Nutritional geometry and macronutrient variation in the diets of gannets: the challenges in marine field studies

Abstract

Foraging theory proposes that the nutritional driver of food choice and foraging in carnivores is energy gain. In contrast, recent laboratory experiments have shown that several species of carnivore select prey that provides a diet with a specific balance of macronutrients, rather than the highest energy content. It remains, however, to be determined how nutritionally variable the foods of predators in the wild are, and whether they feed selectively from available prey to balance their diet. Here, we used a geometric method named the right-angled mixture triangle (RMT) for examining nutritional variability in the prey and selected diets of a group of wild carnivores and marine top predators, the gannets (Morus spp.). A prey-level diet analysis was performed on Australasian gannets (M. serrator) from two New Zealand locations, and the macronutrient composition of their chosen prey species was measured. We use RMT to extend the comparison in the compositions of foods and diets from Australasian gannets from Australia as well as Northern gannets (M. bassanus) and Cape gannets (M. capensis). We found nutritional variability at multiple scales: intra- and interspecific variability in the pelagic fish and squid prey themselves; and intra- and interspecific variability in the diets consumed by geographically disparate populations of gannets. This nutritional variability potentially presents these predatory seabirds with both opportunity to select an optimal diet, and constraint if prevented from securing an optimal diet.

Full text

1 23
Marine Biology
International Journal on Life in Oceans
and Coastal Waters
ISSN 0025-3162
Volume 161
Number 12
Mar Biol (2014) 161:2791-2801
DOI 10.1007/s00227-014-2544-1
Nutritional geometry and macronutrient
variation in the diets of gannets: the
challenges in marine field studies
Alice H.Tait, David Raubenheimer,
Karen A.Stockin, Monika Merriman &
Gabriel E.Machovsky-Capuska

1 23
Your article is protected by copyright and
all rights are held exclusively by Springer-
Verlag Berlin Heidelberg. This e-offprint is
for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 3
Mar Biol (2014) 161:2791–2801
DOI 10.1007/s00227-014-2544-1
ORIGINAL PAPER
Nutritional geometry and macronutrient variation in the diets
of gannets: the challenges in marine field studies
Alice H. Tait · David Raubenheimer · Karen A. Stockin ·
Monika Merriman · Gabriel E. Machovsky‑Capuska
Received: 26 March 2014 / Accepted: 12 September 2014 / Published online: 1 October 2014
© Springer-Verlag Berlin Heidelberg 2014
(M. capensis). We found nutritional variability at multiple
scales: intra- and interspecific variability in the pelagic fish
and squid prey themselves; and intra- and interspecific vari-
ability in the diets consumed by geographically disparate
populations of gannets. This nutritional variability poten-
tially presents these predatory seabirds with both opportu-
nity to select an optimal diet, and constraint if prevented
from securing an optimal diet.
Introduction
It has long been known that herbivores and omnivores
feed on diets that are variable in the ratios of macronutri-
ents (Westoby 1974, 1978) and consequently have evolved
mechanisms for balancing their diet through nutrient-spe-
cific foraging (Chambers et al. 1995; Raubenheimer and
Jones 1996; Rothman et al. 2011). In contrast, it is widely
believed that predators feed on foods that are nutritionally
balanced and therefore have no need to select nutritionally
complementary prey to balance their diet (Westoby 1978;
Stephens and Krebs 1986; Galef 1996; Fryxell and Lund-
berg 1997). Recent studies in the laboratory have shown,
however, that predatory invertebrates (spiders, beetles)
and vertebrates (mink, fish) do feed selectively for prey
that contain specific ratios of macronutrients (Mayntz
et al. 2005, 2009; Raubenheimer et al. 2007; Rubio et al.
2008, 2009). Yet, it remains to be determined how variable
macronutrients in foods of predators in the wild are, and
whether they feed selectively from available prey to bal-
ance their diet (Wilder and Eubanks 2010). This knowledge
is fundamental to our understanding of how ecosystems
work, given the key role that predators play in structuring
ecological communities (Polis et al. 1989; Polis and Holt
1992; Raubenheimer et al. 2007).
