Florin_et_al-2011-Functional_Ecology

The biological basis for understanding and predicting
dietary-induced variation in nitrogen and sulphur
isotope ratio discrimination
Scott T. Florin*,1
, Laura A. Felicetti2
and Charles T. Robbins1,2
1
School of Biological Sciences, Washington State University, Pullman, Washington 99164-4236, USA; and 2
Department
of Natural Resource Sciences, Washington State University, Pullman, Washington 99164-4236, USA
Summary
1. Accurately predicting isotope ratio discrimination is central to using mixing models to esti-
mate assimilated diets of wild animals. This process is complicated when omnivores consume
mixed diets because their discrimination is unlikely to be the weighted average of the various die-
tary constituents as occurs in current models.
2. We sought a basic understanding of how protein quality and quantity determine D15
N and
D34
S in mammals and birds. Dietary protein is the primary source of both elements in many
plants and animals. Low protein quality and high protein content have the potential to increase
D15
N by increasing protein turnover.
3. Protein quality, defined as the relative amount of the most limiting amino acid, accounted for
87–90% of the variation in D15
N when mammals and birds consumed plant matter and mixed
diets of plants and animals with protein of intermediate quality and quantity. However, foods
containing relatively large amounts of high quality protein (i.e. vertebrate flesh) and foods with
exceptionally low quality protein (e.g. lentils, Lens culinaris) had disparate nitrogen discrimina-
tions relative to what would be predicted from protein quality alone.
4. Supplementation of plant and animal diets with nitrogen-free carbohydrates and fats to
reduce protein quantity did not reduce D15
N in three plant-based diets fed to laboratory rats, but
reduced D15
N in two of three meat diets with >50% protein.
5. D34
S was weakly correlated with D15
N (R2
= 0Æ48) but was highly correlated with dietary
d34
S (R2
= 0Æ89). Because methionine, a sulphur amino acid, was the most limiting amino acid
in all diets, sulphur should be highly conserved as indicated by the lack of any change in D34
S
when diets were supplemented with carbohydrates and fat to both provide additional energy and
reduce protein content.
6. Predictive equations incorporating both protein quality and quantity accounted for 81% of
the variation in D15
N and offer the opportunity to create more realistic mixing models to accu-
rately estimate assimilated diets for omnivores.
Key-words: assimilated diet, isotope discrimination, nitrogen, protein quality, stable isotopes,
sulphur
Introduction
Accurately predicting isotopic discrimination is central to
estimating assimilated diets of wild animals when using sta-
ble isotopes (Martinez del Rio et al. 2009). While many
studies have postulated or identified causes of variation in
discrimination (Fantle et al. 1999; Roth & Hobson 2000;
McCutchan et al. 2003; Pearson et al. 2003; Vanderklift &
Ponsard 2003; Gaye-Siessegger et al. 2004, 2007; Robbins,
Felicetti & Sponheimer 2005; Miron et al. 2006; Caut, Ang-
ulo & Courchamp 2008, 2009; Tsahar et al. 2008; Robbins,
Felicetti & Florin 2010; Smith et al. 2010), none have pro-
posed cause-effect biologically based models for accurately
estimating unknown discriminations even though selection
of discrimination values is the single most important
assumption determining assimilated diet estimates. The lack*Correspondence author. E-mail: sflorin@wsu.edu
Ó 2010 The Authors. Functional Ecology Ó 2010 British Ecological Society
Functional Ecology 2011, 25, 519–526 doi: 10.1111/j.1365-2435.2010.01799.x

of such models to accurately predict nitrogen, carbon, or
sulphur discriminations, particularly for foods in mixed
diets, may lead to unacceptable errors in estimating assimi-
lated diets of ancestral humans and wild animals (Caut,
Angulo & Courchamp 2008; Robbins, Felicetti & Florin
2010).
Current approaches to estimating unknown discrimina-
tions for foods consumed by free-ranging animals include:
(i) feeding wild-collected foods to captive animals and directly
measuring their discrimination, which is not always feasible
and may rarely simulate field complexity; (ii) using a grand
mean for all foods (e.g. 2Æ0–3Æ4& for nitrogen and 0& for
sulphur), which ignores the three- to fourfold variation in
D15
N (e.g. c. )2 to 6&) and D34
S (e.g. c. )3 to 8&); or (iii)
using various regressions between dietary isotope values and
discriminations that have been determined with captive wild-
life consuming a wide range of foods, which describe very
general relationships that may not be cause-effect (Peterson
& Fry 1987; McCutchan et al. 2003; Vanderklift & Ponsard
2003; Robbins, Felicetti & Sponheimer 2005; Caut, Angulo &
Courchamp 2009; Martinez del Rio et al. 2009; Robbins,
Felicetti & Florin 2010).
Two major hypotheses have been proposed to explain
much of the dietary-induced variation in D15
N. The protein
quantity hypothesis suggests that as dietary protein content
(%) or intake (g day)1
) increase, D15
N will increase (Pearson
et al. 2003; Martinez del Rio et al. 2009). The protein quality
hypothesis suggests that as protein quality decreases, D15
N
will increase (Roth & Hobson 2000; Robbins, Felicetti &
Sponheimer 2005; Robbins, Felicetti & Florin 2010). Both
are based on the observation or hypothesis that as dietary
protein intake or amino acid scavenging increase, nitrogen
excretion will increase and lead to the preferential retention of
15
N which will elevate the animal’s d15
N value relative to the
diet.
