A Bioenergetic Biomagnification
Model for the Animal Kingdom
A D R I A N M . H . D E B R U Y N * A N D
F R A N K A . P . C . G O B A S
School of Resource & Environmental Management,
Simon Fraser University, Burnaby,
British Columbia, Canada V5A 1S6
Species vary greatly in the degree to which they
accumulate dietary contaminants. Bioenergetic processes
play a key role in chemical uptake and elimination, and
interspecific variation in bioaccumulation can be attributed
in large part to variation in how species feed, digest,
and allocate energy. We present a quantitative treatment
of this relationship for the entire animal kingdom. We
derive a model to predict the biomagnification factor for
nonmetabolizable, slowly eliminated chemicals, BMFmax. We
test the model with observed biomagnification factors
and independently derived bioenergetic parameters for a
heterotherms and homeotherms, vertebrates and inver-
tebrates, adults and juveniles, domestic/laboratory animals
environments. The model successfully predicts species-
ranging from less than 1 in caterpillars to nearly 100 in
some carnivores. In addition, we make novel predictions
of BMFmaxfor several taxa for which no measured
bioaccumulation data are available. Our analysis provides
new insights into the role of ecology in chemical dynamics
across the animal kingdom, providing a general framework
for understanding how characteristics of an organism
and its ecological context influence the degree to which
that organism accumulates chemicals present in its diet.
The 2004 United Nations Environment Programme (UNEP)
persistent organic pollutants (POPs) protocol established
are largely based on bioaccumulation observed in fish and,
hence, are applicable only to fish. Unfortunately, the
variety of organisms (e.g., raptors, pinnipeds, humans), and
it is important to expand the criteria to include these and
other organisms. The need for a broader taxonomic assess-
ment is further demonstrated by observations that certain
substances (e.g., perfluorinated sulfonic acids, ?-hexachlo-
rocyclohexane, endosulfan) can biomagnify in species such
as wolves, seals, and whales to a much greater extent than
the development and application of exposure and risk
have outpaced those for many other classes of organisms.
is a need to develop models that can assess the bioaccu-
mulative nature of commercial substances in a variety of
The purpose of this article is to provide a general
framework for integrating bioenergetic reasoning into the
of organisms. Chemical concentrations in consumers arise
from a complex interplay of processes that promote (e.g.,
gastrointestinal magnification) and counteract (e.g., growth
bioaccumulation. The relevant vital rates, feeding rate,
egestion rate, respiration rate, and growth rate, are highly
variable entitiessany or all can vary greatly with charac-
the diet (e.g., biochemical composition, abundance), or the
environment (e.g., temperature, physical structure). Any
variables that have an important influence on the process
(1). Explicitly linking chemical accumulation models to the
bioenergetic processes underlying bioaccumulation is a
powerful way to do this (2-6). The field of bioenergetics is
the central role of energy provides a common currency with
which to link bioaccumulation models to other aspects of
allows the exploitation of the interrelationships among vital
implausible parameter combinations (4).
This bioenergetically based model is intended to pre-
dict the major patterns of biomagnification as manifesta-
tions of fundamental and universal bioenergetic pro-
cesses. If this can be done, then it should be easier to inter-
pret and investigate the remaining variation due to respira-
tion, metabolic transformation, and other processes. It is
our hope that universal treatments such as the model we
present here will stimulate a more thorough integration of
ecological theory into the analysis of chemical dynamics in
a chemical in a consumer organism achieves a thermody-
namic activity (often measured by the lipid normalized
concentration or fugacity) in excess of that in its diet (7, 8).
or air) via respiratory surfaces. Bioconcentration is often
controlled by passive, reversible diffusion and raises an
organism’s internal chemical activity up to, but not above,
in that it can produce internal concentrations and thermo-
dynamic activities of chemicals well above those in either
the diet or the external environment (7, 8, 10). This process
very high and toxic concentrations in upper-trophic-level
Following the body/gut two-compartment model de-
for the steady-state biomagnification factor (BMF) as the
ratio of chemical fugacities in the consumer’s body (fB) and
* Correspondingauthorphone: 604-268-6813;fax: 604-291-4968;
e-mail: firstname.lastname@example.org or email@example.com.