Abstract Foraging theory proposes that the nutritional
driver of food choice and foraging in carnivores is energy
gain. In contrast, recent laboratory experiments have shown
that several species of carnivore select prey that provides a
diet with a specific balance of macronutrients, rather than
the highest energy content. It remains, however, to be deter-
mined how nutritionally variable the foods of predators in
the wild are, and whether they feed selectively from avail-
able prey to balance their diet. Here, we used a geometric
method named the right-angled mixture triangle (RMT) for
examining nutritional variability in the prey and selected
diets of a group of wild carnivores and marine top preda-
tors, the gannets (Morus spp.). A prey-level diet analysis
was performed on Australasian gannets (M. serrator) from
two New Zealand locations, and the macronutrient com-
position of their chosen prey species was measured. We
use RMT to extend the comparison in the compositions of
foods and diets from Australasian gannets from Australia as
well as Northern gannets (M. bassanus) and Cape gannets
Communicated by Y. Cherel.
Electronic supplementary material The online version of this
article (doi:10.1007/s00227-014-2544-1) contains supplementary
material, which is available to authorized users.
A. H. Tait · K. A. Stockin · M. Merriman ·
G. E. Machovsky-Capuska
Institute of Natural and Mathematical Sciences,
Massey University, North Shore MSC, Private Bag 102 904,
Auckland, New Zealand
D. Raubenheimer · G. E. Machovsky-Capuska (*)
Faculty of Veterinary Science, The Charles Perkins Centre,
School of Biological Sciences, University of Sydney,
Sydney, Australia
e-mail: g.machovsky@sydney.edu.au
Author's personal copy

2792 Mar Biol (2014) 161:2791–2801
1 3
A factor that has limited objective tests of the variability
of prey composition is the common assumption that asso-
ciates prey quality with the levels in foods of one particu-
lar component, usually energy (Stephens and Krebs 1986).
Several studies have estimated energetic dietary demands
for different seabird species using energy values of prey
(Furness 1978; Ellis 1984; Annett and Pierotti 1999; Wan-
less et al. 2005; Grémillet et al. 2008; Votier et al. 2010;
Machovsky-Capuska et al. 2011a). Recent work involv-
ing herbivores, omnivores and carnivores spanning a wide
range of taxa, both in laboratory and field studies, has dem-
onstrated the importance of evaluating food quality in rela-
tion to the ratios of several nutrients, rather than the lev-
els of any one (Simpson and Raubenheimer 2012). Until
recently, however, studies applying this multidimensional
view of nutrition to ecological questions such as variation
in prey quality have been limited by the lack of a frame-
work for characterising food compositions in multiple
dimensions and relating them to pertinent factors such as
predator identity, geographic location and diet composition.
An approach for this, called the right-angled mixture trian-
gle (RMT), was recommended by Raubenheimer (2011)
(Fig. 1a, b).
Our primary aims in this paper were to apply RMT
in an initial examination of the variability across scales
of the macronutrient composition of prey of a group of
marine top predators, gannets (Morus spp.), and introduce
this approach as a framework for addressing such ques-
tions. Seabirds, including gannets, are known to forage
for patchily distributed foods in foraging trips that can
span hundreds of kilometres over several days (Hamer
et al. 2000; Richoux et al. 2010; Machovsky-Capuska
et al. 2013, 2014). Gannets are highly specialised marine
predators that feed mainly on pelagic fish and squid at the
air–water interface (Robertson 1992; Machovsky-Capuska
et al. 2011b; Schuckard et al. 2012). We were interested in
estimating the macronutrient variability among different
individuals of the same prey species, across different prey
species, across the diets (i.e. combined prey selected by
individual gannets) within and between colonies at differ-
ent degrees of geographic separation and between species
of gannets.