Although Robbins, Felicetti & Sponheimer (2005) and
Robbins, Felicetti & Florin (2010) found no support for the
protein quantity hypothesis when plotting either nitrogen
content or carbon : nitrogen ratios against D15
N, such plots
are confounded by lower protein, largely plant-based diets of
poorer protein quality at one end of the regression and higher
protein, largely animal-based diets of higher protein quality
at the other. If both protein quality and quantity are impor-
tant, D15
N may be elevated when plant-based diets are con-
sumed primarily because of their poorer protein quality and
when animal-based diets are consumed primarily because of
their higher protein content. Thus, we hypothesized that both
protein quality and quantity may be important, but the rela-
tionships are more complex than either proposal alone sug-
gests.
Thus far, little use has been made of sulphur isotopes for
estimating assimilated diet, although the consumer’s isotope
value should reflect the dietary isotope value (Felicetti et al.
2003; McCutchan et al. 2003; Arneson & MacAvoy 2005).
D15
N and D34
S may be related in that sulphur amino acids
(methionine, cystine, cysteine and taurine) are important
sources of organic sulphur (Arneson & MacAvoy 2005).
If sulphur amino acids are important in determining protein
quality, dietary sulphur amino acid content may be important
in determining both D15
N and D34
S. Consequently, we sought
a unified concept incorporating both protein quality and
quantity that could be used to understand and accurately pre-
dict D15
N, D34
S and assimilated diets of omnivores.
Materials and methods
QUANTIF YING PROTEIN QUANT ITY AND QUALIT Y
While nitrogen or protein quantity (N · 6Æ25) has been measured in
virtually all studies, protein quality has not. There are many measures
of protein quality. Some are based on feeding studies (e.g. protein effi-
ciency ratio, biological value, or net protein utilization) and others
are based on how well the essential amino acid profile of a particular
food matches a hypothetical perfect protein or the animal’s require-
ments (e.g. chemical score). The latter estimates are appealing in that
amino acid profiles of many foods have been determined and very
extensive effort has been made to define the amino acid requirements
of domestic and laboratory animals [NRC (National Research Coun-
cil) 1995].
The complete amino acid profiles of Chinook salmon (Oncorhyn-
chus tshawytscha) and white-tailed deer (Odocoileus virginianus) fed to
brown bears (Ursus arctos) and American black bears (Ursus americ-
anus) (Hilderbrand et al. 1996; Felicetti et al. 2003) and various diets
composed of corn, wheat, alfalfa, soybean meal, lentils, chicken meal,
pork meat and bone meal and fish meal fed to laboratory rats (Rob-
bins, Felicetti & Florin 2010; current study) were determined at the
University of Missouri Agricultural Experiment Station Chemical
Laboratories. Briefly, acid and alkaline hydrolysates were analysed
using a high-performance liquid chromatographic amino acid ana-
lyzer. Additional amino acid profiles or protein contents of foods not
reported by other investigators were estimated from the compilations
of NRC (1994), Davis et al. (1994), American Casein Co. (Burling-
ton, NJ, USA) and the USDA National Nutrient Database for
Standard Reference, Agricultural Research Service (http://www.nal.
usda.gov/fnic/foodcomp/search/) (see Table S1, Supporting informa-
tion).
The basis for estimating protein quality was to express the concen-
tration of each essential amino acid in the diet as a percent of the diet’s
crude protein (N · 6Æ25) content. This relative concentration of each
amino acid was compared with the estimated dietary requirement of
that amino acid as a percent of the total protein requirement for
growth by laboratory rats (Rattus rattus) (NRC 1995) to determine
which amino acid might be most limiting. Amino acid requirements
for growing laboratory rats were used as the standard for all animals
because (i) the amino acid requirements for wild animals are almost
entirely unknown; (ii) the current and previous study (Robbins, Felic-
etti & Florin 2010) used laboratory rats; and (iii) laboratory rats have
not been heavily selected for meat, milk, or egg production as have
many other domestic animals (e.g. livestock and poultry) and, there-
fore, may be a more appropriate comparison with wild animals.
SELECTING NITR OGEN AND SULPHUR DISCRIMINA-
TION VALUES
D15
N and D34
S values for serum, plasma, whole blood, or red blood
cells were sought for diets that covered the widest possible ranges of
protein quality and quantity, had been fed long enough to ensure diet
to animal equilibration and had been fed by multiple investigators or
Ó 2010 The Authors. Functional Ecology Ó 2010 British Ecological Society, Functional Ecology, 25, 519–526
520 S. T. Florin et al.

in various combinations to ensure that the reported isotope discrimi-
nations were reliable. Unfortunately, results on commercial rodent
and poultry diets as well as several other diets could not be used
because of the impossibility of estimating amino acid profiles. Simi-
larly, feeding studies that used pelleted diets were excluded because of
concern about protein damage (Robbins, Felicetti & Florin 2010),
and studies that fed fungi, crustacea, or insects [e.g. mealworms (Ten-
ebrio molitor)] were excluded because much of their nitrogen occurs as
chitin (a nitrogen-containing carbohydrate) (Claridge et al. 1999;
Pearson et al. 2003). For example, van Tets & Hulbert (1999)
estimated that 69% of the nitrogen in mealworms occurred as non-
protein chitin.