Environ. Sci. Technol. 2006, 40, 1581-1587
10.1021/es051800i CCC: $33.50
Published on Web 01/27/2006
2006 American Chemical SocietyVOL. 40, NO. 5, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY91581
D represents the fugacity-based transport parameters (mol
(DM), respiratory exchange (DR), fecal egestion (DF), and
and body-to-gut (DBG) chemical transport; and ED is the
gross chemical absorption efficiency from the gut, i.e., the
rate of absorption of chemical from the gut relative to the
total rate of elimination of chemical from the gut [ED )
The ratio DBG/DGBin eq 1 has been assumed to be unity
in fish (8, 10) on the basis of the premise that a molecule
diffusing from a consumer’s gut into its body is assumed to
that this ratio is on the order of 0.50-0.33 (a DGB/DBGratio
of 2-3), reflecting a greater resistance for organism-to-gut
transport than for gut-to-organism transfer (11).
To focus on the model’s ability to assess variation in the
extent of biomagnification among a wide array of consumer
organisms, we consider in our analysis a simplified form of
eq 1, restricted to situations in which metabolic transforma-
insignificant. This applies to many POPs that are non- or
poorly metabolizable by many organisms and that, because
of low solubility in the respiratory medium, are not ac-
cumulated or eliminated to an appreciable extent via
the model such that the role of energetic efficiencies can be
elucidated. Under these simplifying assumptions, the maxi-
mum steady-state BMF of a consumer becomes
where DDhas been replaced by the product of the feeding
rate (GD, m3/day) and the sorptive capacity of the diet (ZD,
mol m-3Pa-1), DFhas been replaced by the product of the
DGhas been replaced by the product of the growth rate (g)
(feeding, growth, egestion) occur in all three terms of eq 2,
are unnecessary, and one need only be concerned with the
relative sorptive capacities of diet, feces, and consumer.
In this formulation, biomagnification is promoted by the
process of digestion and macronutrient absorption (8, 10).
The fecal egestion rate, GF, is less than the feeding rate, GD,
by the consumer. Similarly, the sorptive capacity of feces
(ZF) is typically less than that of fresh dietary material (ZD)
because lipids, with a relatively high sorptive capacity for
hydrophobic chemicals, are efficiently and preferentially
absorbed by most consumers. The ratio GDZDDGB/GFZFDBG
always exceed unity for real consumers and will be greatest
and/or the consumer has a highly efficient digestive system.
Any asymmetry that might exist in gut-to-body vs body-to-
gut transport (i.e., DGB > DBG) will further promote the
biomagnification process. Biomagnification is opposed in
eq 2 by limitations on the kinetics of chemical absorption
(1/ED) and by growth dilution (gZB). Organisms with high
and/or low chemical diffusion rates (low gut temperature,
low gut surface area, high resistance to diffusion) will have
will be maximized.
Prediction of BMFmaxcan be greatly simplified by con-
sidering the relationships among these parameters. The
consumer’s growth rate, g, and egestion rate, GF, are related
to the feeding rate, GD, and the digestion and absorption of
ingested materials. Similarly, the sorptive capacity of fecal
matter, ZF, is a function of the sorptive capacity of the diet,
by the consumer. A general energy budget can be used to
define these relationships as
where I is energy ingestion, L is the sum of fecal and urinary
losses, P is production, and R is respiration. Note that
respiration in this context means energy expenditure, not
gas exchange. The terms in eq 3 are expressed in units of
(g day-1) by an energy-biomass interconversion ratio.
the net assimilation of energy and is sometimes expressed
as IRE, where RE is the net efficiency with which ingested
food energy is digested and assimilated. Digestive efficiency
of the diet and the ability of the consumer to digest and
the reduction in the sorptive capacity of the consumed diet
in terms of an efficiency RZ, which expresses the volume-
weighted sorptive capacity of the digesta relative to that of
the ingested diet [i.e., GFZF) (1 - RZ)GDZD].
eq 3 represents the possible fates of assimilated energy. A
fraction is expended on various forms of respiration: basal
and activity. The remainder is then allocated to production
of somatic and reproductive tissue and secretions such as
silk, mucus, or milk. The ratio of production to assimilated
energy (P/IRE) is referred to as net production efficiency,
energy allocated to production of somatic tissue alone
(excluding production of offspring and secretions) is the net
growth efficiency. If the animal is nonreproductive and has
no important secretions, the net production and net growth
efficiencies are equal. Sources of variation in e include
thermoregulatory strategy (homeothermy vs heterothermy)
high e values because of low food-acquisition costs; 12, 17).
e also declines with age in most species (12, 18) because
juveniles expend relatively little energy on reproductive
determinate growth, net production efficiency decreases as
I - L ) R + P(3)
(DG+ DM+ DR) +(
15829ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 5, 2006
can even reach zero in long-lived species. Growth rates in
adult birds and mammals are often small, but even very low
growth rates can have an influence on contaminant con-
centrations (19, 20).
