Content uploaded by Erin E Reardon
Author content
All content in this area was uploaded by Erin E Reardon on May 02, 2016
Content may be subject to copyright.
Volume 4 • 2016 10.1093/conphys/cow012
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/),
which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
© The Author 2016. Published by Oxford University Press and the Society for Experimental Biology. 1
Research article
Themed Issue Article: Conservation Physiology of Marine Fishes
A new analysis of hypoxia tolerance in fishes using
a database of critical oxygen level (Pcrit)
Nicholas J. Rogers1, Mauricio A. Urbina1,†, Erin E. Reardon1, David J. McKenzie2 and Rod W. Wilson1,*
1Biosciences, College of Life and Environmental Sciences, Georey Pope Building, University of Exeter, Stocker Road, Exeter EX4 4QD, UK
2Centre for Marine Biodiversity Exploitation and Conservation (Marbec), UMR 9190 CNRS-Université Montpellier-Ifremer-IRD, Université
Montpellier, Place Eugène Bataillon, Montpellier cedex 5 34095, France
*Corresponding author. Biosciences, College of Life and Environmental Sciences, Georey Pope Building, University of Exeter, Stocker Road,
Exeter EX4 4QD, UK. Tel: +44 (0)1392 724652. Email: r.w.wilson@ex.ac.uk
Hypoxia is a common occurrence in aquatic habitats, and it is becoming an increasingly frequent and widespread environ-
mental perturbation, primarily as the result of anthropogenic nutrient enrichment and climate change. An in-depth under-
standing of the hypoxia tolerance of fishes, and how this varies among individuals and species, is required to make accurate
predictions of future ecological impacts and to provide better information for conservation and fisheries management. The
critical oxygen level (Pcrit) has been widely used as a quantifiable trait of hypoxia tolerance. It is defined as the oxygen level
below which the animal can no longer maintain a stable rate of oxygen uptake (oxyregulate) and uptake becomes dependent
on ambient oxygen availability (the animal transitions to oxyconforming). A comprehensive database of Pcrit values, compris-
ing 331 measurements from 96 published studies, covering 151 fish species from 58 families, provides the most extensive and
up-to-date analysis of hypoxia tolerance in teleosts. Methodologies for determining Pcrit are critically examined to evaluate its
usefulness as an indicator of hypoxia tolerance in fishes. Various abiotic and biotic factors that interact with hypoxia are anal-
ysed for their effect on Pcrit, including temperature, CO2, acidification, toxic metals and feeding. Salinity, temperature, body
mass and routine metabolic rate were strongly correlated with Pcrit; 20% of variation in the Pcrit data set was explained by these
four variables. An important methodological issue not previously considered is the inconsistent increase in partial pressure of
CO2 within a closed respirometer during the measurement of Pcrit. Modelling suggests that the final partial pressure of CO2
reached can vary from 650 to 3500 µatm depending on the ambient pH and salinity, with potentially major effects on blood
acid–base balance and Pcrit itself. This database will form part of a widely accessible repository of physiological trait data that
will serve as a resource to facilitate future studies of fish ecology, conservation and management.
Key words: Carbon dioxide, critical oxygen tension, metabolic rate, oxygen and capacity limitation of thermal tolerance,
physiological trait
Editor: Steven Cooke
Received 17 December 2015; Revised 17 March 2016; accepted 19 March 2016
Cite as: Rogers NJ, Urbina MA, Reardon EE, McKenzie DJ, Wilson RW (2016) A new analysis of hypoxia tolerance in shes using a database of
critical oxygen level (Pcrit). Conserv Physiol 4(1): cow012; doi:10.1093/conphys/cow012.
†Present address: Departamento de Zoología, Facultad de Ciencias Naturales y Oceanográcas, Universidad de Concepción, Casilla 160-C,
Concepción, Chile.
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
Introduction
In recent decades, there has been growing concern regarding
the increasingly widespread and frequent occurrence of
hypoxia in aquatic environments, associated with the increased
discovery of hypoxic zones globally (Diaz, 2001; Diaz and
Breitburg, 2009; Zhang et al., 2010). Although periods of
hypoxia can develop naturally in many aquatic systems,
anthropogenic inuences have been shown to be a major
driver of hypoxic events in both freshwater and marine habi-
tats (Friedrich etal., 2014). In particular, eutrophication asso-
ciated with increased anthropogenic nutrient loading of lakes,
rivers and coastal waters leads to blooms of algae and phyto-
plankton, the death of which subsequently fuels microbial res-
piration and the depletion of dissolved oxygen (Smith, 2003).
Hypoxia has been shown to result in losses of biodiversity and
to trigger widespread mortality events (Vaquer-Sunyer and
Duarte, 2008). In the marine environment, more than 400
coastal systems have been reported as eutrophication-associ-
ated ‘dead zones’ (Diaz and Rosenberg, 2008). Global warm-
ing is likely to exacerbate hypoxia in aquatic systems owing to
increased microbial respiration rates and reduced oxygen solu-
bility with increasing water temperatures (McBryan et al.,
2013). In addition, potential modications to oceanic circula-
tion linked to future climate change are predicted to result in
greater stratication and ‘deoxygenation’ of the oceans
(Keeling and Garcia, 2002; Keeling etal., 2009). In summary,
in the future, reduced oxygen concentrations are predicted to
occur more extensively, more frequently and for longer periods
of time (IPCC, 2014). Fish are among the more hypoxia sensi-
tive of aquatic taxa and, as such, the sequential loss of fauna
from aquatic ecosystems during hypoxic events is commonly
initiated by the loss or relocation of sh populations (Vaquer-
Sunyer and Duarte, 2008). Understanding the physiological
responses of individual organisms to environmental stressors,
such as hypoxia, provides a mechanistic link between environ-
mental change and population-level effects, which may be key
to predicting future ecological impacts (Chown, 2012;
Seebacher and Franklin, 2012; Cooke etal., 2013).
Fishes can show various behavioural responses to hypoxia,
such as rising to the surface to breathe the uppermost layer of
water in contact with air, increasing activity to escape the
hypoxic area or decreasing activity to reduce oxygen demand
(Chapman and McKenzie, 2009; Urbina et al., 2011;
Domenici etal., 2012). Beyond these behavioural responses,
shes can engage numerous profound physiological responses,
such as changes in ventilation, cardiac activity and haemoglo-
bin–O2 binding (Richards, 2009). These physiological
responses work primarily to sustain oxygen extraction from
the environment in order to maintain aerobic ATP produc-
tion. This allows the majority of shes to maintain stable oxy-
gen uptake rates across a wide range of ambient partial
pressures of oxygen (
P
O2), a response known as ‘oxyregulation’
(reviewed by Perry et al., 2009). When, however, oxygen
reduces to a threshold below which oxygen uptake rate
cannot be maintained, oxygen uptake declines linearly with a
decrease in ambient
P
O2, a response known as ‘oxyconform-
ing’ (Pörtner and Grieshaber, 1993; Claireaux and Chabot,
2016). This threshold, when oxygen uptake transitions from
regulation to conforming, is referred to as the critical
P
O
2
(Pcrit;
Beamish, 1964; Ultsch etal., 1978). As a measure of whole-
animal oxygen extraction capacity, which varies extensively
across species and among populations, Pcrit is widely used to
describe the degree of hypoxia tolerance in shes (Ultsch
etal., 1978; Chapman etal., 2002; Nilsson etal., 2007a,b;
Mandic etal., 2009; reviewed by Chapman and McKenzie,
2009; Speers-Roesch etal., 2012).
Oxygen, the key variable in Pcrit measurements, is used by
aerobic organisms as an electron acceptor in order to drive the
production of ATP. As such, the rate of oxygen uptake is
widely considered as a proxy for the rate of aerobic metabo-
lism, at least when in a steady state (Brown et al., 2004;
Nelson, 2016). Standard metabolic rate (SMR) is the oxygen
uptake rate of an entirely inactive, post-absorptive sh and
reects its minimal cost of living at a given temperature
(Beamish and Mookherjii, 1964; Chabot etal., 2016). Routine
metabolic rate (RMR) provides a similar estimate of the cost
of living but takes into account energy expended on maintain-
ing posture and making the small movements that are typical
of most shes even when in a quiescent state (McBryan etal.,
2013). In contrast, maximal metabolic rate (MMR) is the
highest rate of oxygen uptake that can be attained in dened
environmental conditions (Clark et al., 2013; Norin and
Clark, 2016). The difference between SMR and MMR is
referred to as aerobic scope and provides for the oxygen
demands of higher functions, such as locomotion, growth,
behaviour and reproduction (Farrell and Richards, 2009;
Claireaux and Chabot, 2016). In the context of this aerobic
hierarchy, several levels of critical
P
O2 are represented in
Figure1. As this conceptual diagram illustrates, MMR is the
2
Conservation Physiology • Volume 4 2016Research article
Figure 1: Diagram illustrating the conceptual idea of the eects of
hypoxia on the standard metabolic rate (SMR), routine metabolic rate
(RMR), maximal metabolic rate (MMR) and aerobic scope (AS) of an
oxyregulator. This and may not apply to species with facultative metabolic
depression below the critical oxygen level (Pcrit). Pcmax is dened as the
critical exeternal oxygen partial pressure at which oxygen supply no
longer meets the maximum demand for oxygen (Portner, 2010).
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
rst rate to become limited as ambient oxygen decreases
(Pcmax), from which point a decline in MMR leads to a reduc-
tion in aerobic scope. Secondly, the Pcrit for RMR is reached,
whereby oxygen supply cannot sustain even minimal levels of
aerobic activity. Finally, the Pcrit for SMR indicates that oxy-
gen supply cannot meet even basic oxygen demands (Pörtner
and Lannig, 2009; Claireaux and Chabot, 2016). Below this
threshold, anaerobiosis or suppression of metabolic rate are
required to sustain life (Richards, 2009). Each of the three
levels of Pcrit may indicate the difference between mortality
and survival. If so, Pcrit may have major implications for the
tness of shes living in environments prone to hypoxia and,
as such, each of these levels can be considered as functional
traits (McGill etal., 2006; Claireaux and Chabot, 2016).
The examination of trait variation across populations and
communities, and its ecological implications, are increasingly
becoming the basis for predicting and potentially mitigating the
effects on biodiversity of environmental change (Chown, 2012).
Such trait-based approaches are facilitated by the collection
and dissemination of trait data. Large-scale multi-trait data-
bases have been compiled for various taxa, including plants
(Kattge etal., 2011), mammals (Jones et al., 2009), marine
polychaetes (Faulwetter et al., 2014) and North American
freshwater shes (Frimpong and Angermeier, 2009). As a quan-
tiable measure of hypoxia tolerance that is measured on indi-
viduals and is applicable at population level, Pcrit is useful for
incorporation into trait-based approaches to the conservation
physiology of shes (Frimpong and Angermeier, 2009).
The eld of sh physiology has generated a large body of
literature on Pcrit, across a wide range of species and in highly
variable abiotic and biotic conditions (Perry et al., 2009).
Owing to the discrete and nuanced nature of each study, it is
challenging to make broad generalizations. The aims of the
present work were as follows: (i)to assemble a database of the
Pcrit values reported for shes, from published literature, in a
format suitable for future incorporation into multi-trait-based
analyses; (ii)to analyse the data to identify how biotic and
abiotic factors (particularly temperature) interact with
hypoxia and affect Pcrit; and (iii)to appraise methodologies
for measuring Pcrit critically, and thereby evaluate its useful-
ness for quantifying hypoxia tolerance in shes. This new
analysis not only provides an opportunity for further quanti-
tative considerations but also serves as a tangible link between
the physiology and the conservation of shes.
Methods
Literature search
The citation and abstract indexes, Scopus® and Web of
Science®, were used to collect relevant peer-reviewed literature.
The literature search was conducted in December 2014 using
the following terms: ‘critical oxygen’, ‘critical PO2’, ‘oxygen
threshold’, ‘Pcrit’, ‘oxyregulate’, ‘oxyconform’ or ‘hypoxia toler-
ance’. Approximately 400 papers from relevant subject areas
were identied. Each of these articles was individually assessed
for relevance based on their title and abstract. Finally, 144
papers were downloaded for a full read of the manuscript. Of
these, only 96 papers reported Pcrit measurements in at least
one sh species.
Database construction
In order to maximize the future usefulness of the database and
to ensure that it fully reects the variation in abiotic/biotic con-
ditions in which Pcrit has previously been measured in shes, it
was necessary to extract multiple parameters from each study.