Fig. 1 Right-angled mixture triangles provide a means to plot three
components in 2D graphs. In a each point represents a mixture of
protein (P), lipid (L) and carbohydrate (C). By convention the third,
implicit, variable (in this case carbohydrate) is denoted in square
brackets (Raubenheimer 2011). % P and L increase in the normal
way along the X‐axis and Y-axis, respectively, and the P:L balance of
a mixture is given by the slope of the radial that connects the point
to the origin. % C of a point is determined as the difference between
100 % and the value at which a negative through the point intersects
with the two axes. For example, in a point i contains 60 % P, 20 % L
and 20 % C, with a 3:1 P:C ratio. Point ii has the same % C (20 %),
but a lower P:L (1:3) than i, and point iii contains the same P:L ratio
but higher % C (60 %) than ii. b In this case the third, implicit, vari-
able plotted on the I-axis is 100 − (% protein + % lipid). This repre-
sents all components of the diet that effectively dilute macronutrient
concentration (here assuming that carbohydrate content is minimal,
as is the case for many predator diets). Since dilution of macronu-
trients is the inverse of energy concentration, this model integrates
lipid, protein, lipid/protein balance and total energy content in the
diets of gannets. The plot illustrates the fact that if an animal mixes
its intake from two foods (e.g. foods iv, v), the composition of the
diet is constrained to lie somewhere on the line that connects the two
foods (e.g. point x lies on dashed line connecting points iv, v). If its
intake is derived from three foods (e.g. v–vii), then the resulting diet
composition is constrained to lie within the triangle connecting these
foods (e.g. point y)
▸
Author's personal copy

2793Mar Biol (2014) 161:2791–2801
1 3
To achieve this range of scales in our comparison,
we have combined our original data with relevant data
extracted from the literature. In so doing, it became appar-
ent that data are often not collected in a way that facilitates
comparative analyses of diet variability across scales. Our
secondary aim is thus to recommend strategies for data col-
lection that will facilitate data pooling for the analysis of
broad-scale ecological questions concerning food quality.
We show that a range of significant questions in nutritional
ecology can be addressed if appropriate relevant data are
collected for a given system.
Materials and methods
The right-angled mixture triangle
Right-angled mixture triangles enable mixtures of three
components to be graphed in a 2D plot (Fig. 1a, b)
(Raubenheimer 2011). Data are prepared for plotting by
summing the percentages of the three components (e.g. the
macronutrients protein, fat and carbohydrate) in the parent
mixture (e.g. the food) and then expressing each compo-
nent as a percentage of the sum of all three. For example,
if protein (P), lipid (L) and carbohydrate (C) were present
in a food at 10, 20 and 30 % by weight, respectively, then
each would be expressed as follows:
To construct an RMT to display this mixture, % P is
plotted against % L. Because the three nutrients in the mix-
ture sum to 100 %, plotting % P on % L will automatically
result in the point positioning to also reflect the value of %
C. Such plots not only provide a visual representation of a
3D mixture, but also enable various parameters of differ-
ent mixtures to be compared and inter-related (Fig. 1a). A
particularly useful feature of RMTs is their use for mod-
elling meta-mixtures (mixtures of mixtures), for example,
the nutritional composition of diets that result from mixing
multiple foods (Fig. 1b).
Source of data
An examination of the macronutrient variability of the
prey of gannets (or any marine predator) at several scales
involves (1) identifying the foods eaten by individual gan-
nets; (2) collecting representative samples of the prey spe-
cies in the diet over the same temporal and spatial scale as
the gannet diet analysis was conducted; (3) measuring the
macronutrient content of the prey samples using proximate
analysis; (4) examining the data using RMTs; and (5) using
appropriate statistics to test hypotheses of interest. In the
P
=
10/(10
+
20
+
30)%
=
10/60 %
=
16.7 %;
L=20
/
60 =33.3 %;C=30
/
60 % =50 %
present study, a complete data set was collected and ana-
lysed for Australasian gannets (M. serrator) from two New
Zealand locations—Hauraki Gulf (HG), North Island and
Farewell Spit (FS), South Island. We also combined pre-
viously published partial data sets to enable a comparison
between colonies within and between other gannet species,
i.e. data on the diet (prey species) consumed by Australa-
sian gannets from Australia, as well as Northern gannets
(M. bassanus) and Cape gannets (M. capensis) from differ-
ent locations were combined with other published data on
the proximate composition of the relevant prey species.