The total production rate, expressing the combined
offspring, and secretions, can be derived from the feeding
rate, energetic efficiencies, and relative energy density of
consumer and diet as:
where δDand δBare the energy densities of the diet and the
consumer, respectively. Expressing g as a function of both
digestive efficiency and net production efficiency permits
and diet composition (RE) and the effect of consumer life
Expressing g as a rate of total production (rather than
elimination resulting from reproduction and secretion are
included in the bioaccumulation model. In this way, repro-
this approach recognizes that, although parturition is as-
sociated with an immediate loss of contaminant mass, the
chemical fugacity in the organism undergoes no change
(assuming that mother and offspring are at equifugacity).
We can now reformulate our expression for BMFmax in
terms of the bioenergetic efficiencies derived above. Sub-
stituting eq 4 into eq 2 and dividing through by GDZDgives
The second term in the denominator of eq 5, (DBG/DGB)(1 -
RZ), or ?, contains the driving forces for biomagnification.
The first term, (1/ED)REe(δD/δB)(ZB/ZD), or γ, represents
processes that counteract the biomagnification process.
Model Parametrization. With the exception of ED,
DBG/DGB, and e, all of the terms in eq 5 can be expressed as
simple functions of biochemical composition and charac-
teristic values of R, δ, and Z for the major constituents (i.e.,
The energy densities of the diet (D) and the consumer (B)
can be derived as
biochemical constituents of the diet or consumer and φi
represents the volume fraction (cm3cm-3) of the diet or
As energy densities of lipid, protein, and carbohydrate vary
little among organisms, we can use standard values for δiof
35.6 kJ cm-3for lipid, 26.8 kJ cm-3for protein, 26.2 kJ cm-3
for carbohydrate, and 0 kJ cm-3for water [corresponding to
characteristic values of 39.5 kJ g-1for lipid, 19.7 kJ g-1for
protein, 17.2 kJ g-1for carbohydrate, and 0 kJ g-1for water
(B) can be derived as
where Zi represents the fugacity capacity of the major
biochemical constituents. Zican be expressed as a fraction
0.05, 0.1, and 1.0 for simple hydrophobic organic chemicals
in proteins, carbohydrates, and lipids, respectively (10, 22).
The absolute value for Zlipid is not required for the model
calculations. In many organisms or tissues, the sorptive
capacity for hydrophobic organic chemicals is dominated
by lipids. However, in organisms or tissues with low lipid
content (e.g., algae, mussels, feces), the contribution of
nonlipid constituents can also be important (10, 19, 23).
Differences in sorptive capacities among different types of
lipids, proteins, and carbohydrates have been observed,
but the differences are relatively small in most cases
Digestive efficiencies on a dry-matter basis, RD, and an
energy basis, RE, are
where Rirepresents the dry-matter digestibility of each of
the major biochemical constituents of the diet by the
consumer. The digestive efficiency scaled to the sorptive
capacity of the diet for a chemical substance is measured by
RZ, calculated as
The dietary absorption efficiency of a chemical, ED, is, in
many cases, available from empirical observations. ED is
related to RD and RE. The same processes that produce a
large R (e.g., long gut residence time, high surface area for
ED. Thus, broad taxa (fish, birds, mammals) tend to have
characteristic values of EDfor simple nonionic hydrophobic
organic chemicals up to log KOWvalues of approximately 7.5
that are not metabolized in the gastrointestinal tract
(27-30). The ratio DBG/DGB, which reflects asymmetry in a
chemical’s gut-to-body exchange, appears to be approxi-
mately 1 in fish and possibly aquatic invertebrates. There is
evidence that DBG/DGB can be somewhat less than 1 in
for DBG/DGBin different taxa. Characteristic values of e and
net growth efficiency for major taxa have been reviewed
(13-16), and species-specific values are available in the
literature for many taxa.