For each Pcrit entry, 66 columns summarize information on the
species and origin, acclimation parameters, animal character-
istics, experimental method, results, statistical analyses, gen-
eral comments and bibliographic information (Table1). The
database was constructed as a single Microsoft Excel le, with
individual columns for each parameter and rows for each Pcrit
determination in a particular species or treatment group. As
such, a single study may occupy several rows depending on the
number of treatment groups and/or species for which Pcrit is
reported. Values for Pcrit were reported in a variety of different
oxygen units across the literature (millimetres of mercury, torr,
percentage air saturation, milligrams of oxygen per litre and
micromolar), but were converted here to a partial pressure of
oxygen (in kilopascals) based on oxygen solubility values
reported by Green and Carrit (1967) and assuming standard
atmospheric pressure at sea level (760 mmHg), if not other-
wise reported. Likewise, all values of oxygen uptake rate were
converted to milligrams of oxygen per kilogram per hour. To
enable unbiased inter-species comparison, a subset of the full
database was produced, which included only those Pcrit mea-
surements made in shes meeting the following conditions: (i)
in an unfed or post-absorptive state; (ii) undergoing no addi-
tional (to hypoxia) abiotic stressor; and (iii) where tempera-
ture acclimation lasted for >2 days.
Database analysis
The frequency of Pcrit measurements across families and cli-
mate zones was calculated based on the full database.
However, comparisons of Pcrit values were made using the
subset ‘control’ database described above. Based on the lati-
tude of where the studies were conducted, each entry was
labelled as tropical, sub-tropical, temperature or polar.
Analysis of variance was used to test for an effect of climate
zone on Pcrit using the Sidak post hoc test.
Potential inuences of varying respirometry methodologies
and hypoxia exposure methods on Pcrit were explored using
the subset ‘control’ database, in which there are 297 data
points. Similar to the full database, the majority of studies
measured Pcrit using closed static respirometry on individual
sh, where oxygen is reduced via the oxygen consumption of
the sh (n = 202). Where there were sufcient data to compare
methods between respirometry methods within a species, a
Student’s unpaired t-test was used to compare between
groups. It was not possible to test for differences in hypoxia
exposure methods within species because there were insuf-
cient data from at least two methods.
3
Conservation Physiology • Volume 4 2016 Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
Stepwise multiple linear regression analysis was used to
develop a model for predicting Pcrit based on biotic (body
mass, RMR) and abiotic (temperature, salinity) variables.
Earlier analysis detected no signicant within-species effect of
respirometry method (closed or ow through) on Pcrit, and it
was therefore not included in the linear regression model.
Acclimation variables such as temperature, P
O2 and salinity
were not included in this analysis because they were very
highly correlated with the equivalent variables reported dur-
ing the trials. Minimal
P
O
2
was not included in the model
because it is driven by Pcrit.
As the multivariate model identied salinity as a relevant
factor, the potential effect of salinity on Pcrit was explored fur-
ther by comparing Pcrit values measured in seawater (150
entries from 82 species) with Pcrit values measured in freshwa-
ter (116 entries from 50 species). This approach was taken
because most of the studies were conducted either in freshwa-
ter [∼0.1 practical salinity units (PSU)] or seawater (∼30–
38 PSU). Values of Pcrit were calculated as the partial pressure
of oxygen (in kilopascals) and as the concentration of oxygen
(in milligrams per litre), using the solubility coefcient based
on experimental temperature and salinity (Green and Carrit,
1967). Potential differences between groups were then tested
by a Mann–Whitney U-test, because normality assumptions
were violated.
Results and discussion
Database coverage
Of the 96 studies reviewed, 331 measurements of Pcrit across
151 species were incorporated into the database. Across the
global database, 58 families are represented, with Cyprinidae
(n = 44), Pomacentridae (n = 41), Gobiidae (n = 24), Cichildae
(n = 23), Salmonidae (n = 19), Cottidae (n = 18), Apogonidae
(n = 17), Percidae (n = 13) and Sparidae (n = 12) the most fre-
quently represented. Freshwater and marine (including eury-
haline) species account for 40 and 60% of Pcrit entries,
respectively. Water temperatures at which Pcrit values were
determined ranged between −1.5 and 36°C, with a mean
(±SD) of 21.7 ± 7.6°C. Values for Pcrit over the entire data set
ranged between 1.02 kPa (Pseudocrenilabrus multicolor
victoriae; Reardon and Chapman, 2010) and 16.2 kPa (Solea
4
Conservation Physiology • Volume 4 2016Research article
Table 1: List of the parameters incorporated into the database alongside each reported critical oxygen level value
Species and origin Stock
acclimation
Sample
characteristics
Experimental
method Results Statistical
analysis
Comments
and reference
Family Holding time Sample size Respirometry type Oxy regulating or
conforming
Statistical
method
Comments
Genus Acclimation
temperature
Mean mass BMR/RMR/SMR/MMR
M
O
2
Pcrit calculation
method
Reference
Species Acclimation
salinity
Mass SD Determination
method
Critical
P
O
2
SMR
determination
Year
Origin
P
O
2
units Mass SEM Swimming speed Critical
P
O2
range Corresponding
Author
Latitude and longitude Acclimation
P
O
2
Mass range upper Hypoxia method Critical
P
O
2
SD DOI
Acclimation pH Mass range lower Rate of hypoxia onset Critical
P
O2
SEM Full citation
Acclimation time Mean length
P
O
2
set-point time Critical
P
O
2
units
Diet Length SD Minimal
P
O
2
Air breathing
threshold
Energy content Length SEM
P
O
2
unit Common
P
O2
units
Ration unit Length range
upper
Salinity
Ration size Length range
lower
Temperature
Photoperiod
(light:dark)
Life stage pH
Feeding regimen Sex
P
CO2
Last feed Photoperiod
(light:dark)
Access to air
Abbreviations: BMR, basal metabolic rate; DOI, digital object identier; MMR, maximal metabolic rate;
MO2
, oxygen uptake rate;
P
CO2
, partial pressure of carbon diox-
ide; Pcrit, critical oxygen level;
P
O
2
, partial pressure of oxygen; RMR, routine metabolic rate; SMR, standard metabolic rate.
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
solea larvae; McKenzie etal., 2008) with a mean (±SD) Pcrit in
the ‘control’ data set of 5.15 ± 2.21 kPa. Plots of species and
their reported Pcrit values from the subset data set are provided
in the Supplementary Data (Supplementary Fig.1).
The geographical coverage of the database includes at least
one entry from every continent, although North America,
Europe and Australasia are by far the most heavily represented
and, when combined, account for 87% of Pcrit entries. Perhaps
unsurprisingly, most studies of Pcrit in shes have been concen-
trated around the major sh physiology research groups in
Europe, North America and Australia. Arguably, this intro-
duces an element of bias into the database, given the incom-
plete representation of all habitats and species at a global scale.
Based on the full database, tropical studies are the most fre-
quently represented (n = 125 Pcrit measurements, dominated by
Lizard Island Research Station, Australia, n = 98), followed by
subtropical (n = 104) and temperate regions (n = 100), domi-
nated by Canada and Europe. The polar regions are the most
under-represented (n = 2). Within the subset ‘control’ database,
there was a signicant difference in mean Pcrit across climatic
regions (ANOVA, F2,297 = 4.054, P = 0.018), where tropical
shes had the lowest Pcrit (mean ± SEM: 4.92 ± 0.190 kPa) <
sub-tropical fishes (5.0 ± 0.24 kPa) < temperate fishes
(5.74 ± 0.24 kPa). However, the Sidak post hoc test suggested
that Pcrit values for tropical shes were signicantly lower only
than temperate shes (P = 0.021). There was no difference in
mean Pcrit between subtropical and either tropical (P = 0.991)
or temperate Pcrit (P = 0.085). Owing to low sample size, the
polar Pcrit values were not included in the ANOVA across tem-
peratures but, interestingly, had a higher mean Pcrit than the
other three climatic zones (7.9 ± 1.6 kPa).
Additionally, the species studied tend to be those conducive
to respirometry trials. In particular, large, active or highly sen-
sitive species, such as those of the Scombridae family (tuna,
mackerels and bonitos) are generally under-represented in the
literature (Blank etal., 2007). For example, the majority of
Pcrit values reported in the database were measured on sh
<1 kg body mass.
Methodology used to determine critical
oxygen level
The relationship between ambient
P
O2 and oxygen uptake in
shes has been investigated since the study of Keys (1930).
Even at that early stage, there was considerable discussion
among physiologists regarding the validity of different meth-
odologies. Technological developments, particularly methods
for measuring dissolved oxygen content such as galvanic oxy-
gen electrodes and, more recently, bre-optic sensors, have
made the performance of high-resolution measurements of
oxygen uptake in shes increasingly common (Clark et al.,
2013; Nelson, 2016). Nevertheless, the literature examined
for the purpose of building this database is characterized by
considerable variation in terms of methods used to determine
Pcrit. For example, the majority of studies (56%) used closed
respirometry for Pcrit estimates, 21% used ow-through respi-
rometry, 20% used intermittent respirometry, and 3% used
other approaches, such as indirect estimation of gill oxygen
uptake (Table2). Most studies (70%) depleted ambient oxy-
gen through the sh’s own respiration, whereas 30% of stud-
ies bubbled nitrogen gas into the water to reduce ambient
oxygen levels. The majority of studies (80%) measured RMR
for Pcrit estimates; the remaining 20% measured SMR. These
methodological differences and their implications are impor-
tant to consider when interpreting collated Pcrit data.
Closed respirometry, whereby the sh is placed within a
sealed chamber from which water is intermittently sampled
for measurement of dissolved oxygen content, provides the
simplest method of measuring oxygen uptake rate (Steffensen,
1989), as follows:
MVVO t
Or
f
2
bw=−×÷×[( )](),∆∆
2
where
M
O
2
represents oxygen uptake rate, Vr is respirometer
volume, Vf is sh volume, ΔO2 is change in ambient oxygen con-
tent, t is time, and bw is sh mass (‘body weight’). Importantly,
water needs to be recirculated within the chamber to ensure
adequate mixing, thus preventing the stratication of dissolved
5
Conservation Physiology • Volume 4 2016 Research article
Table 2: The breakdown of the number of data points representing each respirometry type and oxygen removal method in the subset database
Oxygen depletion method
Respirometry type Fish respiration N2 equilibration N2 and O2
equilibration
N2 and CO2
equilibration
N2, O2 and air
equilibration Total
Closed static (individual) 202 1 0 0 0 203
Closed static (grouped) 13 0 0 0 0 13
Closed ow-through (individual) 13 14 0 0 3 30
Intermittent ow (individual) 13 26 2 1 0 42
Mesocosm (grouped, large tuna) 0 1 0 0 0 1
Open ow-through (grouped) 7 0 0 0 0 7
Opercular mask 1 0 0 0 0 1
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
oxygen within the chamber (Keys, 1930). Whether spontaneous
movements and ventilation are sufcient to provide mixing
depends on the species and achieving the correct sh-to-respi-
rometer volume ratio. For closed determinations of Pcrit, hypoxia
is generated by allowing the sh to deplete available oxygen
through its own respiration, therefore negating the need to strip
dissolved oxygen from the water articially through equilibra-
tion with nitrogen. For this reason, closed respirometry is par-
ticularly useful for conducting measurements of Pcrit in the eld
or at remote locations where facilities such as a supply of N2 may
not be readily available (Rosenberger and Chapman, 2000;
Nilsson etal., 2007b).
However, there are several important considerations
regarding the use of closed respirometry for determination of
Pcrit. For instance, the rate of oxygen depletion during closed
respirometry is determined by the ratio of sh size (or oxygen
uptake rate) to respirometer volume. A lack of control over
the development of hypoxia can be problematic in compara-
tive studies that use the same respirometer to measure Pcrit in
sh of different size and/or metabolic rate. As an illustrative
example, the depletion of oxygen levels from 20 to 1 kPa by
Australian barramundi (Lates calcarifer) took between 1.5
and 4 h depending on the temperature (26 or 36°C; Collins
etal., 2013). From our database, it is evident that there is very
little, if any, standardization in terms of the rate of oxygen
depletion between Pcrit studies, irrespective of which respirom-
etry method is employed. This is in contrast to measurements
of other physiological threshold traits, such as the determina-
tion of critical temperature, which tends to be made at consis-
tent warming or cooling rates among studies (0.2–0.3°C min−1;
Beitinger etal., 2000; Mora and Maya, 2006; Murchie etal.,
2011). It is unclear whether the rate of decline in ambient
oxygen will signicantly affect Pcrit, but it is likely that a lon-
ger time scale would allow for greater respiratory adjust-
ments, and hence, reveal lower Pcrit values than more acute
hypoxic exposures. Indeed, our own anecdotal observations
in European ounder (Platichthys esus) suggest that these
sh tend to oxyconform across the entire range of ambient
P
O
2
when exposed to a very rapid reduction of oxygen (from 21 to
2 kPa in <2 h).