Prey composition of diet
The diet of Australasian gannets was measured using gut
contents of carcasses opportunistically collected from the
waters of HG (36°51′S, 174°46′E) on the East Coast of the
North Island of New Zealand, and regurgitations collected
from individuals from the FS gannetry (40°33′S, 173°02′E)
on the West Coast of the South Island of New Zealand. In
HG, carcasses were collected between August 2010 and
January 2011. HG is a shallow (60 m maximum depth),
semi-enclosed body of temperate water (Manighetti and
Carter 1999) that exhibits a high diversity of marine fauna,
including four gannet colonies with an estimated popula-
tion of 7,000 breeding pairs (Nelson 2005). The carcasses
were typically stored frozen and later examined post-
mortem following avian necropsy protocols (Work 2000).
Individual prey items were removed from the oral cavity,
oesophagus and stomach. Prey items that were ingested
ante-mortem were individually weighed to 0.1 g, and stom-
ach contents were washed through a 0.25-mm-mesh sieve
to examine for otoliths and cephalopod beaks (Wingham
1985; Duffy and Jackson 1986). Digestion codes were
assigned to retrieved prey items (following Meynier et al.
2008), and prey species were identified to the lowest pos-
sible taxonomic level using published guides (Paulin et al.
1989). In FS, regurgitations were collected following
Wingham (1985) from individuals in January 2011 during
the chick-rearing period. The FS gannetry in Golden Bay
was established in 1983 with c.75 breeding pairs and is one
of four breeding sites in the South Island (Hawkins 1988).
Since then, the population has increased by an average of
11.5 % per annum, to an estimated at 3,900 pairs in 2011
(Schuckard et al. 2012). Prey items from the regurgitations
were processed as described above for HG samples.
Macronutrient composition of prey
Samples of the prey species found in the diets of gannets
from HG and FS were selected from undigested material
collected from the carcasses and regurgitations. In the HG,
samples of all prey species represented in the diet were
Author's personal copy

2794 Mar Biol (2014) 161:2791–2801
1 3
collected including anchovy (Engraulis australis, n = 5),
jack mackerel (Trachurus novaezelandiae, n = 6), kahawai
(Arripis trutta, n = 5), pilchard (Sardinops neopilchardus,
n = 6), yellow-eye mullet (Aldrichetta forsteri, n = 5) and
arrow squid (Nototodarus spp., n = 5). In FS, samples of
prey species comprising 82.6 % of the mass of the diet
were collected including kahawai (n = 4), pilchard (n = 7)
and arrow squid (n = 3). This enabled us to measure the
proximate composition of 20 out of the 24 regurgitations
(the remaining four regurgitations contained prey species
that were not collected for analysis). All prey samples were
frozen within five hours of collection and stored at −20 °C
until proximate composition analysis.
Prior to analysis, each sample was partially thawed and
weighed to 0.1 g, dried overnight in a convection oven at
60 °C and ground in a coffee grinder. Total nitrogen was
measured by Kjeldahl analysis and crude protein estimated
by multiplying N by a factor of 6.25 (AOAC 981.10, AOAC
2005). Total lipid (ether extract) was measured by the
Mojonnier method (AOAC 954.02). Moisture was meas-
ured by drying the sample in a convection oven at 125 °C
(AOAC 950.46) and combining this moisture loss with ini-
tial loss from the overnight dry down. Ash was measured
by ignition in a furnace at 550 °C (AOAC 920.153).
