.E.D. (1986b) Oxygen consumption of two teleosts Lithognathus mormyrus
(Linnaeus, 1758) and Lithognathus lithognathus (Cuvier,
1830) (Teleosti: Sparidae). Comparative Biochemistry and
Physiology 85A,313^331.
Dunbrack, R.L. (1988) Feeding of juvenile coho salmon
(Oncorhynchus kisutch): maximum appetite, sustained
feeding rate, appetite return, and body size. Canadian
Journal of Fisheries and Aquatic Sciences 45,1191^1196.
Dunbrack, R.L. and Dill, L.M. (1983) A model of size dependent surface feeding in a stream dwelling salmonid.
Environmental Biology of Fishes 8, 203^216.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Egglishaw, H.J. and Shackley, P.E. (1977) Growth, survival


and production of juvenile salmon and trout in a Scottish
stream. Journal of Fish Biology11,647^672.
Elliott, J.M. (1972) Rates of gastric evacuation in brown trout,
Salmo trutta L. Freshwater Biology 2,1^18.
Elliott, J.M. (1975a) Weight of food and time required to satiate brown trout, Salmo trutta L. Freshwater Biology 5,
51^64.
Elliott, J.M. (1975b) Number of meals in a day, maximum
weight of food consumed in a day and maximum rate of
feeding for brown trout Salmo trutta. Freshwater Biology
5, 287^303.
Elliott, J.M. (1976) Energy losses in the waste products of
brown trout (Salmo trutta L.). Journal of Animal Ecology
45,561^580.
Emerson, S.B. (1986) Heterochrony and frogs: the relationship of a life history trait to morphological form. American Naturalist127,163^183.
Farbridge, K.J. and Leatherland, J.F. (1992a) Plasma growthhormone levels in fed fasted rainbow trout (Oncorhynchus mykiss) are decreased following handling stress.
Fish Physiology and Biochemistry10,67^73.
Farbridge, K.J. and Leatherland, J.F. (1992b) Temporal
changes in plasma thyroid-hormone, growth-hormone
and free fatty-acid concentrations, and hepatic 50 -monodeiodinase on refeeding in rainbow trout (Oncorhynchus
mykiss). Fish Physiology and Biochemistry10, 245^257.
Farbridge, K.J. Flett, P.A. and Leatherland, J.F. (1992)Temporal
eects of restricted diet and compensatory increased dietary intake on thyroid-function, plasma growth-hormone
levels and tissue lipid reserves of rainbow trout (Oncorhynchus mykiss). Aquaculture104,157^174.
Farrell, A.P. Bennett, W.B. and Devlin, R.H. (1997) Growthenhanced transgenic salmon can be inferior swimmers.
CanadianJournal of Zoology 75,335^337.
Ferreri, C.P. and Taylor,W.W. (1996) Compensation in individual growth-rates and its inuence on lake trout population-dynamics in the Michigan waters of Lake Superior.
Journal of Fish Biology 49,763^777.
Fletcher, D.A. (1984) The physiological control of appetite
in sh. Comparative Biochemistry and Physiology 78A,
617^628.
Foss, A. and Imsland, A.K. (2002) Compensatory growth in
the spotted wolsh Anarhichas minor (Olafsen) after a
period of limited oxygen supply. Aquaculture Research
33,1097^1101.
From, J. and Rasmussen, G. (1984) A growth model, gastric
evacuation and body composition in rainbow trout,
Salmo gairdneri Richardson,1836. Dana 3,61^139.
Gardiner, W.R. and Geddes, P. (1980) The inuence of body
composition on the survival of juvenile salmon. Hydrobiologia 69,67^72.
Gaylord,T.G. and Gatlin,D.M. III (2001) Dietary protein and
energy modications to maximize compensatory
growth of channel catsh (Ictalurus punctatus). Aquaculture194,337^348.

183

Compensatory growth in shes M Ali et al.

Gaylord, T.G. MacKenzie, D.S. and Gatlin,D.M. III (2001)


Growth performance, body composition and plasma
thyroid hormone status of channel catsh (Ictalurus
punctatus) in response to short-term feed deprivation
and refeeding. Fish Physiology and Biochemistry 24,
73^79.
Gerking, S.D. (1994) Feeding Ecology of Fish. Academic Press,
London.
Gerking, S.D. and Raush, R.R. (1979) Relative importance of
size and chronological age in the life programme of
shes. Ergebnis Limnologie1979,181^194.
Goddard, J.S. (1974) An x-ray investigation of the eects of
starvation and drugs on intestinal motility in the plaice,
Pleuronectes platessa L. Icthyologica 6, 49^58.
Godin, J.-G.J. (1981) Eect of hunger on the daily pattern of
feeding rates in juvenile pink salmon (Oncorhynchus gorbuscha). Journal of Fish Biology19,63^71.
Gorman, G.C. and Nielson, L.A. (1982) Piscivory by stocked
brown trout (Salmo trutta) and its impact on the nongame sh community of Bottom Creek, Virginia. CanadianJournal of Fisheries and Aquatic Sciences 39,862^869.
Gross, M.R. (1987) The evolution of disdromy in shes. American Fisheries Society of Symposium1,14^25.
Gross, M.R. Coleman, R.M. and McDowall, R.M. (1988)Aquatic productivity and the evolution of diadromous sh
migration. Science 239,1291^1293.
Grove, D.J. Lozoides, L. and Nott, J. (1978) Satiation amount,
frequency of feeding and gastric emptying rate in Salmo
gairdneri. Journal of Fish Biology12,507^516.
Grove, D.J. Mortezuma, M.A. Flett, H.R.J. Foot, J.S.Watson,T.
and Flowerdew, M.W. (1985) Gastric emptying and the
return of appetite in juvenile turbot, Scophthalmus maximus L. fed on articial diets. Journal of Fish Biology 26,
339^354.
Haweet, A. and Ozawa, T. (1996) Age and growth of ribbon
sh Trichiurus japonicus in Kagoshima Bay, Japan. Fisheries Science 62,529^533.
Hawkins, D.K. and Quinn, T.P. (1996) Critical swimming
velocity and associated morphology of juvenile coastal
cutthroat trout (Oncorhynchus clarki clarki), steelhead
trout (Oncorhynchus mykiss), and their hybrids. Canadian
Journal of Fisheries and Aquatic Sciences 53,1487^1496.
Hayward, R.S. and Wang, N. (2001) Failure to induce overcompensation of growth in maturing yellow perch. Journal of Fish Biology 59,126^140.
Hayward, R.S. Noltie, D.B. and Wang, N. (1997) Use of compensatory growth to double hybrid sunsh growth
rates. Transactions of American Fisheries Society of 126,
316^322.
Hayward, R.S.Wang, N. and Noltie, D.B. (2000) Group holding impedes compensatory growth of hybrid sunsh.
Aquaculture183, 299^305.
Healey, M.C. (1982) Timing and relative intensity of size
selective mortality of juvenile chum salmon (Oncorhynchus keta) during early sea life. Canadian Journal of
Fisheries and Aquatic Sciences 39,952^957.