A further issue associated with closed respirometry is the
build-up of the waste products of metabolism, in particular
CO2 (Keys, 1930; Steffensen 1989; Urbina etal., 2012). It has
been argued that the level of CO2 accumulation within a
closed respirometer is unlikely to impact on CO2 excretion by
shes signicantly, given that they normally exhibit a blood
partial pressure of CO2 (
P
CO
2
) of around 2–4 mmHg, much
higher than normal ambient levels (Ishimatsu etal., 2005;
Nilsson et al., 2007a). However, a precedent has been set,
albeit at more severe levels of hypercarbia (2.25–20 mmHg),
to show that elevated
P
CO
2
can increase Pcrit in European eels
(Anguilla anguilla; Cruz-Neto and Steffensen, 1997), although
no effect on Pcrit was observed when eels were given enough
time to acclimate fully in terms of acid–base regulation
(McKenzie et al., 2003), or in spot sh (Leiostomus xan-
thurus) and mummichog (Fundulus heteroclitus; Cochran and
Burnett, 1996). Given the potential inuence of hypercarbia,
it would be prudent to report any change in water
P
CO2 along-
side values for Pcrit that have been determined through closed
respirometry, but this has rarely been the case throughout the
existing literature. A single study so far has evaluated this
potential confounding factor in determining Pcrit, but in this
unusual oxyconforming species (inanga, Galaxias maculatus)
elevated
P
CO2 had no effect on oxygen uptake rate at any level
of ambient oxygen (Urbina etal., 2012). Furthermore, the
authors pointed out that the effect of CO2 on
M
O2 in shes
appears to be species specic (Gilmour, 2001; Ishimatsu etal.,
2008).
An important issue that does not appear to have been con-
sidered previously is that the extent to which
P
CO2 increases
within a closed respirometer will be highly dependent on the
starting water chemistry, in particular pH and salinity (Fig.2).
A higher seawater pH indicates a greater total alkalinity (TA).
In turn, this gives increased capacity for buffering added CO2
and limiting the increase in
P
CO
2
for a given increase in total
CO2 attributable to net excretion by the sh in a respirometer.
Therefore, the lower the starting water pH, the larger the
6
Conservation Physiology • Volume 4 2016Research article
Figure 2: Model of the estimated partial pressure of carbon dioxide
(
P
CO
2
) reached, in water of dierent salinities and starting pH values,
after the addition of 140 µM CO2. The value of 140 µM approximates
the increase in total CO2 attributable to excretion by a sh at 15°C
during a closed respirometry experiment. In this theoretical example,
the oxygen level is allowed to decline as a result of respiration from a
normoxic partial pressure of >20 kPa (∼245 µM) to a common Pcrit value
of ∼6 kPa (∼74 µM), and we have assumed a respiratory quotient (CO2
excreted ÷ O2 consumed) of 0.85 for sh (Kieer etal., 1998). At each
starting pH, the total alkalinity (TA) and total CO2 were calculated from
the pH and assuming equilibration with atmospheric
P
CO
2
(395 µatm).
When excreted CO2 is dissolved in water, the total CO2 increases
accordingly (in this case, by 140 µM) but TA remains unchanged
(Riebesell etal., 2010). For each starting pH, we therefore used the
CO2sys program (for the national bureau of standards pH scale) to
calculate the nal
P
CO
2
that would result from increasing total CO2 by
140 µM while TA remained constant. This was repeated for salinities of
20, 25, 30, 35 and 40 practical salinity units (PSU) and starting pH values
of 7.5–8.5 to cover ranges experienced in many marine laboratories.
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
overall change in
P
CO2 over the course of the Pcrit measure-
ment. From the models shown in Figure2, it is clear that pH
has a massive inuence on the ambient
P
CO
2
reached within
such a closed respirometry scenario, with nal
P
CO
2
values
ranging by 5-fold, from ∼650 µatm (0.49 mmHg) to
∼3500 µatm (2.66 mmHg) at the highest (8.5) and lowest
(7.5) starting pH values shown, respectively. Note that even
the lowest of these nal
P
CO
2
values has been shown (in exper-
iments designed to mimic future ‘ocean acidication’ scenar-
ios) to have signicant detrimental effects in shes (Munday
etal., 2009). When the starting pH is low, the highest
P
CO
2
values of ∼3500 µatm occur, which are more than 3.5 times
higher than the ‘business as usual’ for end-of-century global
CO2 projections (representative concentration pathway sce-
nario 8.5; Meinshausen etal., 2011). It is also relevant to note
that salinity has a major modulating effect, in particular
within the middle of the range of starting pH values. For
example, at a starting pH of 8.0, the nal
P
CO
2
will vary from
slightly <1500 µatm (1.14 mmHg) at the highest salinity
(40 PSU) to >2500 µatm (1.90 mmHg) at the lowest salinity
(20 PSU).
The larger ambient
P
CO
2
values indicated above would cer-
tainly be expected to cause signicant blood acid–base distur-
bance during the time scale of a typical closed respirometry
experiment (minutes to hours) and thus have the potential to
inuence Pcrit via alterations in the oxygen binding afnity of
haemoglobin. It is therefore important to recognize this vari-
ability in
P
CO
2
when conducting closed respirometry experi-
ments to determine hypoxia tolerance, and particularly, when
interpreting Pcrit measurements.
Flow-through respirometry is a technique whereby oxygen
content of the inowing (O2,in) and outowing (O2,out)
water is continuously measured at a xed water ow rate
through the respirometer (Fw). By application of the Fick prin-
ciple, oxygen uptake (
M
O
2
) is determined by:
MFOO
Ow
2out
=−÷
(, ,) .
22
in bw
Although ow-through respirometry avoids the accumulation
of metabolites in the chamber, it suffers from problems pri-
marily related to the ‘wash-out’ effect, whereby a signicant
lag can develop between changes in the sh’s real
M
O
2
and
changes in observed O2,out. The degree of wash-out depends
on the dilution factor, which is a function of water mixing,
volume and ow rate (Steffensen, 1989).
Intermittent ow-through respirometry is generally consid-
ered the ideal method of
M
O
2
determination in shes because
it involves none of the problems associated with closed or
ow-through techniques (Steffensen, 1989; Clark etal., 2013).
The term ‘intermittent’ or ‘semi-closed’ in this context refers to
the transitioning between a closed phase for determination of
M
O
2
and a ush phase for restoring O2 to a set level and remov-
ing metabolites from the respirometer. As the equipment and
software for automating ush–recirculation cycles and simul-
taneous data acquisition from multiple chambers have become
more sophisticated and widely available, intermittent
flow-through respirometry has been increasingly used
(Svendsen etal., 2016). However, Pcrit measurements via this
preferred technique account for only 20% of values incorpo-
rated into the present database.
Flow-through techniques allow for the supply of hypoxic
water to the respirometry chamber. This hypoxic water can be
produced by bubbling with N2 via a solenoid valve linked to
an O2 probe (Schurmann and Steffensen, 1997) or by bub-
bling with set gas mixtures of variable O2 and N2 content.
Both methods allow for ner control of the hypoxic exposure
compared with allowing the sh to deplete ambient oxygen
levels dependent on its own
M
O2. Progressive hypoxia can be
generated in a stepwise fashion such that multiple
M
O
2
mea-
surements can be made at a specic
P
O
2
, thereby increasing the
likelihood of determining an
M
O
2
that is representative of true
SMR or RMR (Rantin etal., 1993).
Using the present database, we were able to explore differ-
ences in respirometry methods within three species, Atlantic
salmon (Salmo salar), common carp (Cyprinus carpio) and
Nile tilapia (Oreochromis niloticus), for which the sample size
for at least two methods was greater than n > 2. Between
closed static or closed ow-through respirometers, there was
no difference in Pcrit of common carp (Student’s unpaired
t-test, t = 1.429, d.f. = 6, P = 0.203). Likewise, between closed,
static respirometers (individual sh) and open ow respirom-
etry (with grouped sh), there was no difference in Pcrit in
Atlantic salmon (Student’s unpaired t-test, t = −0.678, d.f. = 8,
P = 0.517). There was no difference in Pcrit between closed,
ow-through or intermittent ow-through respirometry
within Nile tilapia (Student’s unpaired t-test, t = −0.644,
d.f. = 6, P = 0.543). In both Atlantic salmon and common
carp, oxygen levels were reduced by the respiration of the sh,
whereas in Nile tilapia the oxygen was reduced by nitrogen
equilibration. A direct comparison in the shiner perch
(Cymatogaster aggregata) found, however, that Pcrit measured
by intermittent ow-through respirometry was signicantly
lower than that measured by closed respirometry (Snyder
etal., 2016). Thus, more direct comparisons are needed to
investigate whether the two most common methodologies
might provide different estimates of Pcrit.
To determine Pcrit,
M
O
2
is plotted against ambient
P
O
2
in
order to identify the inection point at which
M
O2 transitions
from being independent of ambient oxygen to dependent on
ambient oxygen. Within this procedure, a great deal of subtle
variation exists among studies. Most obvious is the differential
use of SMR or RMR, with the majority (84%) of studies
reporting a Pcrit for RMR. Arguably, the Pcrit exhibited for
RMR is more ecologically relevant, given that this level of
M
O2
is likely to be exhibited most of the time in the eld (Ultsch
etal., 1978; Pörtner, 2010). Indeed, for some highly active spe-
cies, such as salmonids, Pcrit determined during active swim-
ming may be most useful in considering the ecological
implications of hypoxia (Fry, 1957). Activity level may affect
Pcrit in unexpected ways, such as in the Adriatic sturgeon
(Acipenser naccarii), which exhibits a well-developed ability to
7
Conservation Physiology • Volume 4 2016 Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
oxyregulate (Pcrit = 4.9 ± 0.5 kPa) when permitted to swim at a
low sustained speed but oxyconforms across the entire range
of declining ambient oxygen when its activity is restricted in a
static respirometer (McKenzie et al., 2007). Some species
exhibit a relatively high Pcrit for RMR at a
P
O
2
that is well
above the P50 (half of the hemoglobin oxygen binding sites are
saturated with oxygen) of their haemoglobin. In these
instances, Pcrit may indicate a behavioural change and not sim-
ply a physical limitation of oxygen supply (McBryan etal.,
2013).
Of the studies that determine the Pcrit for SMR, the meth-
ods used for quantifying SMR vary considerably. Some stud-
ies use the single lowest
M
O
2
value recorded at normoxia,
whereas others take the average of a set number of the lowest
M
O
2
values (Iversen etal., 2010). More sophisticated and
robust methods involve extrapolating the average
M
O
2
mea-
sured at specied swimming speeds back to zero activity
(Wilson etal., 1994; Cook etal., 2014) or the use of percen-
tiles and frequency distributions to assess all normoxic
M
O
2
data (Dupont-Prinet et al., 2013). As the critical level for
basal metabolism, Pcrit determinations based on SMR should
theoretically reect a true physiological limitation of oxygen
extraction capacity (McBryan etal., 2013), although this may
not be true in species for which metabolic depression below
Pcrit has a facultative component. Given that the Pcrit for RMR
is likely to be encountered at higher
P
O
2
than that for SMR
(Fig.1), intra- or inter-species comparisons among studies
reporting different levels of Pcrit may not be entirely valid.
Whether SMR or RMR measurements are used to reect nor-
moxic
M
O
2
, it is essential that sufcient time is allowed for
the sh to acclimate to the respirometry chamber; otherwise,
apparent reductions in
M
O
2
as hypoxia develops may be an
artefact of increasing habituation rather than true oxycon-
forming (Nilsson etal., 2004).