Data analysis
For each bird, the total weight of prey in the anterior gut
or regurgitation was calculated as the sum of the individual
prey items (Duffy and Jackson 1986) and each prey spe-
cies was assigned a mass percentage (M%), calculated as
the percentage of total prey weight that the species contrib-
uted to the overall diet (Duffy and Jackson 1986; Schuck-
ard et al. 2012). For each location, each prey species was
assigned a numerical abundance percentage (N%) calcu-
lated as the percentage of the total number of prey items
contributed by individuals of a particular species, and a
frequency of occurrence percentage (F%) calculated as the
percentage of birds that had a particular species in their
diet. Data on the proximate composition of prey were ana-
lysed using Kruskal–Wallis and Mann–Whitney U tests to
determine whether there were differences between species
within a location or between locations within a species,
respectively, using SPSS (PASW Statistics 19, IBM Corp.,
Somers, NY, USA). RMTs (Raubenheimer 2011) were
used to explore the relationships among the proportional
content of nutrients in the prey and diets. The nutritional
composition of the prey and diets was examined in terms of
protein (X-axis), lipid (Y-axis) and rest, i.e. 100 − (% pro-
tein + % lipid) (I-axis) content, each expressed as a per-
centage of wet weight. This model integrates lipid, protein
and the overall content (the concentration of macronutri-
ents, as represented by I-axis) as opposed to the common
approach of expressing prey composition only by their
caloric content (Raubenheimer 2011).
Reanalysis of published data
To place in broader context our original data collected
from Australasian gannets in HG and FS in New Zea-
land, we also reanalysed published data on the diet of
Australasian gannets from Port Phillip Bay (PPB), Victo-
ria, Australia (Bunce 2001), the diet of Northern gannets
from Funk Island (FI), Newfoundland, Canada (49°45′N,
53°11′W) (Montevecchi et al. 1984; Garthe et al. 2011) and
Bass Rock (BR), southeast Scotland (Wanless et al. 2005;
Hamer et al. 2007), and the diet of Cape gannets from Mal-
gas Island (MI) (33°03′S, 17°55′E) on the Western Cape
of South Africa and Bird Island (BI) (33°50′S, 26°17′E),
on the Eastern Cape of South Africa (Adams et al. 1991;
Pichegru et al. 2007). For each species/location, data
on the prey composition of gannets’ diet and data on the
proximate composition of prey species were necessarily
obtained from different publications, as no one publication
contained both data sets (see Supporting Information S1).
An effort was made to obtain prey proximate composition
data from samples collected at similar spatiotemporal prox-
imity to the study. These data enabled us to calculate the
proximate composition of prey species constituting (as a
percentage of mass) 79 % of the diet of Australasian gan-
nets from PPB, 99.6 and 96.7 % of the diet of Northern
gannets from FI and BR, respectively, and 100 and 83 % of
the diet of Cape gannets from MI and BI, respectively. The
data were plotted on RMTs as described previously for the
New Zealand populations (see Supporting Information S1).
Results
Prey composition of diets
Of the 35 carcasses examined from HG, 30 contained iden-
tifiable prey remains in their anterior gut. The median mass
of prey per carcass was 215.6 g (range 176.2–426.0 g). The
251 individual prey items identified from these 30 samples
included five species of fish and one species of squid, where
71 % of carcasses contained two prey species, 23 % con-
tained one prey species and 6 % contained three prey spe-
cies. Pilchard had the highest frequency of occurrence and
the greatest mass percentage, followed by anchovy, though
anchovy had greater numerical abundance than pilchard
(Table 1). Twenty-four regurgitations were collected from
different individuals from the FS gannetry. The median
mass of prey per regurgitation was 190.0 g (range 100.0–
430.0 g), which equated to an average daily food intake of
9.4 % of body weight (Nelson 2005). The 134 individual
Author's personal copy

2795Mar Biol (2014) 161:2791–2801
1 3
prey items identified from these 24 samples included seven
species of fish and one species of squid, where 75 % of
regurgitations contained one prey species, 21 % contained
two prey species and 4 % contained three prey species. Pil-
chard had the highest frequency of occurrence, the great-
est mass percentage and numerical abundance. Arrow squid
had the second highest frequency of occurrence and the
second greatest mass percentage, while anchovy had the
second greatest numerical abundance (Table 1).