TESTING THE INTERACTION BETWEEN PR OT EIN
QUANTIT Y, QUALITY AND D1 5
N
Two approaches were used to test the interactions between protein
quantity and quality in determining D15
N. The first approach was an
indirect test in which the relative amount of the most limiting essential
amino acid was compared with the D15
N for several diet–animal com-
binations used in the current and previous studies (Hobson & Clark
1992; Hilderbrand et al. 1996; Hobson et al. 1996; Ben-David &
Schell 2001; Jenkins et al. 2001; Bearhop et al. 2002; Lesage, Hammill
& Kavacs 2002; Felicetti et al. 2003; Sponheimer et al. 2003; Ogden,
Hobson & Lank 2004; Arneson & MacAvoy 2005; Cherel, Hobson &
Hassani 2005; Robbins, Felicetti & Sponheimer 2005; Podlesak &
McWilliams 2006; Darr & Hewitt 2008; Tsahar et al. 2008; Robbins,
Felicetti & Florin 2010) (see Table S2). If the protein quality hypothe-
sis is valid, D15
N should decrease as the concentration of the most lim-
iting amino acid increases across diets. Similarly, if the protein
quantity hypothesis is valid, D15
N should increase above the relation-
ship determined by protein quality alone once the most limiting
amino acid is no longer the sole determinant of dietary protein turn-
over.
The second approach was a direct test in which foods ranging in
both protein quality and quantity were supplemented with additional
energy to dilute the protein concentration and thereby reduce daily
protein intake. The D15
N and D34
S of animals consuming the energy-
supplemented diets should be less than the non-supplemented diets
when protein quantity becomes important in determining discrimina-
tion. Specifically, we hypothesized that D15
N values for plant-based
diets would be less likely to decrease with energy dilution than ani-
mal-based diets.
Thus, the diets used included fish meal (Brevoortia tyrannus),
chicken meal, pork meat and bone meal, soybean meal, lentils and
wheat because they cover a wide range in both protein quantity and
quality in both plants and animals. All feeds were purchased as single
batches, finely ground and mixed thoroughly to ensure that composi-
tion and isotopic values were constant. Each diet was fed in the undi-
luted form followed immediately by the diluted form to the same 10
rats. The diluted diets were supplemented with nitrogen- and sulphur-
free sucrose, starch and corn oil in the ratio of 5 : 2 : 2 : 1, such that
the protein concentration was reduced by 50%. Further dilution was
not attempted because of concern about creating nutritional deficien-
cies.
Ten male, Sprague–Dawley laboratory rats were used in all feeding
trials. Each feeding trial lasted 21 days to ensure that plasma had
equilibrated with the diet and followed the protocol of Robbins,
Felicetti & Florin (2010). Blood samples were collected in heparinized
tubes at the end of each feeding trial. Plasma was separated, frozen,
and freeze-dried. All rats were fed ad libitum to promote positive
energy and protein balance, weight gain, and therefore minimal tissue
mobilization. Rats were weighed weekly.
ISOTOPIC AND STATISTIC AL ANALYSES
d15
N and d34
S values for diets and freeze-dried plasma were deter-
mined with a continuous flow isotope ratio mass spectrometer (Delta
PlusXP; Thermo Finnigan, Bremen, Germany) at the Washington
State University Stable Isotope Core Laboratory. Mean dietary iso-
tope values were based on the analyses of five samples per diet. Nitro-
gen isotope ratios are reported as per mil (&) relative to atmospheric
nitrogen (d15
N). Sulphur isotope ratios are reported as per mil relative
to Vienna Canon Diablo Troilite by assigning a value of )0Æ3& to
IAEA S-1 silver sulphide. Laboratory reference standards (acetanilide
and keratin for nitrogen and sulfanilimide, IAEA S-2, IAEA SO5,
and IAEA S3 for sulphur) were interspersed throughout each analysis
to ensure maintenance of calibration. Analytical errors (±1 SD) for
the above standards were £0Æ1& for nitrogen and £0Æ4& for sulphur.
Linear and curvilinear least squares regressions were used to model
all relationships (SAS 1998). Differences in slopes of regressions were
tested using small sample t-tests (Kleinbaum & Kupper 1978). ANOVA
was used to test for differences in discrimination between diets.
A P-value of <0Æ05 was considered significant. Means are reported
with ±1 SD.
Results
NITROGEN ISOTOPE RAT IO DISCR IMINATION
Protein quality as defined by the relative methionine concen-
tration accounted for 87–90% of the variation in D15
N when
animals consumed diets that ranged from 6Æ9% to 53Æ8% pro-
tein with methionine concentrations ranging from 1Æ3% to
2Æ6% (Fig. 1, see Tables S1 and S2). The inclusion of the
other sulphur-containing amino acids that can partially sub-
stitute for methionine (i.e. cystine, cysteine and taurine) did
not improve the regressions. The pattern of decreasing D15
N
with increasing protein quality occurred for laboratory rats
consuming a wide range of single-item and mixed diets (Rob-
bins, Felicetti & Florin 2010; current study), non-primate
neonates consuming milk (Robbins 1993; Davis et al. 1994;
Jenkins et al. 2001; Robbins, Felicetti & Sponheimer 2005),
wild and domestic ruminants consuming alfalfa or alfalfa and
corn (Jenkins et al. 2001; Sponheimer et al. 2003; Darr & He-
witt 2008), and yellow-vented bulbuls (Pycnonotus xanthopy-
gos) and yellow-rumped warblers (Dendroica coronata)
consuming mixed diets of casein and bananas (Tsahar et al.
2008) or casein, sugar and olive oil (Podlesak & McWilliams
2006).
However, several diets had either higher or lower discrimi-
nations than predicted by the regression equations of Fig. 1.