The method used to establish the point of intersection
between continuous oxyregulation and oxyconforming
M
O
2
data is also inconsistent among studies. The slope of these lines
will determine the Pcrit and vice versa; therefore, determining
which data points should be included within each line is criti-
cal to establishing an accurate estimate of Pcrit (Yeager and
Ultsch, 1989). This can be achieved graphically by tting a
least-squares linear regression through data points that show a
progressive decline in
M
O
2
such that it intersects with a regres-
sion line tted through normoxic
M
O
2
data (Monteiro etal.,
2013). A number of mathematical methods for performing so-
called piece-wise or segmented linear regression analyses are
available, which provide greater robustness to estimates of Pcrit
and are used in the majority of studies incorporated into the
present database (Nickerson etal., 1989; Yeager and Ultsch,
1989; Leiva etal., 2015). These approaches assume that the
response of
M
O
2
to declining
P
O
2
is biphasic and consists of two
entirely linear elements, with an abrupt transition between the
two. Such assumptions are not necessarily met by real-world
data, and indeed, concentration-dependent reaction kinetics
make truly linear relationships between
M
O2 and
P
O2 unlikely
(Marshall et al., 2013). Recent developments in non-linear
regression techniques are now being promoted as a more accu-
rate approach to determining biological thresholds such as Pcrit
(Stinchcombe and Kirkpatrick, 2012; Marshall etal., 2013).
Critical oxygen level as a hypoxia tolerance
trait
A low Pcrit is generally associated with greater hypoxia toler-
ance because it indicates a higher capacity for oxygen extrac-
tion and tissue delivery at low
P
O
2
(Mandic et al., 2009).
Maintaining aerobic metabolism during hypoxia is advanta-
geous because it is up to 30-fold more efcient than anaero-
bic ATP production (per unit substrate consumed) and avoids
accumulation of the deleterious by-products (e.g. H+) of
anaerobic metabolism (Richards, 2009). Hypoxia-induced
physiological modications that increase oxygen extraction
capacity, such as increased gill surface area (Nilsson, 2007)
and haemoglobin –O2 binding (Brix etal., 1999), are observed
in shes that frequently encounter hypoxia, suggesting that
maintaining aerobic metabolism is a primary hypoxia sur-
vival strategy (Mandic etal., 2009). However, when ambient
P
O
2
declines below Pcrit, survival depends on the availability of
substrate for O2-independent ATP production (primarily gly-
colysis) and the ability to reduce metabolic demand
(Richards, 2009).
How long a sh can maintain a balance between ATP
demand and supply below its Pcrit, and thus delay the onset of
cellular dysfunction, necrosis and subsequent death, is a key
component of hypoxia tolerance (Nilsson and Östlund-
Nilsson, 2008; Urbina and Glover, 2012; Speers-Roesch etal.,
2013). Speers-Roesch etal. (2013) showed that Pcrit does not
entirely predict hypoxia tolerance at lower oxygen levels. The
authors used three species of sculpin (Blepsias cirrhosis,
Leptocottus armatushave and Oligocottus maculosus), which
exhibit different Pcrit values (1.76, 1.48 and 1.03 kPa, respec-
tively), and exposed them to hypoxia levels that were 30%
below each of their respective Pcrit values while recording the
time to loss of equilibrium. The loss of equilibrium was con-
sistent between only two of the three species (L. armatushave
and O. maculosus). Similar relative hypoxia exposures in the
epaulette shark (Hemiscyllium ocellatum) and shovelnose ray
(Aptychotrema rostrata) revealed lower lactate accumulation
in epaulette sharks, indicating enhanced metabolic depression
in this species (Speers-Roesch et al., 2012). Furthermore,
Nilsson and Östlund-Nilsson (2008) showed that Pcrit did not
correlate with body mass in juvenile and adult damselsh
(Pomacentridae) ranging between 10 mg and 40 g but that
smaller sh were much less tolerant to hypoxia below Pcrit,
owing to their limited capacity for meeting ATP demand
through anaerobic metabolism. These ndings were further
supported in G. maculatus (Urbina and Glover, 2013). These
results illustrate the benet of considering Pcrit alongside other
methods of determining hypoxia tolerance, such as measure-
ments of tissue-specic lactate accumulation and determina-
tions of the loss of equilibrium of 50% of the sh, in order to
assess overall hypoxia tolerance (Urbina and Glover, 2013;
Speers-Roesch etal., 2013; Claireaux and Chabot, 2016).
8
Conservation Physiology • Volume 4 2016Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
A recent review by Salin etal. (2015) argues that whole-
animal oxygen consumption measurements may provide only
a partial proxy for energy metabolism because of variation,
within and between individuals, in the amount of ATP pro-
duced per molecule of oxygen consumed by mitochondria
(P/O ratio). Environmental factors such as ambient tempera-
ture, food intake and diet composition have been shown both
to increase and to decrease P/O ratios in the mitochondria of
a variety of organisms (Salin etal., 2015). Hence, conclusions
based on oxygen consumption rate alone could lead to mis-
leading conclusions regarding respiratory performance during
environmental changes. To our knowledge, the effect of
hypoxia on P/O ratios in sh has yet to be investigated, and as
such, provides an interesting avenue for further research.
As a hypoxia-tolerance trait, low Pcrit can often, but not
always, indicate an ability to survive in hypoxic water. It does
not consider the use of hypoxia-avoidance strategies, such as
adaptations for emersion, aquatic surface respiration and air
breathing (Chapman and McKenzie, 2009). The inanga
(G. maculatus), which inhabits lowland streams prone to
severe hypoxia, is a rare example of a sh species that appears
to be an entirely obligate oxyconformer and thus demon-
strates no discernible Pcrit (Urbina etal., 2012). Likewise, sev-
eral species of Gymnotiform electric shes from South
America, which inhabit naturally hypoxic oodplain pools,
also appear to be obligate oxyconformers with no Pcrit
(Reardon E. E., personal communication), an observation that
is also anecdotally supported in Brachyhypopmus brevirostris
(Crampton, 1998). In some of these species, such as the
inanga, a lack of scales and a large surface area-to-volume
ratio indicate a high capacity for cutaneous O2 uptake whilst
emersed, and hence, provide a short-term means to escape
aquatic hypoxia (Urbina etal., 2011). The oxygen thresholds
for aquatic surface respiration, air breathing and emergence
were incorporated into the database, but only where they
have been reported alongside Pcrit measurements. Such exam-
ples demonstrate the limitation of Pcrit as a universal and com-
parative measure of hypoxia tolerance between species and
emphasize the benet of multi-trait-based approaches.
Biotic and abiotic interactions
Environmental stressors, such as hypoxia, rarely occur in iso-
lation, and the interaction between stressors is of key concern
in the context of predicting the ecological impacts of future
environmental change (Crain et al., 2008). As a typical
threshold effect, the response of sh to hypoxia is likely to
result in ‘ecological surprises’, whereby seemingly resilient
populations suddenly collapse once a critical threshold is
crossed (McBryan etal., 2013). Additive or synergistic inter-
actions with hypoxia could hasten the arrival of such thresh-
olds, meaning that small environmental shifts could result in
large effects on the performance of a population.
Theoretically, any abiotic or biotic factor that affects either
oxygen supply (cardiorespiratory capacity) or oxygen
demand (metabolic rate) of an individual, and the balance
therein, will have implications for its hypoxia tolerance. As
an indicator of hypoxia tolerance, the effects of a wide range
of abiotic and biotic interactions on Pcrit in sh have been
published (Table3).
The stepwise multiple linear regression found that biotic
(body mass, RMR) and abiotic (temperature, salinity) vari-
ables were highly correlated with Pcrit (see Table4). A signi-
cant regression (F4,1154 = 10.565, P < 0.001) predicted 19.5%
of the variation in the data, based on an adjusted r2 (multiple
linear regression). Predicted Pcrit is equal to 5.689 + 0.047
(salinity) − 0.083(temperature) + 1.931(body mass) + 0.001
(RMR), where salinity is measured in practical salinity units,
temperature in degrees Celsius, body mass in kilograms, and
RMR in milligrams of oxygen per litre. All four variables were
signicant predictors of Pcrit in the full model (Table4).
Temperature is by far the most widely studied abiotic fac-
tor potentially interacting with hypoxia (reported in 30 spe-
cies) and is particularly relevant, given ongoing global climate
change (Ficke etal., 2007; Pörtner, 2010). As ectotherms,
oxygen demand in shes increases in a roughly exponential
manner with temperature (inter-species mean Q10 of 1.83;
Clarke and Johnston, 1999), and the intrinsic link between
temperature and environmental hypoxia has become the
basis of an overarching concept termed ‘oxygen and capacity
limitation of thermal tolerance’ (Pörtner, 2001, 2010).
Essentially, this concept suggests that the thermal tolerance of
ectotherms is dictated by their capacity to meet the oxygen
demands of aerobic metabolism. Increased temperature both
elevates basal oxygen demand (SMR) and reduces oxygen
supply (via its effect on oxygen solubility), whereas hypoxia
reduces the oxygen supply. Hence, temperature and hypoxia
are likely to act synergistically in shes. Within species,
increasing temperature generally results in a higher Pcrit, but
among species, the slope of the relationship between temper-
ature and Pcrit is highly variable (Fig.3). For example, the
Atlantic salmon (S.salar) exhibits a steep linear increase of
Pcrit in comparison to the shallower slope seen in the common
carp (C. carpio) across a similar temperature range (Ott etal.,
1980; Remen etal., 2013). A surprising exception to the gen-
erally positive intra-species correlation between temperature
and Pcrit was observed in four out of six species of darter
(Etheostoma), for which Pcrit was lower at 20 than 10°C
(Ultsch etal., 1978). Variation in the sensitivity of species to
temperature in terms of hypoxia tolerance may arise because
of differences in their potential for thermal acclimation.
Explanations for this variation may include reducing the
metabolic impact of increased temperature or enhancing oxy-
gen extraction capacity (Ott et al., 1980; Pörtner, 2010).
Species exhibit highly contrasting capacities for plastic accli-
mation responses. At opposite ends of this spectrum, crucian
carp (Carassius carassius) can dramatically increase respira-
tory surface area through gill remodelling in response to tem-
perature and hypoxia (Sollid etal., 2005), whereas certain
tropical reef sh species (Ostorhinchus doederleini and
Pomacentrus moluccensis) demonstrate no thermal acclima-
tion ability even over a relatively modest temperature range
(29–32°C; Nilsson etal., 2010).