184

Henderson, M.A. and Cass, A.J. (1991) Eect of smolt size on


smolt-to-adult survival for Chilko Lake sockeye salmon
(Oncorhynchus nerka). Canadian Journal of Fisheries and
Aquatic Sciences 48,988^994.
Henderson, P.A., Holmes, R.H.A. and Bamber, R.N. (1988)
Size-selective overwintering mortality in the sand smelt,
Atherina boyeri Risso, and its role in population regulation. Journal of Fish Biology 33, 221^233.
Hepher, B. Liao, I.C. Cheng, S.H. and Hsieh, C.S. (1983) Food
utilization by tilapia-eect of diet composition, feeding
level and temperature on utilization eciency for maintenance and growth. Aquaculture 32, 255^275.
Hoebel, B.G. (1997) Neuroscience and appetitive behaviour
research: 25 years. Appetite 29,119^133.
Houlihan, D.F. Mathers, E.M. and Foster, A. (1993) Biochemical correlates of growth rate in sh. In: Fish Ecophysiology (eds J.C. Rankin and F.B. Jensen), Chapman & Hall,
London, pp. 45^71.
Houston, A.I. and McNamara, J.M. (1999) Models of Adaptive
Behaviour. Cambridge University Press, Cambridge.
Houston, A.I. McNamara, J.M. and Hutchinson, J.M.C. (1993)
General results concerning the trade-o between gaining energy and avoiding predation. Philosophical Transactions of the Royal Society of London B 341,375^397.
Hubbell, S.P. (1971) Of sowbugs and systems: the ecological
bioenergetics of a terrestrial isopod. In: Systems Analysis
and Simulation in Ecology (ed. B.C. Patten), Academic
Press, London, pp. 269^323.
Hunt, R.L. (1969) Overwinter survival of wild ngerling
brook trout in Lawrence Creek,Wisconsin. Journal of the
Fisheries Research Board of Canada 26,1473^1483.
Hutchings, J.A. (1994) Age-specic and size-specic costs of
reproduction within population of brook trout Salvelinus
fontinalis. Oikos 70,12^20.
Ince, B.W. and Thorpe, A. (1976) The eects of starvation and
force-feeding on the metabolism of the Northern Pike,
Esox lucius L. Journal of Fish Biology 8,79^88.
Jezierska, B. Hazel, J.R. and Gerking, S.D. (1982) Lipid mobilization during starvation in the rainbow trout, Salmo
gairdneri Richardson, with attention to fatty-acids. Journal of Fish Biology 21,681^692.
Jobling, M. (1980) Eects of starvation on proximate chemical composition and energy utilization of plaice, Pleuronectes platessa L. Journal of Fish Biology17,325^334.
Jobling, M. (1982) Some observations on the eects of feeding frequency on the food intake and growth of plaice,
Pleuronectes platessa L. Journal of Fish Biology 20,
431^444.
Jobling, M. (1993) Bioenergetics: feed intake and energy partitioning. In: Fish Ecophysiology (eds J.C. Rankin and F.B.
Jensen), Chapman & Hall, London, pp.1^44.
Jobling, M. (1994) Fish Bioenergetics. Chapman & Hall,
London.
Jobling, M. and Johansen, S.J.S. (1999) The lipostat, hyperphagia and catch-up growth. Aquaculture Research 30,
473^478.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