For example, lentils containing relatively low quality protein
had a much lower D15
N than predicted by the regressions. At
the other extreme, high-protein meat diets containing rela-
tively high quality protein (e.g. fish, fish meal, chicken meal,
ungulates, and quail) fed to various mammals (Canis latrans,
Halichoerus grypus, Mustela vison, Pagophilus groenlandicus,
Phoca hispida, Phoca vitulina, U. americanus, U. arctos) and
Understanding nitrogen and sulphur discrimination 521

birds (Calidris alpine pacifica, Catharacta skua, Corvus brac-
hyrhynchos, Falco peregrines, Larus delawarensis) had higher
nitrogen discriminations than predicted from protein quality
alone (Fig. 1, see Tables S1 and S2) (Hobson & Clark 1992;
Hilderbrand et al. 1996; Hobson et al. 1996; Ben-David &
Schell 2001; Bearhop et al. 2002; Lesage, Hammill & Kavacs
2002; Felicetti et al. 2003; Ogden, Hobson & Lank 2004;
Arneson & MacAvoy 2005; Cherel, Hobson & Hassani 2005;
Robbins, Felicetti & Florin 2010). Discriminations for five
of six meat diets averaged 1Æ1 ± 0Æ4& higher (range =
0Æ5–1Æ6&) (Fig. 1) than predicted from protein quality alone.
The exception to this trend occurred when laboratory rats
consumed pork meat and bone meal that contained relatively
low quality protein (Fig. 1, see Tables S1 and S2). Its discrim-
ination (5Æ0 ± 0Æ1&) was similar to what would be predicted
from the more general protein quality regressions of Fig. 1
(5Æ1–5Æ2&).
The protein content of the plant and animal foods used in
the protein dilution study ranged from 12Æ5% to 72Æ0%
(Fig. 2, see Table S2); and protein quality in those foods was
limited by the amino acid methionine, which ranged from
0Æ85% to 2Æ61% of the crude protein (see Tables S1 and S2).
Average daily protein intake was reduced by 48Æ7 ± 3Æ5%
when rats consumed the diluted diets relative to the undiluted
diets. Rats gained weight on all plant-based diets with and
without dilution (1Æ7 ± 1Æ2 g day)1
, range = 0Æ7–3Æ7) and
on five of six animal-based diets (1Æ3 ± 0Æ5 g day)1
, ran-
ge = 0Æ8–2Æ0). The exception was some rats lost weight on
the undiluted pork meat and bone meal ()1Æ0 ± 1Æ5 g day)1
),
but they gained weight on diluted pork meat and bone meal
(1Æ5 ± 0Æ35 g day)1
). However, there was no difference in the
nitrogen discrimination for rats that lost weight when con-
suming the pork meat and bone meal as compared with those
that maintained or gained weight (t = 0Æ58, P = 0Æ59).
Nitrogen discrimination did not decrease with energy dilu-
tion in any of the plant-based diets (lentils, F = 0Æ43,
P = 0Æ52 and soybean meal F = 0Æ40, P = 0Æ54), although
D15
N slightly increased (0Æ17 ± 0Æ17&) when wheat was
diluted (F = 7Æ53, P = 0Æ01) (Figs 1 and 2). The mean differ-
ence in discrimination due to dilution for the three plant diets
was 0Æ06 ± 0Æ13& and did not differ from 0 (t = 0Æ77,
P = 0Æ52). D15
N decreased in two of three meat diets (fish
meal, F = 91Æ4, P < 0Æ0001 and pork meat and bone meal,
F = 34Æ9, P < 0Æ001), but did not decrease when chicken
meal was diluted (F = 1Æ69, P = 0Æ21) even though seven of
the ten rats had lower discriminations when consuming the
diluted diet (Figs 1 and 2).
Fig. 1. The relationship between dietary protein quality as defined by
the limiting amino acid (methionine) and D15
N for the plasma, serum
or red blood cells of laboratory rats consuming various diets of corn,
wheat, alfalfa, lentils, soybean meal, fish meal, pork meat and bone
meal, chicken meal, and their mixtures (Robbins, Felicetti & Florin
2010; current study) and various wild birds and mammals consuming
fish (Hobson & Clark 1992; Hilderbrand et al. 1996; Hobson et al.
1996; Ben-David & Schell 2001; Bearhop et al. 2002; Lesage, Ham-
mill & Kavacs 2002; Felicetti et al. 2003; Cherel, Hobson & Hassani
2005), fish meal (Ogden, Hobson & Lank 2004; Arneson & MacAvoy
2005; Robbins, Felicetti & Florin 2010), quail (Hobson & Clark
1992), ungulates (Hilderbrand et al. 1996; Ben-David & Schell 2001;
Bearhop et al. 2002), alfalfa or alfalfa and corn (Jenkins et al. 2001;
Sponheimer et al. 2003; Darr & Hewitt 2008), non-primate milks
(Davis et al. 1994; Jenkins et al. 2001; Robbins, Felicetti & Sponhei-
mer 2005) and casein-supplemented diets (Podlesak & McWilliams
2006; Tsahar et al. 2008) (Table S2). Although results for vertebrate
flesh with high quality protein (i.e. ungulates, fish meal, fish, chicken
meal and quail) and lentils as the entire diet are plotted, they are not
included in the regressions.
–1
–0·5
0
0·5
1
0·5 1 1·5 2 2·5
ChangeinΔ15NorΔ34S(‰)
Protein quality (methionine content as a % of dietary protein)
Lentils
Soybean meal
Wheat
Fish mealPork meal
Chicken meal
Fig. 2. The effect on nitrogen and sulphur discrimination when diets
composed of fish meal, chicken meal, pork meat and bone meal, soy-
bean meal, lentils and wheat were fed to laboratory rats with and
without dilution. The diluted diets were created by supplementing
each of the above foods with sucrose, starch and corn oil in the ratio
of 5 : 2 : 2 : 1 to reduce protein concentration to half of that in the
undiluted diet. The change in D15
N and D34
S is the difference between
when laboratory rats were fed the diluted diet minus the undiluted
diet.