9
Conservation Physiology • Volume 4 2016 Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
10
Conservation Physiology • Volume 4 2016Research article
Table 3: Summary of biotic and abiotic factors and their interactions with the intra-species critical oxygen level as reported by studies included
in the database
Variable Species Eect on Pcrit Reference
Increasing temperature
Gadus morhua Increase Schurmann and Steensen (1997)
Lates calcarifer Increase Collins etal. (2013)
Scyliorhinus canicula Increase Butler and Taylor (1975)
Salmo salar Increase Barnes etal. (2011)
S. salar Increase Remen etal. (2013)
Dentex dentex Increase Cerezo Valverde et al. (2006)
Tautogolabrus adspersus Increase Corkum and Gamperl (2009)
Gadus ogac Increase Corkum and Gamperl (2009)
Bellapiscis medius Increase Hilton etal. (2008)
Bellapiscis lesleyae Increase Hilton etal. (2008)
Morone saxatilis Increase Lapointe etal. (2014)
Carassius carassius Increase Sollid etal. (2005)
Gobiodon histrio Increase Sørensen et al. (2014)
Gobiodon erythrospilus Increase Sørensen et al. (2014)
Oreochromis niloticus Increase Fernandes and Rantin (1989)
Cyprinus carpio Increase Ott etal. (1980)
Oncorhynchus mykiss Increase Ott etal. (1980)
Pomacentrus moluccensis Increase Nilsson etal. (2010)
Ostorhinchus doederleini Increase Nilsson etal. (2010)
Carassius auratus grandoculis No eect Yamanaka etal. (2013)
Etheostoma boschungi Decrease Ultsch etal. (1978)
Etheostoma fusiforme Decrease Ultsch etal. (1978)
Etheostoma abellare Decrease Ultsch etal. (1978)
Etheostoma rulineatum Decrease Ultsch etal. (1978)
Increasing salinity
Cottus asper Decrease Henriksson etal. (2008)
Leptocottus armatus No eect Henriksson etal. (2008)
Cyprinus carpio Increase De Boeck et al. (2000)
Cyprinodon ariegatus Increase Haney and Nordlie (1997)
Increased
P
CO2
Fundulus heteroclitus No eect Cochran and Burnett (1996)
Leiostomus xanthurus No eect Cochran and Burnett (1996)
Anguilla anguilla Increase Cruz-Neto and Steensen (1997)
Platichthys esus Increase Rogers (2015)
Hypoxic acclimation
Pagrus auratus No eect Cook etal. (2013)
S. salar No eect Remen etal. (2013)
Hemiscyllium ocellatum Decrease Routley etal. (2002)
Spinibarbus sinensis Decrease Dan etal. (2014)
C. auratus Decrease Fu etal. (2011)
Poecilia latipinna Decrease Timmerman and Chapman (2004 a,b)
(Continued)
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
11
Conservation Physiology • Volume 4 2016 Research article
Table 3: continued
Variable Species Eect on Pcrit Reference
Reared in hypoxic environment
Pseudocrenilabrus multicolor Decrease Reardon and Chapman (2010)
Exercise pre-conditioning
C. auratus Decrease Fu etal. (2011)
Fed
Astronotus ocellatus Increase De Boeck et al. (2013)
Oreochromis niloticus Increase Mamun etal. (2013)
Perca uviatilis Increase Thuy etal. (2010)
Fatty acid-enriched diet
Solea solea (larvae) Decrease McKenzie etal. (2008)
S. solea (juveniles) Decrease McKenzie etal. (2008)
Increasing body mass
Hypostomus plecostomus Decrease Perna and Fernandes (1996)
Astronotus ocellatus Decrease Sloman etal. (2006)
Pomacentridae No eect Nilsson and Östlund-Nilsson (2008)
Pre- to post-settlement (larvae)
Chromis atripectoralis Decrease Nilsson etal. (2007a,b)
Pomacentrus amboinensis Decrease Nilsson etal. (2007a,b)
Larvae to juveniles
C. auratus grandoculis Decrease Yamanaka etal. (2013)
Juveniles to adults
Reinhardtius hippoglossoides Decrease Dupont-Prinet etal. (2013)
Increasing brood size
(mouthbrooders) Zoramia fragilis Increase Östlun-Nilsson and Nilsson (2004)
Zoramia leptacantha Increase Östlun-Nilsson and Nilsson (2004)
Mycobacteriosis infection
Morone saxatilis Increase Lapointe etal. (2014)
Acidied water
Salmo gairdneri Increase Ultsch etal. (1980)
Cyprinus carpio Increase Ultsch etal. (1980)
Metal exposure
Brycon amazonicus Increase Monteiro etal. (2013) (Hg2+)
C. carassius Increase Schjolden etal. (2007) (Cu2+)
Perca uviatilis Increase Bilberg etal. (2010) (AgNO3)
P. uviatilis Increase Bilberg etal. (2010) (nano-Ag)
Organophosphate exposure
Oreochromis niloticus Increase Thomaz etal. (2009)
Anaemia
Pagrus auratus Increase Cook etal. (2011)
Abbreviations:
P
CO2
, partial pressure of carbon dioxide; Pcrit, critical oxygen level.
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
Unlike intra-species Pcrit, there is no apparent relationship
between temperature and inter-species Pcrit (Fig.3), suggesting
that evolution may have nullied the thermal sensitivity of
hypoxia tolerance across species. It has been shown that the
difference in RMR between a typical cold-water and warm-
water sh is less than expected, given the thermal sensitivity of
RMR within individual species (intra-species median
Q10 = 2.4; Clarke and Johnston, 1999). In addition, gill surface
area appears to scale in a linear manner with metabolic rate,
implying that natural selection equips shes with the oxygen
extraction capacity required to match demand at higher tem-
peratures (Nilsson and Östlund-Nilsson, 2008). Selective pres-
sures for small gills, such as the osmorespiratory compromise
(Nilsson, 1986; Gonzalez and McDonald, 1992; Urbina and
Glover, 2015), gill parasites and risks associated with gill
injury, are likely to limit respiratory surface area so that oxy-
gen extraction capacity does not exceed that required by a par-
ticular species for survival in its natural range (Nilsson, 2007).
Thus, generalizations regarding hypoxia tolerance across tem-
peratures cannot be established rmly at the inter-species level.
Although salinity has long been recognized as a key envi-
ronmental factor, studies evaluating the effects of salinity on
Pcrit are scarce. A previous study in the euryhaline sheephead
minnow (Ciprinodon variegatus), acclimated to salinities
from freshwater (0 PSU) to hypersaline waters (100 PSU),
showed a marked effect on Pcrit (Haney and Nordlie 1997) as
environmental salinity rose. Inter-specic comparisons in the
database agree with this previous intra-specic nding; that is,
salinity had a signicant inuence on Pcrit, whereby freshwater
12
Conservation Physiology • Volume 4 2016Research article
Table 4: Results of the stepwise linear regression analysis where salinity, body mass, routine metabolic rate (RMR) and temperature had zero-
order r correlations with Pcrit (P < 0.05) and with each other, where values were reported
Zero-order r (n = 159)
Variable Salinity
(psu) Temperature (°C) Body mass (kg) RMR (mg O2 l−1)Pcrit
(kPa) βsr2b
Salinity 0.317 −0.165 0.354 0.279 0.346 0.099 0.047
Temperature 0.366 −0.141 −0.314 0.081 −0.083
Body mass −0.166 0.166 0.242 0.056 1.931
RMR 0.17 0.202 0.032 0.001
Mean 23.54 23.1 0.1 323.84 5.4 Intercept = 4.027
SD 15.36 7.9 0.3 434.04 2.1 Adjusted r2 = 0.195 P < 0.001
Abbreviations: Pcrit, critical oxygen level; RMR, routine metabolic rate. In the full model, all four variables were signicant predictors of Pcrit.
Figure 3: The eect of temperature on inter-species critical oxygen level (Pcrit; black dashed line) and intra-species Pcrit (continuous lines).
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
species (including a few euryhaline species) presented a 23%
lower Pcrit than seawater species (also including a few euryha-
line species; Fig.4A; P ≤ 0.001).
As explained in earlier sections, any factor inuencing the
oxygen demand (metabolic rate) of an individual will be likely
to have implications for its hypoxia tolerance. Given that tele-
ost shes must maintain a tight regulation of their internal
salts and water composition (osmolality), as external salinity
changes or becomes extreme, shes must expend increased
efforts to maintain internal homeostasis (Urbina and Glover,
2015). As many of the mechanisms of osmoregulation involve
the action of ATP-driven pumps (i.e. Na+,K+-ATPase) in order
to pump ions against a concentration gradient, increased costs
of osmoregulation may explain, in part, some of these differ-
ences in Pcrit, at least for intra-specic comparisons. However,
from our database (inter-specic), where more freshwater vs.
seawater species comparison are presented, it is likely that
other mechanisms are explaining differences in Pcrit. Given
that seawater species separated million years ago from a fresh-
water ancestor (actinoptyergians, 300–180 million years ago;
Vega and Wiens, 2012), both fresh- and seawater species have
adapted to their respective environments, and therefore, have
also optimized their energy allocated to osmoregulation.
Thus, the differences in Pcrit found in the present study, rather
than being explained by energy-related/oxygen demand issues,
could be associated with intrinsic characteristics of both
media (freshwater vs. seawater). Owing to differences in size,
organic matter load and stability, hypoxia is much more prev-
alent and common in freshwater than in seawater environ-
ments. As such, the driver for an enhanced hypoxia tolerance
(lower Pcrit) could potentially explain the lower Pcrit found in
freshwater species. A future phylogenetic analysis might con-
tribute to test this hypothesis.
It is also worth noting that the difference found in Pcrit
when presented as the partial pressure of oxygen (in kilopas-
cals) was no longer found when Pcrit was calculated as the
concentration (in milligrams per litre; (Fig.4B; P > 0.05). This
could potentially highlight the importance of working with
partial pressure, because this is what drives diffusion when
considering gases. Alternatively, it could indicate that the oxy-
gen concentration is more relevant when considering Pcrit val-
ues, because it determines the total amount of oxygen that is
potentially available for diffusion as water ows over the gills,
i.e. for the same oxygen uptake, salinity (through its effect on
solubility) will have a big effect on the difference between
inspired and expired
P
O
2
.
The biological processes that consume O2 also produce
CO2; therefore, hypoxia and hypercarbia can often co-occur
in aquatic environments (Ultsch, 1996; Cruz-Neto and
Steffensen, 1997; Gilmour, 2001). Despite this, the interactive
effect of environmental hypercarbia on hypoxia tolerance has
been relatively understudied. As previously discussed
(Table3), there are conicting reports within the available
literature regarding to the effect of hypercarbia on the Pcrit of
shes (Cochran and Burnett, 1996; Cruz-Neto and Steffensen,
1997; McKenzie etal., 2003). The most likely mechanism by
which hypercarbia could negatively impact hypoxia tolerance
is through respiratory acidosis, leading to Bohr/Root effects
on haemoglobin and reduced oxygen transport capacity
(Jensen etal., 1993; Cruz-Neto and Steffensen, 1997). In this
respect, hypercarbia is partly akin to the far more extreme
acidosis that can occur in poorly buffered freshwater environ-
ments subjected to acid precipitation or drainage. Acidication
of the surrounding water by addition of sulphuric acid (water
pH range 7.4–4.0, at constant atmospheric
P
CO2) increases Pcrit
in both rainbow trout (Oncorhynchus mykiss) and common
carp (Cyprinus carpio; Ultsch etal., 1980). The time required
to compensate for acid–base disturbance is highly variable
among species (10–24 h during moderate hypercarbia;
Melzner etal., 2009), and as such, the effect of hypercarbia
and acidication on hypoxia tolerance is likely to be dependent
13
Conservation Physiology • Volume 4 2016 Research article
Figure 4: The eect of environmental salinity on inter-species critical
oxygen level (Pcrit), expressed as partial pressure of oxygen (in
kilopascals; A) and concentration of oxygen (in milligrams per litre; B).
Data are shown as means + SEM, including data from 82 species in
seawater and 50 species in freshwater. *Unpaired t-test, signicant
when P < 0.050.
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
largely on the species in question as well as the severity and
duration of the hypercarbic or acid exposure (Jensen etal.,
1993).
Exposure to toxicants, such as trace metal contamination,
appears to reduce hypoxia tolerance in shes. Specically,
exposure to elevated concentrations of copper (300 µg l−1),
mercury (150 µg l−1) and silver (63 µg l−1) have been demon-
strated to increase Pcrit in various species (Table3). The accu-
mulation of toxic metals on the gills can stimulate the
hypersecretion of mucus, which acts as a barrier to diffusion
of external toxicants into the blood (McDonald and Wood,
1993; Wilson et al., 1994). In addition, some trace metals
cause hyperplasia and hypertrophy of gill epithelia cells that
results in the fusing and thickening of gill lamellae (Schjolden
etal., 2007; Bilberg etal., 2010). As a consequence, respira-
tory function is compromised as a result of reduced diffusion
area and increased diffusion distance (McDonald and Wood,
1993). The organophosphate insecticide trichlorfon has been
shown to increase Pcrit by inducing similar changes in gill mor-
phology as well as by promoting vasoconstriction that reduces
lamellar blood ow in Nile tilapia (Oreochromis niloticus;
Thomaz et al., 2009). These potential interactions between
toxic contaminants and hypoxia in shes are clearly of con-
cern, particularly given that both stressors predominantly
threaten freshwater and coastal marine systems and are there-
fore likely to coincide (McDonald and Wood, 1993; Diaz and
Rosenberg, 2008).
Determinations of Pcrit in shes have almost universally
been made in unfed, post-absorptive individuals which,
although providing a useful basis for comparing absolute
hypoxia tolerance among species and individuals, does not
fully account for the digestive state typical of shes in their
natural setting. An increase in oxygen uptake following inges-
tion of food, termed specic dynamic action (SDA), is required
in order to meet the energetic costs associated with mechani-
cal and biochemical digestion and assimilation (Jobling,
1993). Shortly after a meal, oxygen uptake in sh typically
rises rapidly, reaching a peak two to three times higher than
pre-fed levels within a few hours. The shape and duration of
the SDA is highly dependent on the species in question as well
as the meal size and composition (Secor, 2009). Measurements
of Pcrit in shes undergoing SDA have revealed signicant
increases in Pcrit compared with unfed control shes, showing
that increased aerobic demand during digestion has negative
consequences for hypoxia tolerance (Table 3). In common
perch (Perca uviatilis) force-fed a 5% body mass ration, Pcrit
at 20 h post-feeding was increased by 1.44-fold compared
with sham-fed individuals (Thuy et al., 2010). Likewise,
oscars (Astronotus ocellatus) fasted for 14 days showed a 1.6-
fold lower Pcrit than individuals fed a daily 1% body mass
ration up to 24 h prior to Pcrit determination (De Boeck etal.,
2013). In such experiments, the requirement for a stable
M
O
2
on which to base a determination of Pcrit means that measure-
ments at peak SDA are not feasible, and thus, are likely to
underestimate the effect of digestion on hypoxia tolerance
(Thuy etal., 2010).