General. To test the performance of the model to make
estimates of the BMF values of poorly metabolizable sub-
DGB)(1 - RZ)
γ + ?
VOL. 40, NO. 5, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY91583
stances in a range of taxa, we calculated BMFmax using
published bioenergetic and chemical transport parameters.
We then compared the model-calculated BMFmax to BMF
that DMcan be assumed to be 0.
Biomagnification Factors. We compiled studies that
measured under field or laboratory conditions. For inver-
tebrates and fish, we included group I and II (i.e., poorly
metabolizable) PCBs (31), mirex, and ∑DDT. For birds and
mammals, we included group I PCBs, mirex, ∑DDT, and
pentachlorodibenzodioxin, although it is metabolized to
some degree. When congener-specific PCB values were not
and higher chlorinated PCBs. When concentrations were
to lipid content reported in the study (animals) or to a lipid
equivalence value calculated from the biochemical compo-
sition of the diet (plants). Laboratory observations were
included only if the authors reported that the chemical
concentrations in the consumer were near steady state with
those in the diet. If data were reported separately for males
and females, we considered males only so that we could use
net growth efficiency as a surrogate for net production
efficiency. We used the geometric mean of reported or
as our estimate of the observed BMFmaxfor each species. A
total of 169 BMFs from 35 published studies met the above
criteria. These study results produced estimates of BMFmax
for 35 species-age combinations, including 7 species of
invertebrates (both aquatic and terrestrial), 6 species of fish,
three species (great tit, ringed seal, wolf), we were able to
estimate observed BMFmax values for both juveniles and
Bioenergetics Parameters. We next compiled data to
parametrize eq 5 for (i) all species-age combinations for
which estimates of observed BMFmaxwere obtained and (ii)
data appear to exist (wolf spider, python, oilbird, vampire
bat, and insectivorous bat), to illustrate how our model can
be used to make predictions of the BMF in unstudied taxa
(Table S1). The only values required to parametrize eq 5 are
and the proximate composition of the species and its diet.
We included fish in the model’s performance analysis only
if the study from which BMFs were obtained also reported
enough information to estimate net growth efficiency. Fish
growth is plastic within and among species (32), largely
because of variation in the energetics of food acquisition
in captivity (34) and 10-20% in wild fish (13, 14), but can
approach 0 in lakes without appropriately sized prey (35).
Only six of 32 studies reported the necessary data and are
to accurate net growth efficiency data to correctly calculate
BMFmaxvalues in fish.
absorption efficiencies (ED) of 0.50 in fish (25) and inverte-
brates (36), 0.95 in birds (28), and 0.90 in mammals (29, 30).
To explore the role of asymmetry in gut/body chemical
diffusion rates, we used values for DGB/DBGof 1 (8, 10) and
Uncertainty Analysis. To assess the effect of error and
When multiple values were available for a model input
S1) and coefficient of variation (CV) assuming a log-normal
distribution. When only a single value was available, we
assumed a CV equal to the mean CV of that parameter for
all species with multiple values. In this way, distributions
used in the simulations.
Results and Discussion
Model-Calculated BMF Values. Figure 1 shows isopleths of
the BMFmaxvalues predicted by eq 5 as a function of ? and
γ. In this general survey of parameter space, BMFmax is
predicted to range from approximately 1 for species with
inefficient digestion and highly efficient growth to greater
than 100 for species with highly efficient digestion and
negligible growth. Invertebrates, juvenile birds, and cows
are predicted to have BMFmax< 10, whereas adult birds and
wild mammals are predicted to have BMFmax >10 and in
some cases (e.g., adult wolves) approaching 100. Estimates
100 in adult wolves (Table S1). In general, invertebrates
exhibited lower observed BMFmax values than birds and
mammals. The exceptions were cows, juvenile birds, and
clam-eating tufted ducks because of their low-fat diets and
high growth efficiencies. Juvenile great tits, common eiders,
adult counterparts because of greater growth efficiency at
the juvenile stage.