Nitrogen discriminations can be predicted (R2
= 0Æ81,
F ‡ 68Æ4, P < 0Æ0001, Fig. 3) across the breadth of dietary
data by either one of two equations utilizing both protein
quality and quantity:
D15
N ¼ 7Á62 À 2Á11X þ 0Á015Z eqn 1
D15
N ¼ À6Á02 þ 0Á14X þ 0Á015Z eqn 2
where X is protein quality [eqn 1, methionine content as a
per cent of total dietary protein (Fig. 1a) or eqn 2, the rela-
tive deficit of the most limiting amino acid as a per cent of
the requirement for growth by laboratory rats (Fig. 1b)]
and Z is dietary protein content (% of total dietary dry mat-
ter). The equations utilized all data of Table S2 and Fig. 1
with the exception of the values for lentils (see Discussion).
Protein quality accounted for 75% of the variation
(F ‡ 98Æ9, P < 0Æ0001) and protein quantity for 7%
(F = 2Æ5, P = 0Æ12). More complex regressions, such as
curvilinear regressions or linear and curvilinear regressions
with thresholds for a protein quantity effect (e.g. ‡50%),
produced similar overall predictive capabilities (R2
=
0Æ82–0Æ84, F ‡ 78Æ9, P < 0Æ0001) and estimates of the rela-
tive importance of protein quality (74–76%) and quantity
(5–6%) (eqns 1 and 2).
SULPHUR ISOTOPE RATIO DISCRIMINAT ION
Sulphur amino acids accounted for 84 ± 20% of the dietary
sulphur in corn, wheat, alfalfa, soybean meal, fish meal,
chicken meal, and pork meat and bone meal. However, D34
S
was not highly correlated with D15
N (Fig. 4). D34
S did not
change when sucrose, starch and corn oil were added to any
of the six feeds relative to the undiluted diets (mean change in
D34
S values with dilution = 0Æ02 ± 0Æ13, F = 0Æ34–1Æ26,
P = 0Æ08–0Æ77) (Fig. 2). Dietary d34
S accounted for 89% of
the variation in D34
S (see Table S3, Fig. 5). Regressions
between various measures of sulphur amino acid content,
including total sulphur amino acid content, methionine con-
tent, cystine and cysteine content, and methionine to cystine
ratio, had lower correlation coefficients that ranged from
0Æ46 to 0Æ77.
Discussion
Numerous animal and dietary factors have been proposed to
affect nitrogen discrimination by specific tissues. The animal
factors include intake rate, growth rate, metabolic rate, iso-
tope routing, and type of nitrogen excretion (ureotelic or uri-
cotelic), and the dietary factors include protein quality and
quantity (Martinez del Rio et al. 2009; Kelly & Martinez del
Rio 2010; Smith et al. 2010). The animal factors create con-
cern when trying to estimate the assimilated diets of both
ancient and extant animals because they are rarely known.
The current relationships (Fig. 1 and eqns 1 and 2), which
were developed for mammals and birds that were either main-
taining or gaining weight, suggest that most of the variation
in nitrogen discriminations under these conditions is due to
1
3
5
7
1 3 5 7
PredictedΔ15N(‰)
Observed Δ15N (‰)
Y = 0·78 + 0·81X
R2
= 0·81
N = 35
t = 33·8, P < 0·0001
1:1
Fig. 3. The relationship between the observed discriminations of the
diets in Table S2 and their predicted discriminations when solving
eqns 1 and 2 utilizing their respective protein qualities and quantities.
Dashed line is the 1 : 1 relationship between the variables.
Fig. 4. The relationship between nitrogen and sulphur discrimination
for a range of foods fed to laboratory rats (Robbins, Felicetti & Florin
2010; current study) and grizzly bears (Felicetti et al. 2003).
Fig. 5. The relationship between dietary d34
S, plasma or serum d34
S,
and D34
S for diets fed to laboratory rats (Robbins, Felicetti & Florin
2010; current study) and grizzly bears (Felicetti et al. 2003)
(Table S3).

dietary protein quality and quantity (R2
= 0Æ81–0Æ90). As we
hypothesized, D15
N values for animals consuming plant-
based diets of lower protein quality, even when containing rel-
atively large amounts of protein (e.g. lentils and soybean
meal), were not reduced by dietary energy dilution. D15
N val-
ues for animals consuming two of three high protein meat
diets were reduced when diluted with additional energy,
although the reductions were relatively small in both the dilu-
tion feeding studies (Figs 1 and 2) and in eqns 1 and 2 when
using data from this and other studies.
Nevertheless, the predictive power of the regressions based
on these two variables exceeds that determined by the more
common regressions between dietary d15
N and D15
N, which
explain from 0 to 67% (mean = 40 ± 32%) of the variation
in birds and mammals (Caut, Angulo & Courchamp 2009;
Robbins, Felicetti & Florin 2010). Furthermore, the use of
laboratory rat nutrient standards for growth as a basis for
comparing a wide variety of birds and mammals of various
sizes, gastrointestinal tracts, and productivity suggests that
dietary-induced metabolic relationships determining discrim-
ination are quite conservative.