Several studies have investigated the effect of hypoxia accli-
mation on Pcrit (Table3). Broadly, short-term physiological
acclimation to hypoxia appears to be achieved through either
enhanced O2 extraction capacity or metabolic depression. In
goldsh (Carassius auratus), 48 h of severe (0.63 kPa) hypoxia
induced dramatic increases in both lamellar surface area and
blood haemoglobin content, leading to a 49% reduction in Pcrit
compared with individuals held at normoxia (Fu etal., 2011).
Likewise, sailn molly (Poecilia latipinna) demonstrated
increased haemoglobin and red blood cell concentrations and
a reduced Pcrit following a 6 week exposure to severe hypoxia
(Timmerman and Chapman, 2004a). Depression of RMR at
normoxia and a subsequent reduction in Pcrit following chronic
hypoxic exposure has been observed in the epaulette shark
(H.ocellatum; Routley etal., 2002) and qingbo (Spinibarbus
sinensis; Dan etal., 2014). However, some less hypoxia-toler-
ant species appear to demonstrate no physiological acclima-
tion potential through hypoxic pre-conditioning. Daily
exposure to 6 h of moderate hypoxia (10.5 kPa) for 33 days
had no effect on Pcrit in post-smolt Atlantic salmon (S. salar;
Remen etal., 2013). Additionally, chronic (6 week) moderate
hypoxia produced no change in the Pcrit of juvenile snapper
(Pagrus auratus; Cook etal., 2013).
As hypoxia is likely to become an increasingly predomi-
nant aquatic perturbation in the future (Vaquer-Sunyer and
Duarte, 2008; Keeling etal., 2009), the degree of physiologi-
cal plasticity for hypoxia tolerance will be a key determinant
of species performance. The potential for long-term and trans-
generational hypoxia acclimation with respect to Pcrit has been
largely unstudied. A transgenerational transfer of hypoxia
tolerance has been demonstrated in zebrash (Danio rerio)
larvae after 2–4 weeks of parental hypoxia exposure, but this
was based on determinations of time to loss of equilibrium
(4 kPa O2) rather than through measurement of Pcrit (Ho and
Burggren, 2012). Reardon and Chapman (2010) demon-
strated a strong element of developmental plasticity in the Pcrit
of the Egyptian mouthbrooder (Pseudocrenilabrus multi-
color) when reared in hypoxic conditions. In addition, intra-
species population effects on Pcrit across habitats of differing
O2 regimens have been observed in several species, indicating
that a high degree of phenotypic plasticity for Pcrit exists
within these populations (Timmerman and Chapman, 2004b;
Reardon and Chapman 2010; Fu etal., 2011).
Future applications
The comprehensive Pcrit database presented here provides the
opportunity for a variety of further analyses with potential to
offer fundamental physiological, as well as wider ecological,
insights. For example, further analyses could involve compar-
ing species Pcrit values within a phylogenetic context as a
means to investigate the evolutionary relationships of hypoxia
tolerance among species (Mandic etal., 2009). Likewise, com-
bining species Pcrit data with information on the spatial distri-
bution of populations would help to rene our understanding
of the ecological relevance of Pcrit as a physiological trait. Such
an analysis would be particularly relevant to predicting the
14
Conservation Physiology • Volume 4 2016Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
impacts on sh populations likely to arise from the increas-
ingly widespread occurrence of hypoxic zones in aquatic envi-
ronments around the globe (Friedrich etal., 2014). Given the
variability found in the reported Pcrit for different sh species,
it is likely that hypoxic events will have consequences that are
very dependent on individual species. This highlights the com-
plexity of predicting the effects that hypoxia will have at
community and ecosystem levels, and the potential for
hypoxia to have differential effects on predator-prey interac-
tions, migrations, and ultimately, global sheries.
The integration of the present database with similar data-
bases of other widely measured physiological parameters in
shes should offer useful insights into interactions among
traits. Such physiological data are of great value for improv-
ing the predictive capacity of models as an aid to the manage-
ment and conservation of aquatic systems (Jørgensen etal.,
2012; Cooke etal., 2013). Traits for which databases are cur-
rently under construction include the metabolic response to
feeding (SDA), aerobic scope, growth rate and critical tem-
perature. On completion, the combined data set will be made
widely accessible via an online data repository facility, such as
that provided by Dryad (http://datadryad.org/). Thus, it is
envisaged that these data will prove to be a tangible link
between the eld of sh physiology and future studies of ecol-
ogy, conservation and management.
Supplementary material
Supplementary material is available at Conservation
Physiology online.
Acknowledgements
The authors wish to thank Silvana Birchenough and Julian
Metcalfe (Cefas, Lowestoft, UK) for their mentoring and
encouragement in the creation of the Pcrit database.
Funding
This work was supported by a Natural Environment Research
Council (NERC, UK) PhD studentship awarded to
N.J.R./R.W.W. and NERC and Biotechnology and Biological
Sciences Research Council (BBSRC) research grants (NE/
H010041/1, BB/D005108/1 and BB/J00913X/1) awarded to
R.W.W. The physiological database is a contribution of the
European Union Cooperation in Science and Technology
(COST) Action (FA1004) on the ‘Conservation Physiology of
Marine Fishes’. The same EU COST Action supported this
work as a Short Term Scientic Mission (STSM). For more
information, see: http://sh-conservation.nu/.
References
Barnes R, King H, Carter CG (2011) Hypoxia tolerance and oxygen regula-
tion in Atlantic salmon, Salmo salar from a Tasmanian population.
Aquaculture 318: 397–401.
Beamish FWH (1964) Respiration of shes with special emphasis on stan-
dard oxygen consumption: II. Inuence of weight and temperature
on respiration of several species. Can J Zool 42: 177–188.
Beamish FWH, Mookerjii PS (1964) Respiration of shes with special emphasis
on standard oxygen consumption: I. Inuence of weight and tempera-
ture on respiration of goldsh, Carassius auraus L. Can J Zool 42: 161–175.
Beitinger T, Bennett W, McCauley R (2000) Temperature tolerances of
North American freshwater shes exposed to dynamic changes in
temperature. Environ Biol Fish 58: 237–275.
Bilberg K, Malte H, Wang T, Baatrup E (2010) Silver nanoparticles and sil-
ver nitrate cause respiratory stress in Eurasian perch (Perca uviatilis).
Aquat Toxicol 96: 159–165.
Blank JM, Morrissette JM, Farwell CJ, Price M, Schallert RJ, Block BA (2007)
Temperature eects on metabolic rate of juvenile pacic bluen tuna
Thunnus orientalis. J Exp Biol 210: 4254–4261.
Brix O, Clements KD, Wells RMG (1999) Haemoglobin components and
oxygen transport in relation to habitat distribution in triplen shes
(Tripterygiidae). J Comp Physiol B 169: 329–334.
Brown JH, Gillooly J, Allen AP, Savage VM, West GB (2004) Toward a meta-
bolic theory of ecology. Ecology 85: 1771–1789.
Butler PJ, Taylor EW (1975) The eect of progressive hypoxia on respira-
tion in the dogsh (Scyliorhinus canicula) at dierent seasonal tem-
peratures. J Exp Biol 63: 117–130.
Cerezo Valverde J, Martínez López F-J, García García B (2006) Oxygen
consumption and ventilatory frequency responses to gradual
hypoxia in common dentex (dentex dentex): Basis for suitable oxygen
level estimations. Aquaculture 256: 542–551.
Chabot D, Steensen JF, Farrell AP (2016) The determination of standard
metabolic rate in shes. J Fish Biol 88: 81–121.
Chapman LJ, McKenzie D (2009) Behavioural responses and ecological
consequences. In Richards JG, Farrell AP, Brauner CJ, eds, Hypoxia in
Fishes. Elsevier, San Diego.
Chapman LJ, Chapman CA, Nordlie FG, Rosenberger AE (2002)
Physiological refugia: swamps, hypoxia tolerance, and maintenance
of sh biodiversity in the Lake Victoria region. Comp Biochem Physiol
A Mol Integr Physiol 133: 421–437.
Chown SL (2012) Trait-based approaches to conservation physiology:
forecasting environmental change risks from the bottom up. Philos
Trans R Soc Lond B Biol Sci 367: 1615–1627.
Claireaux G, Chabot D (2016) Responses by shes to environmental
hypoxia: integration through Fry’s concept of aerobic metabolic
scope. J Fish Biol 88: 232–251.
Clark TD, Sandblom E, Jutfelt F (2013) Aerobic scope measurements of
shes in an era of climate change: respirometry, relevance and rec-
ommendations. J Exp Biol 216: 2771–2782.
Clarke A, Johnston NM (1999) Scaling of metabolic rate with body mass
and temperature in teleost sh. J Anim Ecol 68: 893–905.
Cochran RE, Burnett LE (1996) Respiratory responses of the salt marsh
animals, Fundulus heteroclitus, Leiostomus xanthurus, and
15
Conservation Physiology • Volume 4 2016 Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
Palaemonetes pugio to environmental hypoxia and hypercapnia and
to the organophosphate pesticide, azinphosmethyl. J Exp Mar Biol
Ecol 195: 125–144.
Collins GM, Clark TD, Rummer JL, Carton AG (2013) Hypoxia tolerance is
conserved across genetically distinct sub-populations of an iconic,
tropical Australian teleost (Lates calcarifer). Conserv Physiol 1:
doi:10.1093/conphys/cot029.
Cook DG, Wells RMG, Herbert NA (2011) Anaemia adjusts the aerobic
physiology of snapper (Pagrus auratus) and modulates hypoxia
avoidance behaviour during oxygen choice presentations. J Exp Biol
214: 2927–2934.
Cook DG, Iftikar FI, Baker DW, Hickey AJR, Herbert NA (2013) Low-O2
acclimation shifts the hypoxia avoidance behaviour of snapper
(Pagrus auratus) with only subtle changes in aerobic and anaerobic
function. J Exp Biol 216: 369–378.
Cook DG, Brown EJ, Lefevre S, Domenici P, Steensen JF (2014) The
response of striped surfperch Embiotoca lateralis to progressive
hypoxia: swimming activity, shoal structure, and estimated meta-
bolic expenditure. J Exp Mar Biol Ecol 460: 162–169.
Cooke SJ, Sack L, Franklin CE, Farrell AP, Beardall J, Wikelski M, Chown SL
(2013) What is conservation physiology? Perspectives on an increas-
ingly integrated and essential science. Conserv Physiol 1: doi:10.1093/
conphys/cot001.
Corkum CP, Gamperl AK (2009) Does the ability to metabolically down-
regulate alter the hypoxia tolerance of shes? A comparative study
using cunner (T. adspersus) and greenland cod (G. ogac). J Exp Zool A
Ecol Genet Physiol 311: 231–239.
Crain CM, Kroeker K, Halpern BS (2008) Interactive and cumulative
eects of multiple human stressors in marine systems. Ecol Lett 11:
1304–1315.
Crampton WGR (1998) Eects of anoxia on the distribution, respiratory
strategies and electric signal diversity of gymnotiform shes. J Fish
Biol 53: 307–330.
Cruz-Neto AP, Steensen JF (1997) The eects of acute hypoxia and
hypercapnia on oxygen consumption of the freshwater European
eel. J Fish Biol 50: 759–769.
Dan XM, Yan GJ, Zhang AJ, Cao ZD, Fu SJ (2014) Eects of stable and diel-
cycling hypoxia on hypoxia tolerance, postprandial metabolic
response, and growth performance in juvenile qingbo (Spinibarbus
sinensis). Aquaculture 428–429: 21–28.
De Boeck G, Vlaeminck A, Van Der Linden A, Blust R (2000) Salt stress and
resistance to hypoxic challenges in the common carp (Cyprinus car-
pio L.). J Fish Biol 57: 761–776.
De Boeck G, Wood CM, Iftikar FI, Matey V, Scott GR, Sloman KA, De Nazaré
Paula da Silva M, Almeida-Val VMF, Val AL (2013) Interactions
between hypoxia tolerance and food deprivation in Amazonian
oscars, Astronotus ocellatus. J Exp Biol 216: 4590–4600.