Jobling, M. and Koskela, J. (1996) Interindividual variations


in feeding and growth in rainbow trout during restricted
feeding and in subsequent period of compensatory
growth. Journal of Fish Biology 49,658^667.
Jobling, M. Jorgensen, E.H. and Siikavuopio, S.I. (1993) The
inuence of previous feeding regime on the compensatory growth response of maturing and immature Arctic
charr, Salvelinus alpinus. Journal of Fish Biology 43, 409^
419.
Jobling, M. Meloy, O.H. dos Santos, J. and Christiansen, B.
(1994) The compensatory growth response of the Atlantic cod: eects of nutritional history. Aquaculture International 2,75^90.
Johal, M.S. and Tandon, K.K. (1986) Phenomenon of growth
compensation in Cyprinus carpio. IndianJournal ofEcology
13,350^352.
Johansen, S.J.S. Ekli, M. Stanges, B. and Jobling, M. (2001)
Weight gain and lipid deposition in Atlantic salmon,
Salmo salar, during compensatory growth: evidence for
lipostatic regulation? Aquaculture Research 32,963^974.
Johansen, S.J.S. Ekli, M. and Jobling, M. (2002) Is there lipostatic regulation of feed intake in Atlantic salmon Salmo
salar L.? Aquaculture Research 33,515^524.
Johnson,T.B. and Evans, D.O. (1990) Size-dependent winter
mortality of young-of-the-year white perch: climate
warming and invasion of the Laurentian Great Lakes.
Transactions of the American Fisheries Society 119, 301^
313.
Johnson, T.B. and Evans, D.O. (1991) Behaviour, energetics,
and associated mortality of young-of-the-year white
perch (Morone americana) and yellow perch (Perca avescens) under simulated winter conditions. Canadian Journal of Fisheries and Aquatic Scicences 48,672^680.
Johnsson, J.I. (1993) Big and brave: size selection aects foraging under risk of predation in juvenile rainbow trout,
Oncorhynchus mykiss. Animal Behaviour 45,1219^1225.
Johnsson, J.I. and Bjornsson, B.T. (1994) Growth hormone
increases growth rate, appetite and dominance in juvenile rainbow trout, Oncorhynchus mykiss. Animal Behaviour 47,177^186.
Johnsson, J.I. Petersson, E.J. Jonsson, E. Bjornsson, B.T. and
Jarvi,T. (1996) Domestication and growth hormone alter
antipredator behaviour and growth patterns in juvenile
brown trout, Salmo trutta. Canadian Journal of Fisheries
and Aquatic Sciences 53,1546^1554.
Johnsson, J.I. Petersson, E.J. Jonsson, E. Jarvi,T. and Bjornsson, B.Th (1999) Growth hormone-induced eects on
mortality, energy status and growth: a eld study on
brown trout. Functional Ecology13,514^522.
Johnston, I.A. (1981) Quantitative analysis of muscle breakdown during starvation in the marine atsh Pleuronectes platessa. Cell andTissue Research 214,369^386.
Jonsson, E. Johnsson, J.I. and Bjornsson, B.Th (1996) Growth
hormone increases predation exposure of rainbow
trout. Proceedings of the Royal Society of London B 263,
647^651.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Karlsen, O. Holm, J.C. and Kjesbu, O.S. (1995) Eects of periodic starvation on reproductive investment in 1st-time
spawning Atlantic cod (Gadus morhua L). Aquaculture
133,159^170.
Kim, M.K. and Lovell, R.T. (1995) Eect of feeding regimes on
compensatory weight gain and body tissue changes in
channel catsh, Ictalurus punctatus in ponds. Aquaculture135, 285^293.
Kindschi, G.A. (1988) Eect of intermittent feeding on
growth of rainbow trout, Salmo gairdneri Richardson.
Aquaculture and Fisheries Management19, 213^215.
Kinne, O. (1960) Growth, food intake and food conversion in
an euryplastic sh exposed to dierent temperatures
and salinities. Physiological Zoology 33, 288^317.
Kitchell, J.F. Stewart, D.J. and Weininger, D. (1977) Applications of bioenergetics model to yellow perch (Perca avescens) and walleye (Stizostedion vitreum vitreum).
Journal of Fisheries Research Board of Canada 34, 1922^
1935.
Koppe,W. Pockrandt, J. Meyer-Burgdor, K.H. and Gunther,
K.D. (1993) Eects of realimentation after a period of
restricted feeding on food intake, growth and body composition in Piaractus brachypomus (Cuvier 1818), a South
American characoid sh. In: Fish Ecotoxicology and Ecophysiology (eds T. Braunbeck, W. Hanke and H. Segner),
VCH, NewYork, pp. 263^268.
Lagardere, F. (1989) Inuence of feeding conditions and temperature on growth rate and otolith increment deposition of larval Dover sole (Solea solea L.). Rapports Process:
Verbaux Reunion Conseil International Pour lExploration
de la Mer191,390^399.
Lankford, T.E., Jr, Billerbeck, J.M. and Conover, D.O. (2001)
Evolution of intrinsic growth and energy acquisition
rates. II. Trade-os with vulnerability to predation. Evolution 55,1873^1881.
Larsson, A. and Lewander, K. (1973) Metabolic eects of
starvation in the eel, Anguilla anguilla L. Comparative
Biochemistry and Physiology 44A,367^374.
Le Bail, P.-Y. and Boeuf, G. (1997) What hormones may
regulate food intake in sh? Aquatic Living Resources 10,
371^379.
Le Bail, P.-Y. Perez-Sanchez, J. Yao, K. and Maisse, G. (1993)
Eect of GH treatment on salmonid growth, study of the
variability of the response. In: Aquaculture, Fundamental
and Applied Research (eds B. Lalhou and P.Vitiello), AGU,
Washington, DC, pp.137^197.
Leamy, L. and Atchley, W. (1985) Directional selection and
developmental stability: evidence from uctuating
asymmetric of morphometric characters in rats. Growth
49,8^18.
Leatherland, J.F. and Farbridge, K.J. (1992) Chronic fasting
reduces the response of the thyroid to growth-hormone
and TSH, and alters the growth hormone-related
changes in hepatic 50 - monodeiodenase activity in rainbow trout, Oncorhynchus mykiss. General and Comparative Endocrinology 87,342^353.

185

Compensatory growth in shes M Ali et al.