Macronutrient composition of prey and diets
Within location, prey species collected from HG differed
in their proportion of protein (Kruskal–Wallis, H = 27.39,
df = 5, p < 0.001), lipid (Kruskal–Wallis, H = 26.14,
df = 5, p < 0.001) and rest (Kruskal–Wallis, H = 22.91,
df = 5, p < 0.001), and those collected from FS differed in
their proportion of lipid (Kruskal–Wallis, H = 6.54, df = 5,
p < 0.038). Of the three prey species found in the diet of
both the HG and FS gannet populations whose macronutri-
ent composition was measured, two species showed differ-
ences between locations. Compared to collections from FS,
pilchard collected from HG had lower proportions of pro-
tein (Mann–Whitney, U = 0.01, Z = −3.00, p = 0.003) and
lipid (Mann–Whitney, U = 0.01, Z = −3.00, p = 0.003)
and a higher proportion of rest (Mann–Whitney, U = 0.01,
Z = −3.00, p = 0.003) and kahawai collected from HG had
a higher proportion of lipid (Mann–Whitney, U = 0.00,
Z = −2.44, p = 0.014) and a higher proportion of rest
(Mann–Whitney, U = 0.00, Z = −2.44, p = 0.014). There
was no difference between locations in the macronutrient
composition of arrow squid. RMTs depict these differences
within and between prey species with regard to the balance
of protein to lipid and the level to which these energetic
macronutrients are diluted by other dietary constituents
(rest including moisture, inorganic matter etc.) (Fig. 2a).
The macronutrient composition of gannets’ diets differed
between locations, where gannets from HG had a diet that
contained a lower proportion of protein (Mann–Whitney
U = 29.00, Z = −4.65, p < 0.001) and rest (Mann–Whitney
U = 117.00, Z = −2.55, p < 0.05) but did not differ in the
proportion of lipid (Mann–Whitney U = 204.00, Z = −0.48,
p = 0.633) to the diet of gannets from FS. In plotting the
nutrient space accessible to each gannet population, there was
no overlap between the two spaces implying a lack of oppor-
tunity for the two populations to secure the same balance
of nutrients in their diet (Fig. 2b). The nutrient space acces-
sible to the FS gannets is potentially larger than that shown
(17.4 % of the mass of the diet was not analysed for nutrient
content). However, the entire nutrient content of 20 out of the
24 regurgitations was calculated, so the average composition
of all 24 regurgitations from FS gannets is unlikely to be very
different from that of the 20 regurgitations shown here.
Reanalysis of published data
By combining partial data sets from the literature, we were
able to construct RMTs to investigate differences between
populations within a gannet species in the accessible nutri-
ent space and selected diet, as well as to compare diet
between different gannet species.
For Australasian gannets, we compared our data from the
HG and FS populations in New Zealand with data from the
PPB population in Australia (Fig. 3a). While there was no
overlap between the New Zealand populations in nutrient
space, the nutrient space accessible to the PPB population
overlapped with the HG population. Despite this, the diet
consumed by the PPB population fell outside of the nutrient
space of HG population, comprising a greater proportion of
protein (similar to that of the FS population), a much greater
proportion of lipid and a greater proportion of protein and
lipid combined than either of the New Zealand populations.
For Northern gannets, we compared data from popu-
lations from FI, Canada and BR, Scotland (Fig. 3b). The
nutrient space accessible to each of the two populations
shows an area of overlap, and the average diet consumed by
birds of each population both fall within this area of over-
lap and contain a remarkably similar nutrient balance.