Lentils, soybean meal, and pork meat and bone meal were
chosen as test foods because of their high protein content but
generally low protein quality. Both of these characteristics
should produce relatively high D15
N values, with the lentil
value being extremely high. For example, the expected D15
N
for lentils based on the regressions of Fig. 1 would have ran-
ged from 7Æ4% to 9Æ8&. However, the D15
N for lentils
(5Æ6 ± 0Æ2&), soybean meal (5Æ7 ± 0Æ1&), and pork meat
and bone meal (5Æ0 ± 0Æ1&) did not exceed 6Æ0&.
In a compilation of 134 D15
N values for various tissues
from mammals, birds, crustacea, insects and fish (Vanderklift
& Ponsard 2003), <4% of the values were above 5Æ5& and
none exceeded 6&. In a more recent compilation of 142 D15
N
values for mammals and birds (Caut, Angulo & Courchamp
2009), only four were above 6&, although three of the four
were incorrectly estimated from Felicetti et al. (2003) and
actually ranged from 4Æ3& to 5Æ8&. Thus, the aggregate of
these observations suggests an upper limit to D15
N of c. 6&
for mammals and birds consuming foods that do not contain
significant amounts of non-protein nitrogen. Therefore, D15
N
estimates produced by eqns 1 and 2 should be capped at a
maximum of 6& unless a particular food–animal combina-
tion is known to produce a higher discrimination.
The regression between dietary d34
S and D34
S has a higher
correlation coefficient than those measured for similar carbon
and nitrogen regressions and, therefore, may be all that is
needed to estimate assimilated diet (Hilderbrand et al. 1996;
Felicetti et al. 2003; McCutchan et al. 2003; Vanderklift &
Ponsard 2003; Robbins, Felicetti & Florin 2010). We hypoth-
esize that this high correlation coefficient occurred in this
study because methionine was the primary, limiting, essential
amino acid in all diets. Therefore, sulphur and the sulphur
amino acids should be highly conserved during animal metab-
olism as demonstrated by the lack of any change in D34
S dur-
ing the dietary dilution study. The relatively low correlation
coefficient (0Æ48) between D34
S and D15
N is similar to earlier
results (0Æ44) for insects and fish (McCutchan et al. 2003),
which suggests a more complex relationship between the two
variables. Presumably, D15
N reflects the metabolism of all
amino acids and varies with both protein quality and quan-
tity, whereas D34
S reflects the metabolism of only sulphur
amino acids. Therefore, the two variables are not directly
related and the lower correlation coefficient should be
expected.
If the above results linking protein quality, protein
quantity, and D15
N are confirmed or refined by further
studies, estimating nitrogen discriminations for omnivores
without detailed knowledge of the animal factors may not
limit accurate estimates of assimilated diet. However,
Fig. 6. Illustration of assimilated diet estimates for an omnivore consuming a two-component diet (plants and animals) using either linear (no
dietary interaction) or curvilinear (metabolically mixed diet with complementary amino acid profiles) solutions. The assumptions were that:
(i) the plant component of the diet had a d15
N signature of )1Æ0&, a protein quality of 1Æ4% methionine and a protein content of 24%, which gave
a discrimination estimate of 5Æ0& (eqn 1); and (ii) the animal component had a d15
N signature of 4Æ0&, a protein quality of 2Æ5% methionine and
a protein content of 77%, which gave a discrimination estimate of 3Æ5&. Intermediate discriminations for the metabolically mixed diets were
determined by solving eqn 1 for various dietary mixtures. Because the discriminations at a given dietary mixture were lower when the two foods
were consumed in a metabolically mixed diet than when there was no dietary interaction, the linear model underestimates the importance of
animal matter and overestimates the importance of plant matter in the diet when the foods were consumed in a metabolically mixed diet. The
maximum error in the assimilated diet estimates for each dietary component in this example was 14%.

current limitations to this approach include the lack of: (i)
a broad understanding of amino acid profiles in the wide
range of foods consumed by wild animals and the time
course of their metabolic interaction within the consumer
that will determine if they are complementary or non-
complementary; (ii) an understanding of how mixtures of
protein and non-protein nitrogen (e.g. chitin) in insects,
crustacea, and fungi determine nitrogen discrimination;
and (iii) mixing models in which the discriminations of the
individual foods vary from being independent, additive,
and linear for foods consumed in metabolically distinct
meals to dependent and curvilinear when foods with com-
plementary amino acid profiles are consumed in metaboli-
cally mixed diets (DeGabriel, Foley & Wallis 2002;
Robbins, Felicetti & Florin 2010) (Fig. 6).
This latter point means that discriminations may need to
be predicted parameters in mixing models based on addi-
tional animal, dietary, and temporal inputs rather than the
current a priori estimates. However, investigators working
with piscivores, other carnivores, and herbivores may have a
much easier task in estimating discriminations as many of
these groups do not consume foods that vary as extensively in
protein quality and quantity as do the diets consumed by
some omnivores. Although there may be other diets with
nitrogen discriminations that are outside the bounds of our
current understanding, the equations developed in this study
offer the opportunity to begin developing more complex and
realistic mixing models for omnivores that more accurately
estimate assimilated diets.
Acknowledgements
The project was approved by the Washington State University Institutional
Animal Care and Use Committee (#03762) and funded by the Nutritional Ecol-
ogy Research Endowment and the US Fish and Wildlife Service.
References
Arneson, L.S. & MacAvoy, S.E. (2005) Carbon, nitrogen, and sulfur diet-
tissue discrimination in mouse tissues. Canadian Journal of Zoology, 83,
989–995.