Lee, D.J. and Putnam, C.B. (1973) The response of rainbow


trout to varying protein/energy ratios in a test diet. Journal of Nutrition103,916^922.
Letcher, B.H. and Terrick,T. (2001) Eects of developmental
stage at stocking on growth and survival of Atlantic salmon fry. North American Journal of Fisheries Management
21,102^110.
Letcher, B.H. Rice, J.A. and Crowder, L.B. (1996) Size-dependent eects of continuous and intermittent feeding on
starvation time and mass loss in starving yellow perch
larvae and juveniles. Transactions of the American Fisheries Society125,14^26.
Litvak, M.K. and Leggett,W.C. (1992) Age and size-selective
predation on larval shes: the bigger-is-better hypothesis
revisited. Marine Ecology: Progress Series 81,13^24.
Love, R.M. (1970) The Chemical Biology of Fishes. Academic
Press, London.
Love, R.M. (1980) The Chemical Biology of Fishes, Vol. 2.
Academic Press, London.
Lovell, R.T. (1979) Factors aecting voluntary food consumption of the channel catsh. Proceedings of Annual
Conference S.E. Association of Fisheries andWildlife Agency
33,563^571.
Ludwig, D. and Rowe, L. (1990) Life-history strategies for
energy gain and predator avoidance under time constraints. American Naturalist135,686^707.
Lundqvist, H. McKinnell, S. Fangstam, H. and Berglund, I.
(1994) The eect of time, size and sex on recapture rates
and yield after river releases of Salmo salar smolts. Aquaculture121, 245^257.
Luquet, P. Oteme, Z.J. and Cisse, A. (1995) Evidence for compensatory growth and its utility in the culture of Heterobranchus-longilis. Aquatic Living Resources 8,389^394.
Macchiusi, F. and Baker, R.L. (1991) Prey behaviour and sizeselective predation by sh. Freshwater Biology 25,
533^538.
Machiels, M.A.M. and van Dam, A.A. (1987) A dynamic
simulation ^ model for growth of the African catsh,
Clarias gariepinus (Burchell 1822). Part 3. The eect of
body composition on growth and feed intake. Aquaculture 60,55^71.
Mackas, D.L. Denman, K.L. and Abbott, M.R. (1985) Plankton patchiness: biology in the physical vernacular. Bulletin of the of Marine Science 37,652^674.
Maclean, A. and Metcalfe, N.B. (2001) Social status, access
to food and compensatory growth in juvenile Atlantic
salmon. Journal of Fish Biology 58,1331^1346.
Magnusson, J.J. (1969) Digestion and food consumed by
skipjack tuna, Katsuwanus pelamis. Transactions of the
American Fisheries Society 98,379^392.
Marais, J.K.J. and Kissil, G.W. (1979) The inuence of energy
level on the feed intake, growth, food conversion and
composition of Sparus aurata. Aquaculture17, 203^220.
McDowall, R.M. (2001) Anadromy and homing: two life-history traits with adaptive synergies in salmonid shes.
Fish and Fisheries 2,78^85.

186

McGurk, M.D. (1986) Natural mortality of marine pelagic


eggs and larvae: role of spatial patchiness. Marine Ecology: Progress Series 34, 227^242.
McMillan, D.N. and Houlihan, D.F. (1992) Protein synthesis
in trout liver is stimulated by both feeding and fasting.
Fish Physiology and Biochemistry10, 23^34.
McLean, E. Teskeredzic, E. Donaldson, E.M. et al. (1992)
Accelerated growth of coho salmon Oncorhynchus
kisutch following sustained release of recombinant porcine somatotropin. Aquaculture103,377^387.
Mendez, G. andWieser,W. (1993) Metabolic responses to food
deprivation and refeeding in juveniles of Rutilus rutilus
(Teleostei: Cyprinidae). Environmental Biology of Fishes
36,73^81.
Metcalfe, N.B. and Monaghan, P. (2001) Compensation for a
bad start: grow now, pay later.Trends in Ecology andEvolution16, 255^260.
Metcalfe, N.B. andThorpe, J.E. (1992) Anorexia and defended
energy levels in over-wintering juvenile salmon. Journal
of Animal Ecology 61,175^181.
Metcalfe, N.B., Bull, C.D. and Mangel, M. (2002) Seasonal
variation in catch-up growth reveals state-dependent
somatic allocations in salmon. Evolutionary Ecology
Research 4,871^881.
Miglavs, I. and Jobling, M. (1989a) The eect of feeding
regime on proximate body composition and patterns of
energy deposition in juvenileArctic charr, Salvelinus alpinus. Journal of Fish Biology 35,1^11.
Miglavs, I. and Jobling, M. (1989b) Eects of feeding regime
on food consumption, growth rates and tissue nucleic
acids in juvenile Arctic charr, Salvelinus alpinus, with
particular respect to compensatory growth. Journal of
Fish Biology 34,947^957.
Miller,T.J., Crowder, L.B., Rice, J.A. and Marshall, E.A. (1988)
Larval size and recruitment mechanisms in shes:
toward a conceptual framework. Canadian Journal of
Fisheries and Aquatic Science 45,1657^1670.
Mommsen, T.P. (1998) Growth and metabolism. In: The
Physiology of Fishes, 2nd edn. (ed. D.H. Evans), CRC Press,
Boca Raton, pp.65^97.
Monteiro, L.S. and Falconer, D.S. (1966) Compensatory
growth and sexual maturity in mice. Animal Production
8,179^192.
Moores, J.A. and Winters, G.H. (1982) Growth patterns in a
Newfoundland Atlantic herring (Clupea harengus harengus) stock. Canadian Journal of Fisheries and Aquatic
Sciences 39, 454^461.
Morgan, I.J. and Metcalfe, N.B. (2001) Deferred costs of compensatory growth after autumnal food shortage in juvenile salmon. Proceedings of the Royal Society of London B
268, 295^301.
Mortensen, A. and DamsgQrd, B. (1993) Compensatory
growth and weight segregation following light and temperature manipulation of juvenile Atlantic salmon
(Salmo salar L.) and Arctic charr (Salvelinus alpinus L.).
Aquaculture114, 261^272.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