Diaz RJ (2001) Overview of hypoxia around the world. J Environ Qual 30:
275–281.
Diaz RJ, Breitburg DL (2009) Chapter 1 The Hypoxic Environment. In
Jerey G, Richards APF, Colin JB, eds, Fish Physiology, Vol 27. Academic
Press. pp 1–23.
Diaz RJ, Rosenberg R (2008) Spreading dead zones and consequences
for marine ecosystems. Science 321: 926–929.
Domenici P, Herbert NA, LeFrançois C, Steensen JF, McKenzie DJ (2012)
The eect of hypoxia on sh swimming performance and behaviour.
In Palstra AP, Planas JV, eds, Swimming Physiology of Fish. Springer
Verlag, Berlin, pp 129–161.
Dupont-Prinet A, Vagner M, Chabot D, Audet C (2013) Impact of hypoxia
on the metabolism of greenland halibut (Reinhardtius hippoglossoi-
des). Can J Fish Aquat Sci 70: 461–469.
Farrell AP, Richards JG (2009) Chapter 11 Dening Hypoxia: an integrative
synthesis of the responses of sh to hypoxia. In Jerey G, Richards
APF, Colin JB, eds, Fish Physiology, Vol 27. Academic Press, London,
pp 487–503.
Faulwetter S, Markantonatou V, Pavloudi C, Papageorgiou N, Keklikoglou
K, Chatzinikolaou E, Palis E, Chatzigeorgiou G, Vasileiadou K,
Dailianis T et al. (2014) Polytraits: a database on biological traits of
marine polychaetes. Biodivers Data J 2: e1024.
Fernandes MN, Rantin FT (1989) Respiratory responses of Oreochromis
niloticus (Pisces, Cichlidae) to environmental hypoxia under dierent
thermal conditions. J Fish Biol 35: 509–519.
Ficke A, Myrick C, Hansen L (2007) Potential impacts of global climate
change on freshwater sheries. Rev Fish Biol Fisher 17: 581–613.
Friedrich J, Janssen F, Aleynik D, Bange HW, Boltacheva N, Çagatay MN,
Dale AW, Etiope G, Erdem Z, Geraga M et al. (2014) Investigating
hypoxia in aquatic environments: diverse approaches to addressing
a complex phenomenon. Biogeosciences 11: 1215–1259.
Frimpong EA, Angermeier PL (2009) Fishtraits: a database of ecological
and life-history traits of freshwater shes of the United States.
Fisheries 34: 487–493.
Fry FEJ (1957) The aquatic respiration of sh. In Brown M., ed., The
Physiology of Fishes, Vol. I. Academic Press, New York, pp 1–63.
Fu SJ, Brauner CJ, Cao ZD, Richards JG, Peng JL, Dhillon R, Wang YX (2011)
The eect of acclimation to hypoxia and sustained exercise on sub-
sequent hypoxia tolerance and swimming performance in goldsh
(Carassius auratus). J Exp Biol 214: 2080–2088.
Gilmour KM (2001) The CO2/pH ventilatory drive in sh. Comp Biochem
Physiol A Mol Integr Physiol 130: 219–240.
Gonzalez RJ, McDonald G (1992) The relationship between oxygen
consumption and ion loss in a freshwater fish. J Exp Biol 163:
317–332.
Green EJ, Carrit DE (1967) New tables for oxygen saturation of seawater.
J Mar Biol 25: 140–147.
Haney DC, Nordlie FG (1997) Inuence of environmental salinity on rou-
tine metabolic rate and critical oxygen tension of Cyprinodon varie-
gatus. Physiol Zool 70: 511–518.
16
Conservation Physiology • Volume 4 2016Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
Henriksson P, Mandic M, Richards J (2008) The osmorespiratory compro-
mise in sculpins: impaired gas exchange is associated with freshwa-
ter tolerance. Physiol Biochem Zool 81: 310–319.
Hilton Z, Wellenreuther M, Clements KD (2008) Physiology underpins
habitat partitioning in a sympatric sister-species pair of intertidal
shes. Funct Ecol 22: 1108–1117.
Ho DH, Burggren WW (2012) Parental hypoxic exposure confers o-
spring hypoxia resistance in zebrash (Danio rerio). J Exp Biol 215:
4208–4216.
IPCC (2014). Summary for Policymakers. Climate Change 2014: Impacts,
Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. In
Field CB, Barros VR, Dokken DJ, Mach KJ, Mastrandrea MD, Bilir TE,
Chatterjee M, Ebi KL, Estrada YO, Genova RC et al. Contribution of
Working Group II to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge University
Press, Cambridge, UK and New York, NY, USA, pp 1–32.
Ishimatsu A, Hayashi M, Lee K-S, Kikkawa T, Kita J (2005) Physiological
eects on shes in a high-CO2 world. J Geophys Res-Oceans 110:
2156–2202. doi:10.1029/2004JC002564
Ishimatsu A, Hayashi M, Kikkawa T (2008) Fishes in high-CO2, acidied
oceans. Mar Ecol Prog Ser 373: 295–302.
Iversen NK, McKenzie DJ, Malte H, Wang T (2010) Reex bradycardia does
not inuence oxygen consumption during hypoxia in the European
eel (Anguilla anguilla). J Comp Physiol B 180: 495–502.
Jensen FB, Nikinmaa M, Weber RE (1993) Environmental perturbations of
oxygen transport in teleost shes: causes, consequences and com-
pensations. In Rankin JC, Jensen FB, eds, Fish Ecophysiology.
Chapman and Hall, London, pp 162–179.
Jobling M (1993) Bioenergetics: feed intake and energy partitioning. In
Rankin JC, Jensen FB, eds, Fish Ecophysiology. Chapman and Hall,
London, pp. 297–321.
Jones KE, Bielby J, Cardillo M, Fritz SA, O’Dell J, Orme CDL, Sa K, Sechrest
W, Boakes EH, Carbone C et al. (2009) Pantheria: a species-level data-
base of life history, ecology, and geography of extant and recently
extinct mammals. Ecology 90: 2648–2648.
Jørgensen C, Peck MA, Antognarelli F, Azzurro E, Burrows MT, Cheung
WW, Cucco A, Holt RE, Huebert KB, Marras S et al. (2012) Conservation
physiology of marine shes: advancing the predictive capacity of
models. Biol Lett 8: 900–903.
Kattge J, Diaz S, Lavorel S, Prentice IC, Leadley P, Bönisch G, Garnier E,
Westoby M, Reich PB, Wright IJ et al. (2011) TRY – a global database of
plant traits. Glob Change Biol 17: 2905–2935.
Keeling RF, Garcia HE (2002) The change in oceanic O2 inventory asso-
ciated with recent global warming. Proc Natl Acad Sci USA 99:
7848–7853.
Keeling RF, Körtzinger A, Gruber N (2009) Ocean deoxygenation in a
warming world. Annu Rev Mar Sci 2: 199–229.
Keys AB (1930) The relation of the oxygen tension in the external respira-
tory medium to the oxygen consumption of shes. Science 71: 195–196.
Kieer JD, Alsop D, Wood CM (1998) A respirometric analysis of fuel use
during aerobic swimming at dierent temperatures in rainbow trout
(Oncorhynchus mykiss). J Exp Biol 201: 3123–3133.
Lapointe D, Vogelbein WK, Fabrizio MC, Gauthier DT, Brill RW (2014)
Temperature, hypoxia, and mycobacteriosis: eects on adult striped
bass Morone saxatilis metabolic performance. Dis Aquat Organ 108:
113–127.
Leiva FP, Urbina MA, Cumillaf JP, Gebauer P, Paschke K (2015)
Physiological responses of the ghost shrimp Neotrypaea uncinata
(Milne Edwards 1837) (Decapoda: Thalassinidea) to oxygen availabil-
ity and recovery after severe environmental hypoxia. Comp Bioch
Physiol A Mol Integr Physiol 189: 30–37.
McBryan TL, Anttila K, Healy TM, Schulte PM (2013) Responses to tem-
perature and hypoxia as interacting stressors in sh: implications for
adaptation to environmental change. Integr Comp Biol 53: 648–659.
McDonald DG, Wood CM (1993) Branchial mechanisms of acclimation to
metals in freshwater fish. In Rankin JC, Jensen FB, eds, Fish
Ecophysiology. Chapman and Hall, London, pp 297–321.
McGill BJ, Enquist BJ, Weiher E, Westoby M (2006) Rebuilding community
ecology from functional traits. Trends Ecology Evol 21: 178–185.
McKenzie DJ, Dalla Valle AZ, Piccolella M, Taylor EW, Steensen JF (2003)
Tolerance of chronic hypercapnia by the European eel (Anguilla
anguilla). J Exp Biol 206: 1717–1726.
McKenzie DJ, Steensen JF, Korsmeyer K, Whiteley NM, Bronzi P, Taylor
EW (2007) Swimming alters responses to hypoxia in the Adriatic stur-
geon Acipenser naccarii. J Fish Biol 70: 651–658.
McKenzie DJ, Lund I, Pedersen PB (2008) Essential fatty acids inuence
metabolic rate and tolerance of hypoxia in Dover sole (Solea solea)
larvae and juveniles. Mar Biol 154: 1041–1051.
Mamun SM, Focken U, Becker K (2013) A respirometer system to measure
critical and recovery oxygen tensions of sh under simulated diurnal
uctuations in dissolved oxygen. Aquacult Int 21: 31–44.
Mandic M, Todgham AE, Richards JG (2009) Mechanisms and evolution
of hypoxia tolerance in sh. Proc Biol Sci 276: 735–744.
Marshall DJ, Bode M, White CR (2013) Estimating physiological toler-
ances – a comparison of traditional approaches to nonlinear regres-
sion techniques. J Exp Biol 216: 2176–2182.
Meinshausen M, Smith SJ, Calvin K, Daniel JS, Kainuma MLT, Lamarque JF,
Matsumoto K, Montzka SA, Raper SCB, Riahi K et al. (2011) The RCP
greenhouse gas concentrations and their extensions from 1765 to
2300. Clim Change 109: 213–241.
Melzner F, Gutowska MA, Langenbuch M, Dupont S, Lucassen M,
Thorndyke MC, Bleich M, Pörtner HO (2009) Physiological basis for
high CO2 tolerance in marine ectothermic animals: pre-adaptation
through lifestyle and ontogeny? Biogeosciences 6: 2313–2331.
Monteiro DA, Thomaz JM, Rantin FT, Kalinin AL (2013) Cardiorespiratory
responses to graded hypoxia in the neotropical sh matrinxã (Brycon
amazonicus) and traíra (Hoplias malabaricus) after waterborne or tro-
phic exposure to inorganic mercury. Aquat Toxicol 140–141: 346–355.
17
Conservation Physiology • Volume 4 2016 Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
Mora C, Maya MF (2006) Eect of the rate of temperature increase of the
dynamic method on the heat tolerance of shes. J Therm Biol 31:
337–341.
Munday PL, Dixson DL, Donelson JM, Jones GP, Pratchett MS, Devitsina
GV, Døving KB (2009) Ocean acidication impairs olfactory discrimi-
nation and homing ability of a marine sh. Proc Natl Acad Sci USA 106:
1848–1852.
Murchie KJ, Cooke SJ, Danylchuk AJ, Danylchuk SE, Goldberg TL, Suski
CD, Philipp DP (2011) Thermal biology of bonesh (Albula vulpes) in
Bahamian coastal waters and tidal creeks: an integrated laboratory
and eld study. J Therm Biol 36: 38–48.
Nelson JA (2016) Oxygen consumption rate versus rate of energy utilisa-
tion of shes: a comparison and brief history of the two measures. J
Fish Biol 88: 10–25.
Nickerson DM, Facey DE, Grossman GD (1989) Estimating physiological
thresholds with continuous 2-phase regression. Physiol Zool 62:
866–887.
Nilsson GE (2007) Gill remodeling in sh – a new fashion or an ancient
secret? J Exp Biol 210: 2403–2409.
Nilsson GE, Östlund-Nilsson S (2008) Does size matter for hypoxia toler-
ance in sh? Biol Rev 83: 173–189.
Nilsson GE, Hobbs JP, Munday PL, Östlund-Nilsson S (2004) Coward or
braveheart: extreme habitat delity through hypoxia tolerance in a
coral-dwelling goby. J Exp Biol 207: 33–39.