BMFmaxvalues to observed BMFmaxvalues for a range of taxa
FIGURE 1. Predicted maximum biomagnification factors (BMFmax,
dashed isopleths) for all realistic combinations of digestion
efficiency (?) and growth dilution (γ) terms in eq 5 (see text for
details). Symbols indicate where species in our validation set fall
within this parameter space (×, invertebrates; O, fish and reptiles;
2, birds; 0, mammals). Labels are AC, arctic cisco; AP, Adelie
penguin; AS, Atlantic salmon; BG, black guillemot; BW, bowhead
IB, insectivorous bat; MY, freshwater mysid; OB, oilbird; PB, polar
bear; PC, polar cod; PY, python; RO, river otter; SA, ringed seal
adult; SJ, ringed seal juvenile; SQ, squid; ST, three-spined
stickleback; TA, great tit adult; TD, tufted duck; TH, marine mysid;
TJ, great tit juvenile; TM, thick-billed murre; VB, vampire bat; WA,
wolf adult; WJ, wolf juvenile; WL, walrus; WS, wolf spider; ZF,
zebrafish; ZM, zebra mussel.
15849ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 5, 2006
were within a factor of 2 of the model-calculated BMFmax
of 3 of the model-predicted BMFmax values. The good fit
between predicted and observed BMFmaxvalues for nearly
all species in the model performance evaluation supports
our contention that bioenergetic efficiencies play a key role
in the biomagnification process and are useful parameters
for assessing the biomagnification potential of chemicals
among taxa. Differences between calculated and observed
BMFmax values might be due to errors in the model and
uncertainty in model input parameters but might also
originate from errors or uncertainty in observed BMFs. For
example, inaccurate diet concentration, lack of steady state
between consumer and diet, occurrence of metabolic
transformation (e.g., pentachlorodibenzodioxin), and ana-
lytical error are model-unrelated factors that might cause
discrepancies between observed and model-predicted BM-
Fmaxvalues in Figure 2.
We were unable to make predictions of BMFmaxfor most
of the observed BMF values reported for fish because of
model calculations using net growth efficiencies ranging
between 0 and 0.5 (Figure S1) produced calculated BMFmax
and laboratory-reared fish showed low observed BMFmax
values because of the high growth efficiency of animals with
growth. Freshwater fish showed a large range of observed
BMFmaxvalues (between 1.2 and 45), possibly because prey
has important implications for the net growth efficiency of
fish (32, 35).
Asymmetry in Gut-Body Transport. Figure 3 illustrates
model-predicted BMFmax values in relation to observed
and asymmetry, respectively, in gut/body exchange rates.
for some taxa, although the relationship between predicted
and observed BMFmax values was highly significant (Table
S2). Assuming DGB/DBG ) 3 for all taxa or for birds and
mammals only, model-calculated BMFmaxvalues followed a
all species (Table S2, Figure 3). A model scenario using DGB/
DBGvalues of 1 for fish and aquatic invertebrates and 3 for
nearest unity, intercept near 0, high r2) between observed
and predicted BMFmaxvalues. These findings are consistent
body-to-gut transport when expressed in terms of fugacity-
can be best expressed by a DGB/DBGvalue of 3 in birds and
mammals. As the magnitude of the growth dilution term γ
values for five taxa for which no measured bioaccumulation
data are available. The species for which we make novel
combinations. Wolf spiders and pythons are known to have
very efficient digestion and might therefore be expected to
have high BMFmaxvalues. Spiders have extra-oral digestion;
ingest only the most soluble, digestible components of their
and use basking to raise gut temperature and enhance
digestion (Table S1), giving these snakes a digestion ratio,
also have high net growth efficiency, and their predicted
BMFmaxvalues are consequently on the order of 2-3. Bats,
on the other hand, have a very low net growth efficiency.
consumption rates, but their high activity level allows them
is therefore less important for bats, whose BMFmaxvalue is
determined primarily by digestive efficiency. Vampire bats,
adapted to a protein-rich diet, have low lipid digestive
efficiency (Table S1) and, consequently, low (negative) RZ.
Insectivorous bats digest lipids very efficiently (Table S1),
giving them a very high RZ and thus a higher predicted
BMFmax. Oilbirds (Steatornis caripensis) are ecologically
similar to fruit bats, with high feeding rates and low growth
is available. Labels are defined in Table S1.
values of gut/body chemical transport asymmetry. Error bars are
standard error of the mean. Equations and statistics for regression
lines are shown in Table S2.
VOL. 40, NO. 5, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY91585
efficiency. However, oilbirds have a relatively low digestive
efficiency for their lipid-rich fruit diet (Table S1). Fecal
egestion is therefore an efficient elimination route, and the
predicted BMFmaxis relatively low for adult birds.