Nagai, M. and Ikeda, S. (1971) Carbohydrate metabolism in


sh. I. Eects of starvation and dietary composition on
the blood glucose level and the hepatopancreatic glycogen and lipid contents in carp. Bulletin of the of Japanese
Society of Scientic Fisheries 37, 404^409.
Nicieza, A.G. (1993) Estrategias de desarrollo y reproduccion en
el Salmon Atlantico Salmo salar L. PhD Thesis, University
of Oviedo, 206 pages.
Nicieza, A.G. and Brana, F. (1993a) Relationships among
smolt size, marine growth, and sea age at maturity of
Atlantic salmon (Salmo salar) in Northern Spain. CanadianJournal of Fisheries and Aquatic Science 50,1632^1640.
Nicieza, A.G. and Brana, F. (1993b) Compensatory growth
and optimum size in one-year-old smolts of Atlantic
salmon (Salmo salar). In: Production of Juvenile Atlantic
Salmon, Salmo salar, in Natural Waters (eds R.J. Gibson
and R.E. Cutting), Canadian Special Publications in Fisheries and Aquatic Sciences No.118, 225^237.
Nicieza, A.G. and Metcalfe, N.B. (1997) Growth compensation in juvenile Atlantic salmon: Responses to depressed
temperature and food availability. Ecology 78, 2385^
2400.
Nicieza, A.G. and Metcalfe, N.B. (1999) Costs of rapid
growth: the risk of aggression is higher for fast-growing
salmon. Functional Ecology13,793^800.
Nicieza, A.G., Reyes-Gavilan, F.G. and Brana, F. (1994a) Differentiation in juvenile growth and bimodality patterns
between northern and southern-populations of Atlantic
salmon (Salmo salar L). Canadian Journal of Zoology 72,
1603^1610.
Nicieza, A.G., Reiriz, L. and Brana, F. (1994b) Variation in
digestive performance between geographically disjunct
populations of Atlantic salmon: countergradient in passage time and digestion rate. Oecologia 99, 243^251.
Norton, S.F. (1991) Capture success and diet of cottid shes:
the role of predator morphology and attack kinematics.
Ecology 72,1807^1819.
Oliver, J.D., Holeton, G.F. and Chua, K.F. (1979) Overwinter
mortality of ngerling smallmouth bass in relation to
size, relative energy stores, and environmental temperature. Transactions of the American Fisheries Society 108,
130^136.
Owen, R.W. (1989) Microscale and nescale variations of
small plankton in coastal and pelagic environments.
Journal of Marine Research 47,197^240.
Page, G.W. and Andrews, J.W. (1973) Interactions of dietary
levels of protein and energy on channel catsh (Ictalurus
punctatus). Journal of Nutrition103,1339^1346.
Pandian,T.J. (1967) Transformation of food in the sh Megalops cyprinoides. Part II. Inuence of quantity of food.
Marine Biology1,107^109.
Parker, G.A. (1992) The evolution of sexual size dimorphism
in sh. Journal of Fish Biology 41 (Suppl. B),1^20.
Parker, R.R. (1971) Size-selective predation among juvenile
salmonid shes in a British Columbia inlet. Journal of the
Fisheries Research Board of Canada 28,1503^1510.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Paul, A.J., Paul, J.M. and Smith, R.L. (1995) Compensatory


growth in Alaska yellown sole, Pleuronectes asper,
following food deprivation. Journal of Fish Biology 46,
442^448.
Pedersen, B.H. (1993) Growth and mortality in young larval
herring (Clupea harengus); eect of repetitive changes in
food availability. Marine Biology117,547^550.
Pedersen, B.H., Ugelstad, I. and Hjelmeland, K. (1990) Eects
of transitory, low food supply in the early life of larval
herring (Clupea harengus) on mortality, growth and
digestive capacity. Marine Biology107,61^66.
Perez-Sanchez, J. and Le Bail, P.-Y. (1999) Growth hormone
axis as a marker of nutritional status and growth performance in sh. Aquaculture177,117^128.
Peter, R.E. (1979) The brain and feeding behaviour. In: Fish
Physiology, Vol. VIII (eds W.S. Hoar, D.J. Randall and J.R.
Brett), Academic Press, London, pp.121^159.
Pirhonen, J. and Forsman, L. (1998) Eect of prolonged feed
restriction on size variation, feed consumption, body
composition, growth and smolting of brown trout, Salmo
trutta. Aquaculture162, 203^217.
Plank, J., Kirchgessner, M. and Schwarz, F.J. (1984) Einuss
einer Futterrestriktion und ^realimentation auf Wachstum, Futteraufwand und Schlachtkorperzusammensetzung von Karpfen (Cyprinus carpio). Zeitschrift Fur
Tierphysiologie 53,138^149. [In German]
Post, J.R. and Evans, D.O. (1989) Size dependent overwintering mortality of young-of-the-year yellow perch (Perca
avescens): laboratory, in situ enclosure, and eld experiments. Canadian Journal of Fisheries and Aquatic Sciences
46,1958^1968.
Priede, I.G. (1985) Metabolic scope in shes. In: Fish Energetics New Perspectives (eds P. Tytler, and P. Calow),
Croom-Helm, London, pp.33^64.
Purchase, C.F. and Brown, J.A. (2001) Stock-specic changes
in growth rates, food conversion eciencies, and energy
allocation in response to temperature change in juvenile
Atlantic cod. Journal of Fish Biology 58,36^52.
Qian, X., Cui, Y., Xiong, B. and Yang, Y. (2000) Compensatory growth, feed utilization and activity in gibel carp,
following feed deprivation. Journal of Fish Biology 56,
228^2232.
Quinton, J.C. and Blake, R.W. (1990) The eect of feed cycling
and ration level on the compensatory growth response
in rainbow trout, Oncorhynchus mykiss. Journal of Fish
Biology 37,33^41.
Raikova-Petrova, G.N. and Zivkov, M.T. (1998) Growth selfregulation: a reason for the variability of sh condition
indices. International Revue of Hydrobiology 83,599^602.
Reimers, E., Kjorrefjord, G. and Stavostrand, M. (1993) Compensatory growth and reduced maturation in second
sea winter farmed Atlantic salmon following starvation
in February and March. Journal of Fish Biology 43,
805^810.
Reinhardt, U.G. and Healey, M.C. (1998) Predation risk as
an opportunity for compensatory growth in juvenile