Table 1 Composition of the
diet of the Australasian gannet
as reflected by the analysis of
35 carcasses collected from HG
and 24 regurgitations collected
during the chick-rearing period
at FS, New Zealand
Diet is described by percentage
frequency of occurrence (F%),
mass (M%) and numerical
abundance (N%)
Prey species HG FS
M%F%N%M%F%N%
Pilchard 37.3 76.7 37.8 62.1 66.7 84.1
Anchovy 28.1 66.7 56.6 0.8 12.5 5.8
Yellow-eye mullet 15.8 16.7 2.0 – – –
Kahawai 12.5 13.3 1.6 8.3 8.3 1.4
Jack mackerel 4.8 6.7 1.6 3.6 4.2 0.7
Arrow squid 1.5 3.3 0.4 12.1 16.7 3.6
Yellowtail kingfish – – – 8.7 12.5 2.2
Barracouta – – – 4.2 4.2 0.7
Garfish – – – 0.3 4.2 1.4
Author's personal copy

2796 Mar Biol (2014) 161:2791–2801
1 3
For Cape gannets, we compared data from MI and BI on
the Western and Eastern Cape of South Africa, respectively
(Fig. 3c). We were unable to separate the nutrient spaces
accessible to these two populations, as their diets con-
sisted of similar prey species and the sites were too close
geographically for differences within prey species between
locations to be identified. However, the composition of the
average diet consumed by the two populations was mark-
edly different with the BI population having a greater pro-
portion of both lipid and protein and lipid combined than
the MI population.
In comparing the diet of different gannet species, we
plotted species as a function of population (location)
because there were differences between populations within
a species. The data suggest that Australasian and Cape
gannets consume diets containing a greater proportion of
protein than Northern gannets, while Australasian gannets
from PPB, Northern gannets and Cape gannets from BI
consume diets containing a greater proportion of lipid than
Australasian gannets from New Zealand or Cape gannets
from MI. Australasian gannets from PPB and Cape gannets
from BI consume diets with a greater proportion of protein
and lipid combined than the other gannets (Fig. 4).
Discussion
Our study reveals variability in the macronutrient compo-
sition of gannets’ prey on multiple scales. First, we found
nutritional variability among the prey themselves, poten-
tially presenting gannets with both opportunity to select an
optimal diet, and constraint if prevented from securing food
combinations that support an optimal diet. Second, nutri-
tional variability was established among the diets of gannets,
both within and between species. While diversity in nutrient
gain for this top marine predator may reflect differences in
the nutrient requirements of seabirds living in different geo-
graphic locations, there is also evidence to suggest that in
some cases seabirds are prevented from securing an optimal
balance of nutrients, resulting in lower fitness and breeding
success (e.g. Pichegru et al. 2007; Grémillet et al. 2008).
Nutritional variability of gannet prey
It has been suggested that marine prey species can vary in
their macronutrient composition (Lenky et al. 2012). We
found significant differences in the proportion of protein,
lipid and rest in the prey of Australasian gannets at three
Fig. 2 Right-angled mixture triangle showing the composition in
terms of protein, lipid and remaining fresh weight of foods and diets
of Australasian gannets from HG (circles) and FS (triangles), New
Zealand. a Variation in composition of prey items within and across
species. Markers represent individual prey samples, with colour
distinguishing species (grey anchovy, green jack mackerel, purple
kahawai, red pilchard, aqua arrow squid, yellow yellow-eye mullet).
b Variation in the composition of the diets of gannets from HG and
FS. Solid black symbols show the average composition of prey spe-
cies in the diet, and the black polygons show the region of nutrient
space that is accessible to gannets at each site given the composition
of prey. Hollow red markers show the recorded diets of individual
gannets, and solid red markers show the average compositions of the
diets for each site. The prey species shown represent 100 and 82.6 %
of the mass of the recorded diets for HG and FS gannets, respectively.