Bioenergetic Efficiencies vs Vital Rates. The model
developed here is expressed in terms of bioenergetic ef-
ficiencies, and this allows for an examination of the effects
on bioaccumulation of ecological traits such as growth
example, it is clear from eq 5 that absolute growth rate has
ratesthat determines the magnitude of growth dilution. Our
analysis highlights the strong influence of growth dilution
on biomagnification in animals with high growth efficiency
animals. Animals in captivity (e.g., in laboratory bioaccu-
mulation tests or farms) are likely to have unnaturally high
net growth efficiencies, giving these individuals a lower
BMFmaxvalue than wild individuals.
Equation 5 also sheds light on the role of feeding rate in
bioaccumulation. Extrinsic factors such as prey abundance
can influence feeding rate and, consequently, the rate of
dietary intake of contaminants, but if this extrinsic factor
has no effect on the bioenergetic efficiencies e and R, then
egestion and growth vary in proportion to feeding, and the
steady-state BMFmax value remains the same. A similar
argument explains the generally higher BMFs observed in
because a substantial portion of their energy budget is
devoted to homeothermy and activities required to ensure
more than wild snakes, but have similarly low BMFs, in part
because their growth efficiency is inflated by low activity
to more rapidly approach a new steady-state BMF after a
diet shift or change in energetic allocation (e.g., at matura-
tion), but the level of that steady state is determined by
digestive and growth efficiencies, not feeding rate.
Production Dilution: Offspring, Milk, and Slime. In
parametrizing eq 5, we used net growth efficiencies to
approximate net production efficiencies, and thus our pre-
dictions apply to males, juveniles, and nonreproductive
females that have no important secretions (e.g., slime, milk,
or silk). It is conceptually simple to extend the analysis to
include all animals. The key is to consider the additional
production (offspring or secretion) as analogous to growth.
Production efficiencies for offspring and milk are much
higher than growth efficiencies in wild adult mammals (e.g.,
12, 39, but cf 17 for domestic mammals), which has the
steady-state BMF value closer to that of a juvenile. Females
that accumulate fat reserves before breeding also have an
elevated production efficiency during this period, although
total production efficiency can be negative with respect to
the female’s “normal”, pre-reproductive parameters are
restored, her steady-state BMF returns to its previous,
relatively high level. She might approach this level slowly,
however, and a steady-state BMF model might not produce
accurate predictions. Species with high feeding rates (e.g.,
small changes in growth or production efficiency during
reproduction (e.g., invertebrates, fish, reptiles) will be most
likely to return to steady state quickly.
the magnitude of the effect of production dilution on the
animal’s BMF. Offspring and milk might have higher lipid
contents than the female, and so, the effect of this efficient
production on the steady-state BMF might be dispropor-
tionately large (17). Slime and silk probably have very low
sorptive capacities, at least for hydrophobic chemicals, and
so, secretion of these materials might represent a negligible
tissue must also be considered. Animal growth is usually
newly produced tissue will be at equilibrium with the rest of
the animal (e.g., 5). However, if production of offspring or
within the consumer’s body, the produced material might
not be at thermodynamic equilibrium with the consumer
when it leaves the body (40, 41).
Biomagnification of Other Chemicals. Notably absent
from eq 5 is an expression for the chemical’s properties,
such as the octanol-water or octanol-air partition coef-
conditions chosen to evaluate the model. The model
calculates the maximum possible biomagnification in the
absence of significant rates of elimination via respiratory
exchange to water (controlled by KOW) or air (controlled by
KOA) or metabolic transformation. The role of bioenergetics
factor is attained is a function of the chemical’s elimination
and metabolic transformation rates, which are chemical-
values, tends to drop with increasing KOW (27). For the
chemicals used in the current model evaluation, this drop
of an animal and is therefore appropriate only for animals
that are at or near steady state with their diets. Animals in
laboratory experiments are often exposed to contaminants
for short durations and might (if depuration rates are slow)
not reach steady state during the exposure period. Wild
in some cases, magnify the contaminant concentration as
the juvenile absorbs yolk or lipid stores (42, 43) or receives
an additional contaminant load via milk (5). Upon com-
mencing exogenous feeding, an animal begins to egest and
of feeding, egestion, and growth decline, and the steady-
state BMFmaxincreases. If the change in these vital rates is
large and rapid, the animal might spend a period of time
below steady state with its diet. This is most likely to be an
issue for birds (1, 2) and possibly mammals (5). Juvenile
invertebrates, fish, and other heterotherms undergo a slow,
gradual change in specific vital rates and are more likely to
maintain steady state with their diet (1). Other issues that
could prevent an animal from reaching steady state include
ontogenetic and seasonal diet shifts; seasonal lipid cycles;
and other large, rapid changes in bioenergetic parameters.