187

Compensatory growth in shes M Ali et al.

coho salmon? North Pacic Anadromous Fish Commission


Bulletin1, 403.
Ricker,W.E. (1969) Eects of size-selective mortality, production and yield. Journal of the Fisheries Research Board of
Canada 26, 479^541.
Ricker,W.E. (1975) Computation and interpretation of biological statistics of sh populations. Bulletin of the of the
Fisheries Research Board of Canada191,1^382.
Riska, B. (1986) Some models for development, growth, and
morphometric correlation. Evolution 46,1303^1311.
Riska, B., Atchley, W.R. and Rutledge, J.J. (1984) A geneticanalysis of targeted growth in mice. Genetics107,79^101.
Robinson, B.W. and Wardrop, S.L. (2002) Experimentally
manipulated growth rate in threespine sticklebacks:
assessing trade os with developmental stability. Environmental Biology of Fishes 63,67^78.
Ro, D.A. (1992) The Evolution of Life Histories: Theory and
Analysis. Chapman & Hall, London.
Rowe, D.K. and Thorpe, J.E. (1990) Suppression of maturation in male Atlantic salmon (Salmo salar L.) parr by
reduction in feeding and growth during spring months.
Aquaculture 86, 291^313.
Rowe, D.K., Thorpe, J.E. and Shanks, A.M. (1991) Role of fat
stores in the maturation of male Atlantic salmon (Salmo
salar L.) parr. Canadian Journal of Fisheries and Aquatic
Sciences 48, 405^413.
Rozin, P.N. and Mayer, J. (1961) Regulation of food intake
in the goldsh. American Journal of Physiology 201,
968^974.
Rueda, F.M., Martinez, F.J., Zamora, S., Kentouri, M. and
Divanach, P. (1998) Eect of fasting and refeeding on
growth and body composition of red porgy, Pagrus pagrus
L. Aquaculture Research 29, 447^452.
Ruohonen, K. and Makinen,T. (1992) Validation of rainbow
trout (Oncorhynchus mykiss) growth model under Finnish circumstances. Aquaculture105,353^362.
Ruohonen, K., Kettunen, J. and King, J. (2001) Experimental
design in feeding experiments. In: Feed Intake in Fish
(eds D. Houlihan, T. Boujard and M. Jobling), Blackwell
Scientic, Oxford, pp. 88^107.
Russell, N.R. (1991) The dynamics of appetite and growth control in the minnow (Phoxinus phoxinus L.). PhD Thesis,
University of Wales, 224 pages.
Russell, N.R. and Wootton, R.J. (1992) Appetite and growth
compensation in the European minnow, Phoxinus phoxinus (Cyprinidae) following short term of food restriction.
Environmental Biology of Fishes 34, 277^285.
Russell, N.R. and Wootton, R.J. (1993) Satiation, digestive
tract evacuation and return of appetite in the European
minnow, Phoxinus phoxinus (Cyprinidae) following short
periods of pre-prandial starvation. EnvironmentalBiology
of Fishes 38,385^390.
Ryan,W.J. (1990) Compensatory growth in cattle and sheep.
Nutritional Abstract Review of Series B 60,653^664.
Saether, B.-S. and Jobling, M. (1999) The eects of ration level
on feed intake and growth, and compensatory growth