Bearhop, S., Waldron, S., Votier, S.C. & Furness, R.W. (2002) Factors that
influence assimilation rates and fractionation of nitrogen and carbon stable
isotopes in avian blood and feathers. Physiological and Biochemical Zoology,
75, 451–458.
Ben-David, M. & Schell, D.M. (2001) Mixing models in analyses of diet using
multiple stable isotopes: a response. Oecologia, 127, 180–184.
Caut, S., Angulo, E. & Courchamp, F. (2008) Discrimination factors (D15
N
and D13
C) in an omnivorous consumer: effect of diet isotopic ratio. Func-
tional Ecology, 22, 255–263.
Caut, S., Angulo, E. & Courchamp, F. (2009) Variation in discrimination fac-
tors (D15
N and D13
C): the effect of diet isotopic values and applications for
diet reconstruction. Journal of Applied Ecology, 46, 443–453.
Cherel, Y., Hobson, K.A. & Hassani, S. (2005) Isotopic discrimination
between food and blood and feathers of captive penguins: implications
for dietary studies in the wild. Physiological and Biochemical Zoology, 78,
106–115.
Claridge, A.W., Trappe, J.M., Cork, S.J. & Claridge, D.L. (1999) Mycophagy
by small mammals in the coniferous forests of North America: nutritional
value of sporocarps of Rhizopogon vinicolor, a common hypogeous fungus.
Journal of Comparative Physiology, 169B, 172–178.
Darr, R.L. & Hewitt, D.G. (2008) Stable isotope trophic shifts in white-tailed
deer. Journal of Wildlife Management, 72, 1525–1531.
Davis, T.A., Nguyen, H.V., Garcia-Bravo, R., Fiorotto, M.L., Jackson,
E.M., Lewis, D.S., Lee, D.R. & Reeds, P.J. (1994) Amino acid compo-
sition of human milk is not unique. Journal of Nutrition, 124, 1126–
1132.
DeGabriel, J., Foley, W.J. & Wallis, I.R. (2002) The effect of excesses and defi-
ciencies in amino acids on the feeding behavior of the common brushtail pos-
sum. Journal of Comparative Physiology, 172B, 607–617.
Fantle, M.S., Dittel, A.I., Schwalm, S.M., Epifanio, C.E. & Fogel, M.L. (1999)
A food web analysis of the juvenile blue crab, Callinectes sapidus, using sta-
ble isotopes in whole animals and individual amino acids. Oecologia, 120,
416–426.
Felicetti, L.A., Schwartz, C.C., Rye, R.O., Haroldson, M.A., Gunther, K.A.,
Phillips, D.L. & Robbins, C.T. (2003) Use of sulfur and nitrogen stable iso-
topes to determine the importance of whitebark pine nuts to Yellowstone
grizzly bears. Canadian Journal of Zoology, 81, 763–770.
Gaye-Siessegger, J., Focken, U., Muetzel, S., Abel, H. & Becker, K. (2004)
Feeding level and metabolic rate affect d13
C and d15
N values in carp: impli-
cations for food web studies. Oecologia, 138, 175–183.
Gaye-Siessegger, J., Focken, U., Abel, H. & Becker, K. (2007) Starvation and
low feeding levels result in an enrichment of C-13 in lipids and N-15 in pro-
tein of Nile tilapia Oreochromis niloticus L. Journal of Fish Biology, 71,
90–100.
Hilderbrand, G.V., Farley, S.D., Robbins, C.T., Hanley, T.A., Titus, K. &
Servheen, C. (1996) Use of stable isotopes to determine diets of living and
extinct bears. Canadian Journal of Zoology, 74, 2080–2088.
Hobson, K.A. & Clark, R.G. (1992) Assessing avian diets using stable iso-
topes II: factors influencing diet-tissue fractionation. The Condor, 94,
189–197.
Hobson, K.A., Schell, D.M., Renouf, D. & Noseworthy, E. (1996) Stable
carbon and nitrogen isotopic fractionation between diet and tissues
of captive seals: implication for dietary reconstructions involving
marine mammals. Canadian Journal of Fish and Aquatic Science, 52,
528–533.
Jenkins, S.G., Partridge, S.T., Stephenson, T.R., Farley, S.D. & Robbins, C.T.
(2001) Nitrogen and carbon isotope fractionation between mothers, neo-
nates, and nursing offspring. Oecologia, 129, 336–341.
Kelly, L.J. & Martinez del Rio, C. (2010) The fate of carbon in growing fish: an
experimental study of isotopic routing. Physiological and Biochemical Zool-
ogy, 83, 473–480.
Kleinbaum, D.G. & Kupper, L.L. (1978) Applied Regression Analysis and Other
Multivariable Methods. Duxbury Press, Massachusetts.
Lesage, V., Hammill, M.O. & Kavacs, K.M. (2002) Diet-tissue fractionation of
stable carbon and nitrogen isotopes in phocid seals. Marine Mammal Sci-
ence, 18, 182–193.
Martinez del Rio, C., Wolf, N., Carleton, S.A. & Gannes, L.Z. (2009) Isotopic
ecology ten years after a call for more laboratory experiments. Biological
Reviews, 84, 91–111.
McCutchan, J.H., Jr, Lewis, W.M., Jr, Kendall, C. & McGrath, C.C. (2003)
Variation in trophic shift for stable isotopic ratios of carbon, nitrogen, and
sulfur. Oikos, 102, 378–390.