The issue of whether an animal is at steady state is key to
assessing the utility of eq 5 and will be a fruitful avenue for
Thanks to the Natural Sciences and Engineering Research
Council of Canada for support and the members of the
ToxLab group at SFU for helpful discussions.
Supporting Information Available
Derivation of eq 1, biomagnification and bioenergetics
15869ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 5, 2006
predicted and observed log BMFmaxvalues, and calculated
maximum biomagnification factors as a function of net
the Internet at http://pubs.acs.org.
(1) Norstrom, R. Understanding bioaccumulation of POPs in food
webssChemical, biological, ecological and environment con-
siderations. J. Environ. Sci. Pollut. Res. 2002, 9, 300.
(2) Nichols, J. W.; Echols, K. R.; Tillitt, D. E.; Secord, A. L.; McCarty,
in nestling tree swallows from two contaminated sites on the
Upper Hudson River, New York. Environ. Sci. Technol. 2004,
(3) Barber, M. C.; Suarez, L. A.; Lassiter, R. R. Modeling bioaccu-
mulation of organic pollutants in fish with an application to
PCBs in Lake Ontario salmonids. Can. J. Fish. Aquat. Sci. 1991,
(4) Jackson, L. J. Contrast of constant and bioenergetics-based
assimilation efficiencies in describing ecosystems. Ecol. Appl.
1996, 6, 1355.
model for hydrophobic contaminants in marine mammals.
Environ. Toxicol. Chem. 1999, 18, 2622.
(6) Glaser, D.; Connolly, J. P. A model of p,p′-DDE and total PCB
bioaccumulation in birds from the Southern California Bight.
Continental Shelf Res. 2002, 22, 1079.
(7) Connolly, J. P.; Pedersen, C. J. A thermodynamic-based evalu-
ation of organic-chemical accumulation in aquatic organisms.
Environ. Sci. Technol. 1988, 22, 99.
(8) Gobas, F. A. P. C.; Zhang, X.; Wells, R. Gastrointestinal
1993, 27, 2855.
(9) Mackay, D. 1982. Correlation of bioconcentration factors.
Environ. Sci. Technol. 1982, 16, 274.
(10) Gobas, F. A. P. C.; Wilcockson, J. B.; Russell, R. W.; Haffner, G.
field conditions. Environ. Sci. Technol. 1999, 33, 133.
(11) Mackay, D., Fraser, A. Bioaccumulation of persistent organic
(12) Brody, S. Bioenergetics and Growth; Reinhold Publishing
Corporation: New York, 1945.
(13) Welch, H. E. Relationships between assimilation efficiencies
and growth efficiencies for aquatic consumers. Ecology 1968,
(Koch) (Araneae: Lycosidae) with an appraisal of production
efficiency in ectothermic animals. J. Anim. Ecol. 1978, 47, 627.
(15) Humphreys, W. F. Production and respiration in animal
communities. J. Anim. Ecol. 1979, 48, 427.
Press: San Diego, 1987.
(17) McLachlan, M. S. Bioaccumulation of hydrophobic chemicals
in agricultural feed chains. Environ. Sci. Technol. 1996, 30, 252.
(18) Makarieva, A. M.; Gorshkov, V. G.; Li, B. L. Ontogenetic growth:
Models and theory. Ecol. Model. 2004, 176, 15.
(19) Mackay, D.; Fraser, A. Bioaccumulation of persistent organic
(20) Nichols, J. W.; Fitzsimmons, P. N.; Whiteman, F. W. A physi-
ologically based toxicokinetic model for dietary uptake of
hydrophobic organic compounds by fishsII. Simulation of
chronic exposure scenarios. Toxicol. Sci. 2004, 77, 219.
Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and
(22) Otton, S. V. A Method to Measure the Sorptive Capacity of
Thesis, Simon Fraser University, Burnaby, BC, Canada, 2004.
carbon as the sorbing matrix in the modeling of PCB ac-
cumulation in phytoplankton. Environ. Sci. Technol. 1999, 33,
(24) Dobbs, A. J.; Williams, N. Fat solubilitysA property of envi-
ronmental relevance? Chemosphere 1983, 12, 97.
water systems and correlations with fish bioconcentration
factors. Environ. Toxicol. Chem. 1985, 19, 57.
D. A novel method for measuring membrane-water partition
coefficients of hydrophobic organic chemicalssComparison
(27) Buckman, A. H.; Brown, S. B.; Hoekstra, P. F.; Solomon, K. R.;
Fisk, A. T. Toxicokinetics of three polychlorinated biphenyl
technical mixtures in rainbow trout (Oncorhynchus mykiss).
Environ. Toxicol. Chem. 2004, 23, 1725.
of ring doves following initial exposure to an arochlor con-
taminated diet. Environ. Toxicol. Chem. 2000, 19, 2707.
(29) McLachlan, M. S. Mass balance of polychlorinated biphenyls
Food Chem. 1993, 41, 474.
(30) Moser, G. A.; McLachlan, M. S. The influence of dietary
concentration on the absorption and excretion of persistent
lipophilic organic pollutants in the human intestinal tract.
Chemosphere 2001, 45, 201.
(31) Boon, J. P.; vanderMeer, J.; Allchin, C. R.; Law, R. J.; Klunsoyr,
J.; Leonards, P. E. G.; Spliid, H.; StorrHansen, E.; Mckenzie, C.;
in fish-eating mammals: Structural evidence for induction of
cytochrome P450. Arch. Environ. Contam. Toxicol. 1997, 33,
(32) Trudel, M.; Tremblay, A.; Schetagne, R.; Rasmussen, J. B. Why
are dwarf fish so small? An energetic analysis of polymorphism
in lake whitefish (Coregonus clupeaformis). Can. J. Fish. Aquat.
Sci. 2001, 58, 394.
benthic invertebrate communities on the feeding ecology of
yellow perch (Perca flavescens) in metal-contaminated lakes.
Can. J. Fish. Aquat. Sci. 2005, 62, 153.
(34) Brett, J. R.; Groves, T. D. D. Physiological Energetics; In Fish
Press: New York, 1979; Vol. 8.
(35) Sherwood, G. D.; Kovecses, J.; Hontela, A.; Rasmussen, J. B.
Simplified food webs lead to energetic bottlenecks in polluted
lakes. Can. J. Fish. Aquat. Sci. 2002, 59, 1.
(36) Wang, W. X.; Fisher, N. S. Assimilation efficiencies of chemical
contaminants in aquatic invertebrates: A synthesis. Environ.
Toxicol. Chem. 1999, 18, 2034.
(37) Duverge ´, P. L.; Jones, G.; Rydell, J.; Ransome, R. D. Functional
significance of emergence timing in bats. Ecography 2000, 23,
(38) Fisk, A. T.; Hobson, K. A.; Norstrom, R. J. Influence of chemical
and biological factors on trophic transfer of persistent organic
Sci. Technol. 2001, 35, 732.
(39) Korine, C.; Speakman, J.; Arad, Z. Reproductive energetics of
captive and free-ranging Egyptian fruit bats (Rousettus aegyp-
tiacus). Ecology 2004, 85, 220.
(40) Braune, B. M.; Norstrom, R. J. Dynamics of organochlorine
compounds in herring-gulls.s3. Tissue distribution and bio-
accumulation in lake-ontario gulls. Environ. Toxicol. Chem.
1989, 8, 957.
(41) deBruyn, A. M. H.; Ikonomou, M. G.; Gobas, F. A. P. C.
Magnification and toxicity of PCBs, PCDDs, and PCDFs in
upriver-migrating Pacific salmon. Environ. Sci. Technol. 2004,
D. B. Development and validation of a herring gull embryo
toxicokinetic model for PCBs. Ecotoxicology 2003, 12, 55.
D.; Monteverdi, G.; Mastrodone, P. Exposure, uptake and
disposition of chemicals in reproductive and developmental
states of selected oviparous vertebrates. In Reproductive and
DiGiulio, R., Tillett, D., Eds.; SETAC Press/Lewis Publishers:
Ann Arbor, MI, 1999.
Received for review September 9, 2005. Revised manuscript
received December 16, 2005. Accepted December 22, 2005.
VOL. 40, NO. 5, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY91587