188

after restricted feeding, in turbot Scophthalmus maximus


L. Aquaculture Research 30,647^653.
Sanchez-Vazquez, F.J., Yamamoto, T., Akiyama, T., Madrid,
J.A. and Tabata, M. (1998) Selection of macronutrients
by goldsh operating self-feeders. Physiology and Behaviour 65, 211^218.
Sanchez-Vazquez, F.J., Yamamoto, T., Akiyama, T., Madrid,
J.A. and Tabata, M. (1999) Macronutrient self-selection
through demand-feeders in rainbow trout. Physiology
and Behaviour 66, 45^51.
Satoh, S., Takeuchi, T. and Watanabe, T. (1984) Studies on
nutritive-value of dietary lipid in sh. 28. Eects of starvation and environmental-temperature on proximate
and fatty-acids compositions of Tilapia nilotica. Bulletin
of the of theJapanese Society of Scientic Fisheries Research
50,79^84.
Saunders, R.L., Farrell, A.P. and Knox, D.E. (1992) Progression of coronary arterial lesions in Atlantic
salmon (Salmo salar) as a function of growth rate. Canadian Journal of Fisheries and Aquatic Sciences 49,
878^884.
Schultz, E.T. and Conover, D.O. (1999) The allometry of
energy reserve depletion: test of a mechanism for sizedependent winter mortality. Oecologia119, 474^483.
Schultz, E.T., Lankford, T.E. and Conover, D.O. (2002) The
covariance of routine and compensatory juvenile
growth rates over a seasonality gradient in a coastal sh.
Oecologia133,501^509.
Schwarz, F.J., Plank, J. and Kirchgessner, M. (1985) Eects of
protein or energy restriction with subsequent realimentation on performance of carp (Cyprinus carpio L.). Aquaculture 48, 23^33.
Sibly, R.M. and Calow, P. (1986) Physiological Ecology of
Animals. Blackwell Scientic, Oxford.
Siems, D.P. and Sikes, R.S. (1998) Tradeos between growth
and reproduction in response to temporal variation in
food supply. Environmental Biology of Fishes 53,319^329.
Silverstein, J.T. and Plisetskaya, E.M. (2000) The eects of
NPY and insulin on food regulation in sh. American
Zoologist 40, 296^308.
Singh, R.P. and Srivastava, A.K. (1985) Satiation time,
gastric evacuation and appetite revival in Heteropneustes
fossilis (Block) (Siluriformes: Pisces). Aquaculture 49,
307^313.
Skilbrei, O.T. (1990) Compensatory sea growth of male
Atlantic salmon, Salmo salar L. which previously mature
as parr. Journal of Fish Biology 37, 425^435.
Smith, R.R. (1987) Methods of controlling growth of steelhead. Progressive Fish Culturist 49, 248^252.
Sogard, S.M. (1997) Size-selective mortality in the juvenile
stage of teleost shes: a review. Bulletin of the of Marine
Science 60,1129^1157.
Sogard, S.M. and Olla, B.L. (2002) Contrasts in the capacity
and underlying mechanisms for compensatory growth
in two pelagic marine shes. Marine Ecology: Progress
Series 243,165^177.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

Compensatory growth in shes M Ali et al.

Solomon, D.J. and Braeld, A.E. (1972) The energetics of feeding, metabolism and growth of perch (Perca uviatilis L.).
Journal of Animal Ecology 41,699^718.
Speare, D.J. and Arsenault, G.J. (1997) Eects of intermittent
hydrogen peroxide exposure on growth and columnaris
disease prevention of juvenile rainbow trout (Oncorhynchus mykiss). CanadianJournal of Fisheries and Aquatic
Science 54, 2653^2658.
Stirling, H.P. (1976) Eect of experimental feeding and starvation on the proximate composition of the European
bass, Dicentrachus labrax. Marine Biology 34,85^91.
Swain, D.P. (1992) Selective predation for vertebral phenotype in Gasterosteus aculeatus: reversal in the direction
of selection at dierent larval sizes. Evolution 46, 998^
1013.
Takagi, Y. (2001) Eects of starvation and subsequent refeeding on formation and resorption of acellular bone
in tilapia, Oreochromis niloticus. Zoological Science 18,
623^629.
Talbot, C. (1985) Laboratory methods in sh feeding and
nutritional studies. In: Fish Energetics New Perspectives
(eds P. Tytler and P. Calow), London: Croom-Helm, pp.
125^154.
Talbot, C., Higgins, P.J. and Shanks, A.M. (1984) Eects of
pre-and post-prandial starvation on meal size and evacuation rate of juvenile Atlantic salmon, Salmo salar L.
Journal of Fish Biology 25,551^560.
Tandon, K.K. and Johal, M.S. (1983a) Study on age and
growth of Tor putitora (Hamilton) as evidenced by scales.
Indian Journal of Fisheries 30,171^175.
Tandon, K.K. and Johal, M.S. (1983b) Occurrence of the phenomenon of growth compensation in Indian major
carps. Indian Journal of Fisheries 30,180^182.
Tanner, J.M. (1963) Regulation of growth in size in mammals. Nature199,845^850.
Tashima, L. and Cahill, G.F. (1968) Eects of insulin in the
toadsh, Opsanus tau. General and Comparative Endocrinology11, 262^271.
Taylor, E.B. and McPhail, J.D. (1985) Burst swimming and
size-related predation of newly emerged coho samon
Oncorhynchus kisutch. Transactions of the American Fisheries Society114,546^551.
Teskeredzic, Z., Teskeredzic, E., Tomec, M., Hacmanjek, M.
and Mclean, E. (1995) The impact of restricted rationing
upon growth food conversion eciency and body-composition of rainbow-trout. Water Science and Technology
31, 219^223.
Thorpe, J.E. (1986) Age at rst maturity in Atlantic Salmon,
Salmo salar: freshwater period inuences and conicts
with smolting. Canadian Special Publication of the of Fisheries and Aquatic Science 89,7^14.
Thorpe, J.E., Miles, M.S. and Keay, D.S. (1984) Developmental
rate, fecundity and egg size in Atlantic salmon, Salmo
salar L. Aquaculture 43, 289^305.
Thorpe, J.E.,Talbot, C., Miles, M.S. and Keay, D.S. (1990) Control of maturation in cultured Atlantic salmon, Salmo

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

salar, in pumped seawater tanks by restricted food intake.