Miron, M.L.L., Herrera, M.L.G., Ramirez, P.N. & Hobson, K.A. (2006) Effect
of diet quality on carbon and nitrogen turnover and isotopic discrimination
in blood of a New World nectarivorous bat. Journal of Experimental Biol-
ogy, 209, 541–548.
NRC (National Research Council) (1994) Nutrient Requirements of Poultry.
National Academy of Sciences, Washington.
NRC (National Research Council) (1995) Nutrient Requirements of Laboratory
Animals. National Academy of Sciences, Washington.
Ogden, L.J.E., Hobson, K.A. & Lank, D.B. (2004) Blood isotopic (d13
C and
d15
N) turnover and diet-tissue fractionation factors in captive dunlin (Calid-
ris alpine pacifica). The Auk, 121, 170–177.
Pearson, S.F., Levey, D.J., Greenberg, C.H. & Martinez del Rio, C. (2003)
Effects of elemental composition on the incorporation of dietary nitrogen
and carbon isotopic signatures in an omnivorous songbird. Oecologia, 135,
516–523.
Peterson, B.J. & Fry, B. (1987) Stable isotopes in ecosystem studies. Annual
Review of Ecology and Systematics, 18, 293–320.
Podlesak, D.W. & McWilliams, S.R. (2006) Metabolic routing of dietary
nutrients in birds: effects of diet quality and macronutrient composition
revealed using stable isotopes. Physiological and Biochemical Zoology, 79,
534–549.
Robbins, C.T. (1993) Wildlife Feeding and Nutrition. Academic Press, New
York.

Robbins, C.T., Felicetti, L.A. & Florin, S.T. (2010) The impact of protein qual-
ity on stable nitrogen isotope ratio discrimination and assimilated diet esti-
mation. Oecologia, 162, 571–579.
Robbins, C.T., Felicetti, L.A. & Sponheimer, M. (2005) The effect of dietary
protein quality on nitrogen isotope discrimination in mammals and birds.
Oecologia, 144, 534–540.
Roth, J.D. & Hobson, K.A. (2000) Stable carbon and nitrogen isotopic frac-
tionation between diet and tissue of captive red fox: implications for dietary
consideration. Canadian Journal of Zoology, 78, 848–852.
SAS (1998) SAS ⁄ SAT User’s Guide, Version 6.12. SAS Institute, Cary.
Smith, J.J., Millar, J.S., Longstaffe, F.J. & Boonstra, R. (2010) The effect of
metabolic rate on stable carbon and nitrogen isotope compositions in deer
mice, Peromyscus maniculatus. Canadian Journal of Zoology, 88, 36–42.
Sponheimer, M., Robinson, T., Ayliffe, L., Roeder, B., Hammer, J., Passey, B.,
West, A., Cerling, T., Dearing, D. & Ehleringer, J. (2003) Nitrogen isotopes
in mammalian herbivores: hair d15N values from a controlled feeding study.
International Journal of Osteoarchaeology, 13, 80–87.
van Tets, I.G. & Hulbert, A.J. (1999) A comparison of the nitrogen require-
ments of the Eastern pygmy possum, Cercartetus nanus, on a pollen and on a
mealworm diet. Physiological and Biochemical Zoology, 72, 127–137.
Tsahar, E., Wolf, N., Izhaki, I., Arad, Z. & Martinez del Rio, C. (2008) Dietary
protein influences the rate of 15N incorporation in blood cells and plasma of
Yellow-vented bulbuls (Pycnonotus xanthopygos). Journal of Experimental
Biology, 211, 459–465.
Vanderklift, M.A. & Ponsard, S. (2003) Sources of variation in consumer-diet
d15
N enrichment: a meta-analysis. Oecologia, 136, 169–182.
Received 15 June 2010; accepted 16 September 2010
Handling Editor: Adam Kay
Supporting Information
Additional Supporting Information may be found in the online
version of this article.
Table S1. The essential amino acid requirement for growth by labora-
tory rats (NRC 1995) relative to the same amino acids occurring in
various foods used in the current and previous isotope studies (Hil-
derbrand et al. 1996; Darr & Hewitt 2008; Tsahar et al. 2008; Rob-
bins, Felicetti & Florin 2010). Both the requirement and dietary
amino acid profile are expressed as a per cent of total dietary protein.
The subscripts in the diet columns indicate the deficit occurring for
each amino acid in each food relative to the requirement.
Table S2. Estimated or measured dietary protein characteristics and
nitrogen discrimination of whole blood, plasma, serum, or red blood
cells for the diets fed to a wide range of mammals and birds. Protein
quality estimates are from Table S1 (Supporting information) and
based on laboratory rat amino acid requirements for growth. The
numbers following several mixed diets (e.g. 75:25 or 50:50) refer to the
relative contribution of protein by the two ingredients (Robbins, Fe-
licetti & Florin 2010; current study). The ‘diluted’ diets refer to the
addition of sucrose, starch and corn oil to reduce protein content by
50% while holding protein quality constant.
Table S3. Sulphur isotope values for the diets and plasma of labo-
ratory rats and grizzly bears and the corresponding discrimination
(Felicetti et al. 2003; Robbins, Felicetti & Florin 2010; current
study).
As a service to our authors and readers, this journal provides sup-
porting information supplied by the authors. Such materials may be
re-organized for online delivery, but are not copy-edited or typeset.
Technical support issues arising from supporting information (other
than missing files) should be addressed to the authors.

Florin_et_al-2011-Functional_Ecology

More Related Content

What's hot

Viewers also liked

Similar to Florin_et_al-2011-Functional_Ecology

Florin_et_al-2011-Functional_Ecology