Nilsson GE, Hobbs JPA, Östlund-Nilsson S (2007a) Tribute to P. L. Lutz: respi-
ratory ecophysiology of coral-reef teleosts. J Exp Biol 210: 1673–1686.
Nilsson GE, Östlund-Nilsson S, Penfold R, Grutter AS (2007b) From record
performance to hypoxia tolerance: respiratory transition in damsel-
sh larvae settling on a coral reef. Proc Biol Sci 274: 79–85.
Nilsson GE, Östlund-Nilsson S, Munday PL (2010) Eects of elevated tem-
perature on coral reef shes: loss of hypoxia tolerance and inability
to acclimate. Comp Biochem Physiol A Mol Integr Physiol 156: 389–393.
Nilsson S (1986) Control of gill blood ow. In Nielsson S, Holmgren S, eds,
Fish Physiology: Recent Advances. Croom Helm, London, pp 87–101.
Norin T, Clark TD (2016) Measurement and relevance of maximum meta-
bolic rate in shes. J Fish Biol 88: 122–151.
Östlund-Nilsson S, Nilsson GE (2004) Breathing with a mouth full of eggs:
respiratory consequences of mouthbrooding in cardinalsh. Proc Biol
Sci 271: 1015–1022.
Ott ME, Heisler N, Ultsch GR (1980) A re-evaluation of the relationship
between temperature and the critical oxygen tension in freshwater
shes. Comp Biochem Physiol A Physiol 67: 337–340.
Perna S, Fernandes M (1996) Gill morphometry of the facultative air-breath-
ing loricariid sh, Hypostomus plecostomus (Walbaum) with, special
emphasis on aquatic respiration. Fish Physiol Biochem 15: 213–220.
Perry SF, Jonz MG, Gilmour KM (2009) Chapter 5 Oxygen sensing and the
hypoxic ventilatory response. In Jerey G, Richards APF, Colin JB, eds,
Fish Physiology, Vol 27. Academic Press, pp 193–253.
Pörtner HO (2001) Climate change and temperature-dependent bioge-
ography: oxygen limitation of thermal tolerance in animals.
Naturwissenschaften 88: 137–146.
Pörtner HO (2010) Oxygen- and capacity-limitation of thermal tolerance:
a matrix for integrating climate-related stressor eects in marine
ecosystems. J Exp Biol 213: 881–893.
Pörtner HO, Grieshaber MK (1993) Critical PO2(s) in oxyconforming and
oxyregulating animals: gas exchange, metabolic rate and the mode
of energy production. In Bicudo JEPW, eds, The vertebrate gas trans-
port cascade adaptations to environment and mode of life. CRC Press,
Boca Raton, FL.
Pörtner HO, Lannig G (2009) Chapter 4 Oxygen and capacity limited ther-
mal tolerance. In Jerey G, Richards APF, Colin JB, eds, Fish Physiology,
Vol 27. Academic Press, pp 143–191.
Rantin FT, Glass ML, Kalinin AL, Verzola RMM, Fernandes MN (1993)
Cardio-respiratory responses in two ecologically distinct erythrinids
(Hoplias malabaricus and Hoplias lacerdae) exposed to graded envi-
ronmental hypoxia. Environ Biol Fish 36: 93–97.
Reardon EE, Chapman LJ (2010) Energetics of hypoxia in a mouth-brood-
ing cichlid: evidence for interdemic and developmental eects.
Physiol Biochem Zool 83: 414–423.
Remen M, Oppedal F, Imsland AK, Olsen RE, Torgersen T (2013) Hypoxia
tolerance thresholds for post-smolt Atlantic salmon: dependency
of temperature and hypoxia acclimation. Aquaculture 416–417:
41–47.
Richards JG (2009) Chapter 10 Metabolic and molecular responses of sh
to hypoxia. In Jerey G, Richards APF, Colin JB eds, Fish Physiology, Vol
27. Academic Press, pp 443–485.
Riebesell U, Fabry VJ, Hansson L, Gattuso JP (eds.) (2010) Guide to Best
Practices for Ocean Acidification Research and Data Reporting.
Publications Oce of the European Union, Luxembourg, 260 pp.
Rogers NJ (2015) Chapter 4: Respiratory responses and gut carbonate
production during hypoxia and hypercarbia in the European oun-
der (Platichthys esus). In The Respiratory and Gut Physiology of Fish:
Responses to Environmental Change. PhD Dissertation, University of
Exeter, Exeter, UK, pp 95–139.
Rosenberger AE, Chapman LJ (2000) Respiratory characters of three spe-
cies of haplochromine cichlids: implications for use of wetland refu-
gia. J Fish Biol 57: 483–501.
Routley MH, Nilsson GE, Renshaw GMC (2002) Exposure to hypoxia
primes the respiratory and metabolic responses of the epaulette
shark to progressive hypoxia. Comp Biochem Physiol A Mol Integr
Physiol 131: 313–321.
Salin K, Auer SK, Rey B, Selman C, Metcalfe NB (2015) Variation in the
link between oxygen consumption and ATP production, and its
relevance for animal performance. Proc R Soc B Biol Sci 282:
20151028.
Schjolden J, Sørensen J, Nilsson GE, Poléo ABS (2007) The toxicity of cop-
per to crucian carp (Carassius carassius) in soft water. Sci Total Environ
384: 239–251.
18
Conservation Physiology • Volume 4 2016Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
Schurmann H, Steensen JF (1997) Eects of temperature, hypoxia and
activity on the metabolism of juvenile Atlantic cod. J Fish Biol 50:
1166–1180.
Secor SM (2009) Specic dynamic action: a review of the postprandial
metabolic response. J Comp Physiol B 179: 1–56.
Seebacher F, Franklin CE (2012) Determining environmental causes of
biological eects: the need for a mechanistic physiological dimen-
sion in conservation biology. Philos Trans R Soc Lond B Biol Sci 367:
1607–1614.
Sloman KA, Wood CM, Scott GR, Wood S, Kajimura M, Johannsson OE,
Almeida-Val VMF, Val AL (2006) Tribute to R. G. Boutilier: the eect of
size on the physiological and behavioural responses of oscar,
Astronotus ocellatus, to hypoxia. J Exp Biol 209: 1197–1205.
Smith VH (2003) Eutrophication of freshwater and coastal marine ecosys-
tems – a global problem. Environ Sci Pollut Res 10: 126–139.
Snyder S, Nadler LE, Bayley JS, Svendsen MBS, Johansen JL, Domenici P,
Steensen JF (2016) Eect of closed v. intermittent-ow respirome-
try on hypoxia tolerance in the shiner perch Cymatogaster aggregata.
J Fish Biol 88: 252–264.
Sollid J, Weber RE, Nilsson GE (2005) Temperature alters the respiratory
surface area of crucian carp Carassius carassius and goldsh Carassius
auratus. J Exp Biol 208: 1109–1116.
Sørensen C, Munday PL, Nilsson GE (2014) Aerobic vs. anaerobic scope:
sibling species of sh indicate that temperature dependence of
hypoxia tolerance can predict future survival. Glob Change Biol 20:
724–729.
Speers-Roesch B, Richards JG, Brauner CJ, Farrell AP, Hickey AJ, Wang YS,
Renshaw GM (2012) Hypoxia tolerance in elasmobranchs. I. Critical
oxygen tension as a measure of blood oxygen transport during
hypoxia exposure. J Exp Biol 215: 93–102.
Speers-Roesch B, Mandic M, Groom DJE, Richards JG (2013) Critical oxy-
gen tensions as predictors of hypoxia tolerance and tissue metabolic
responses during hypoxia exposure in shes. J Exp Mar Biol Ecol 449:
239–249.
Steensen JF (1989) Some errors in respirometry of aquatic breathers:
how to avoid and correct for them. Fish Physiol Biochem 6: 49–59.
Stinchcombe JR, Kirkpatrick M (2012) Genetics and evolution of func-
tion-valued traits: understanding environmentally responsive phe-
notypes. Trends Ecology Evol 27: 637–647.
Svendsen MBS, Bushnell PG, Steensen JF (2016) Design and setup of an
intermittent-ow respirometry system for aquatic organisms. J Fish
Biol 88: 26–50.
Thomaz JM, Martins ND, Monteiro DA, Rantin FT, Kalinin AL (2009)
Cardio-respiratory function and oxidative stress biomarkers in nile
tilapia exposed to the organophosphate insecticide trichlorfon
(NEGUVON®). Ecotoxicol Environ Saf 72: 1413–1424.
Thuy NH, Tien LA, Tuyet PN, Huong DTT, Cong NV, Bayley M, Wang T,
Lefevre S (2010) Critical oxygen tension increases during digestion in
the perch Perca uviatilis. J Fish Biol 76: 1025–1031.
Timmerman CM, Chapman LJ (2004a) Behavioral and physiological com-
pensation for chronic hypoxia in the sailn molly (Poecilia latipinna).
Physiol Biochem Zool 77: 601–610.
Timmerman CM, Chapman LJ (2004b) Hypoxia and interdemic variation
in Poecilia latipinna. J Fish Biol 65: 635–650.
Ultsch GR (1996) Gas exchange, hypercarbia and acid-base balance,
paleoecology, and the evolutionary transition from water-breathing
to air-breathing among vertebrates. Palaeogeogr Palaeocl 123: 1–27.
Ultsch GR, Boschung H, Ross MJ (1978) Metabolism, critical oxygen ten-
sion, and habitat selection in darters (Etheostoma). Ecology 59: 99–107.
Ultsch GR, Ott ME, Heisler N (1980) Standard metabolic rate, critical oxy-
gen tension, and aerobic scope for spontaneous activity of trout
(Salmo gairdneri) and carp (Cyprinus carpio) in acidied water. Comp
Biochem Physiol A Mol Integr Physiol 67: 329–335.
Urbina MA, Glover CN (2012) Should I stay or should I go? Physiological,
metabolic and biochemical consequences of voluntary emersion
upon aquatic hypoxia in the scaleless sh Galaxias maculatus. J Comp
Physiol B 182: 1057–1067.
Urbina MA, Glover CN (2013) Relationship between sh size and meta-
bolic rate in the oxyconforming inanga Galaxias maculatus reveals
size-dependent strategies to withstand hypoxia. Physiol Biochem
Zool 86: 740–749.
Urbina MA, Glover CN (2015) Eect of salinity on osmoregulation,
metabolism and nitrogen excretion in the amphidromous sh,
inanga (Galaxias maculatus). J Exp Mar Biol Ecol 473: 7–15.
Urbina MA, Forster ME, Glover CN (2011) Leap of faith: voluntary emer-
sion behaviour and physiological adaptations to aerial exposure in a
non-aestivating freshwater sh in response to aquatic hypoxia.
Physiol Behav 103: 240–247.
Urbina MA, Glover CN, Forster ME (2012) A novel oxyconforming
response in the freshwater sh Galaxias maculatus. Comp Biochem
Physiol A Mol Integr Physiol 161: 301–306.
Vaquer-Sunyer R, Duarte CM (2008) Thresholds of hypoxia for marine
biodiversity. Proc Natl Acad Sci USA 105: 15452–15457.
Vega GC, Wiens JJ (2012) Why are there so few sh in the sea? Proc R Soc
B 283: 1826.
Wilson RW, Bergman HL, Wood CM (1994) Metabolic costs and physio-
logical consequences of acclimation to aluminum in juvenile rain-
bow trout (Oncorhynchus mykiss). 2: Gill morphology, swimming
performance, and aerobic scope. Can J Fish Aquat Sci 51: 536–544.
Yamanaka H, Kohmatsu Y, Yuma M (2007) Dierence in the hypoxia toler-
ance of the round crucian carp and largemouth bass: implications for
physiological refugia in the macrophyte zone. Ichthyol Res 54: 308–312.
Yeager DP, Ultsch GR (1989) Physiological regulation and conformation:
a BASIC program for the determination of critical points. Physiol Zool
62: 888–907.
Zhang JD, Gilbert AJ, Gooday L, Levin S, Naqvi WA, Middelburg JJ,
Scranton M, Ekau E, Peña A, Dewitte B et al. (2010) Natural and
human-induced hypoxia and consequences for coastal areas: syn-
thesis and future development. Biogeosciences 7: 1443–1467.
19
Conservation Physiology • Volume 4 2016 Research article
at University of Exeter on April 29, 2016http://conphys.oxfordjournals.org/Downloaded from