Aquaculture 86,315^326.
Toledo, M.M. (1996) Ciclos de vida y estrategias reproductivas
de la trucha comun (Salmo trutta L.). PhD Thesis, University of Oviedo, 203 pages.
Toneys, M.L. and Coble, D.W. (1979) Size-related, rst winter
mortality of freshwater shes. Transactions of the American Fisheries Society108, 415^419.
Travis, J. (1989) The role of optimizing selection in natural
populations. Annual Review of Ecology and Systematics
20, 279^296.
Tyler, A.V. and Dunn, R.S. (1976) Ration, growth and measures of somatic and organ condition in relation to the
meal frequency in winter ounder, Pseudopleuronectes
americanus, with hypotheses regarding population
homeostasis. Journal of Fisheries Research Board of Canada
33,63^75.
Underwood, A.J. (1997) Experiments in Ecology ^ Their Logical
Design and Analysis Using Analysis Of Variance. Cambridge University Press, Cambridge.
Vera-Cruz, E.M. and Mair, G.C. (1994) Conditions for eective
androgen sex reversal in Oreochromis niloticus (L.). Aquaculture122, 237^248.
Wainwright, P.C. (1987) Biomechanical limits to ecological
performance: molluscs-crushing by the Caribbean hogsh, Lachnolaimus maximus (Labridae). Journal of Zoology
213, 283^297.
Wainwright, P.C. and Richard, B.A. (1995) Predicting patterns of prey use from morphology of shes. Environmental Biology of Fishes 44,97^113.
Wang, Y., Cui, Y., Yang, Y.X. and Cai, F.S. (2000) Compensatory growth in hybrid tilapia, Oreochromis mossambicus
 O. niloticus, reared in seawater. Aquaculture 189,
101^108.
Weatherley, A.H. and Gill, H.S. (1981) Recovery growth following periods of restricted rations and starvation in
rainbow trout, Salmo gairdneri Richardson. Journal of
Fish Biology18,195^208.
Weatherley, A.H. and Gill, H.S. (1987) The Biology of Fish
Growth. Academic Press, London.
Webb, P.W., LaLiberte, G.D. and Schrank, A.J. (1996) Does
body and n form aect maneuverability of sh traversing vertical and horizontal slits? Environmental Biology
of Fishes 46,7^14.
Werner, E.E. and Anholt, B.R. (1993) Ecological consequences of the trade-o between growth and mortality
rates mediated by foraging activity. American Naturalist
142, 242^272.
Westerman, M.E. and Holt, G.S. (1988) The RNA-DNA
ratio: measurement of nucleic acids in larval Sciaenops
ocellatus. Contributions in Marine Science 30 (Suppl.),
117^124.
Whitledge, G.W., Hayward, R.S., Nolte, R.B. and Wang, N.
(1998) Testing bioenergetics models under feeding
regimes that elicit compensatory growth. Transactions of
theAmerican Fisheries Society127,740^746.

189

Compensatory growth in shes M Ali et al.

Wieser, W. (1991) Limitations of energy acquisition and


energy use in small poikilotherms: evolutionary implications. Functional Ecology 5, 234^240.
Wieser,W., Krumschnabel, G. and Ojwang-Okwor, J.P. (1992)
The energetics of starvation and growth after refeeding
in juveniles of three cyprinid species. EnvironmentalBiology of Fishes 33,63^71.
Wilson, P.N. and Osbourn, D.F. (1960) Compensatory growth
after undernutrition in mammals and birds. Biological
Review 35,324^363.
Wootton, R.J. (1982) Environmental factors in sh reproduction. In: Reproductive Physiology of Fish (eds C.J.J. Richter
and H.J.T. Goos), Pudoc,Wageningen, pp. 201^219.
Wootton, R.J. (1998) Ecology of Teleost Fishes, 2nd edn.
Kluwer, Dordrecht.
Wootton, R.J., Allen, J.R.M. and Cole, S.J. (1980) Eect of body
weight and temperature on the maximum daily food
consumption of Gasterosteus aculeatus L. and Phoxinus
phoxinus (L.): selecting an appropriate model. Journal of
Fish Biology17,695^705.
Wu, L., Xie, S., Zhu, X., Cui,Y. and Wootton, R.J. (2002) Feeding dynamics in sh experiencing cycles of feed deprivation: a comparison of four species. Aquaculture Research
33, 481^489.
Wu, L., Xie, S., Cui, Y. and Wootton, R.J. (2003) Eects of
cycles of feed deprivation on growth and food consumption of immature three-spined sticklebacks and European minnows. Journal of Fish Biology 62,184^194.
Xie, S., Zhu, X., Cui, Y., Lei, W., Yang, Y. and Wootton, R.J.
(2001) Compensatory growth in the gibel carp following

190

feed deprivation: temporal patterns in growth, nutrient


deposition, feed intake and body composition. Journal of
Fish Biology 58,999^1009.
Zamal, H. and Ollevier, F. (1995) Eect of feeding and lack of
food on the growth, gross biochemical and fatty acid
composition of juvenile catsh. Journal of Fish Biology
46, 404^414.
Zhu, X., Cui,Y., Ali, M. and Wootton, R.J. (2001) Comparison
of compensatory growth responses of juvenile threespined stickleback and minnow following similar food
deprivation protocols. Journal of Fish Biology 58, 1149^
1165.
Zhu, X.,Wu, L., Cui,Y.,Yang,Y. andWootton, R.J. (2003) Compensatory growth in three-spined stickleback in relation
to feed-deprivation protocols. Journal of Fish Biology 62,
195^205.
Zivkov, M.T. (1982) On the eect and nature of growth compensation of sh.Vestnik Ceskoslovenske Spolecnosti Zoologicke 46,142^160.
Zivkov, M.T. (1996) Critique of proportional hypotheses and
methods for back-calculation of sh growth. Environmental Biology of Fishes 46,309^320.
Zivkov, M.T. and Raikova-Petrova, G.N. (1991) Growth compensation of pike-perch, Stizostedion lucioperca (L.), in
the Batak and Ovcharitsa dams. Comptes Rendus de lAcadamie Bulgare des Science 44, 45^58.
Zivkov, M.T., Trichkova, T.A. and Raikova-Petrova, G.N.
(1999) Biological reasons for the unsuitability of growth
parameters and indices for comparing sh growth.
Environmental Biology of Fishes 54,67^76.

# 2003 Blackwell Publishing Ltd, F I S H and F I S H E R I E S, 4, 147^190

You might also like