The negative diagonal represents a total macronutrient content of
26 %—i.e. % protein and % lipid of any point on that line would sum
to 26 % (see Raubenheimer 2011 for the derivation)
◂
Author's personal copy

2797Mar Biol (2014) 161:2791–2801
1 3
different scales (Fig. 2a). First, we observed differences
between species within a similar geographic location that
could be explained by differences in their trophic lev-
els (Simpson and Raubenheimer 2012). Second, kahawai
from HG and pilchard from FS showed variability between
individuals within locations (kahawai: 3.5–6.6 % lipid,
pilchard: 19.8–22.8 % protein). These results are likely
related to differences in body size, sex or reproductive stage
(Lenky et al. 2012). Third, we found significant geographic
variation within species, including the lipid composition of
pilchards (HG 1.2 %, FS 2.5 % and PPB 9.8 %) that could
be linked to differences in the species’ own diet or in the
latter case (PPB) possible effects of the laboratory methods
used between studies.
In addition to geographic effects, seasonal variation
is likely to influence prey macronutrient composition in
response to changes in productivity or as a function of
reproductive effort or migration patterns (Lenky et al. 2012;
Simpson and Raubenheimer 2012). Prey samples for this
study were collected to ensure an accurate representation
of gannets’ nutrient gain at the time the prey composition
of their diet was studied. Overall, these results highlight the
importance to nutritional ecology studies of analysing spa-
tiotemporal fluctuations in an animal’s diet at both the food
and nutrient levels.
Prey and macronutrient composition of diets
Prey composition of gannets’ diets is believe to reflect the
relative abundance of pelagic fish and squid in their forag-
ing grounds (Jarvis 1970; Wingham 1985; Robertson 1992;
Bunce 2001; Schuckard et al. 2012). In terms of mass,
pilchard was the most important prey species in the gan-
nets’ diet from both New Zealand colonies studied as has
been reported in previous studies (Wingham 1985; Schuck-
ard et al. 2012). Despite similarities in the species of prey
observed in the diets of both colonies, we found a signifi-
cant difference in the macronutrient composition of their
diets between sites (Fig. 2b).
Fig. 3 Right-angled mixture triangles comparing the diet composi-
tions (as a % of wet weight) of a Australasian gannets in New Zea-
land (HG and FS, black polygons) and Australia (PPB, grey poly-
gon). b Northern gannets from FI, Canada (green polygon) and BR,
Scotland (blue polygon) and c Cape gannets on MI, Western Cape
of South Africa (star) and BI, Eastern Cape of South Africa (purple
polygon). Solid-coloured symbols show the average composition of
prey species in the diet, and the coloured polygons show the region of
nutrient space that is accessible to gannets at each site given the com-
position of prey. Red solid symbols show the average compositions
of the diets across gannets at each location. In a, proportion of prey
species represented are HG = 100 %, FS = 82.6 % PPB = 79 %; in
b FI = 99.6 % and BR = 96.7 %; in c MI = 100 % and BI = 83 %
▸
Author's personal copy

2798 Mar Biol (2014) 161:2791–2801
1 3
Gannets from HG consumed a lower proportion of
protein and a lower proportion of lipid and protein com-
bined (i.e. a higher proportion of rest) than their conspe-
cifics from FS. While similar lipid intakes were achieved
between colonies, there is no suggestion as to what level
of protein intake is preferred. Since the nutrient space
accessible to gannets from each location did not overlap,
it is possible that one population was constrained from
securing adequate protein intake which could have fitness
consequences. However, this is perhaps unlikely as the
Australasian gannet population in New Zealand has been
considered to be increasing annually by at least 2.3 % since
1947 (Nelson 2005), suggesting that nutrient requirements
are being met for these seabirds.
It has previously been shown that adult chick-rearing
gannets structure their foraging trips to cover their ener-
getic needs first and those of their offspring second (Rop-
ert-Coudert et al. 2004). The energy required by a gannet
parent to meet their own needs and those of their chick is
equivalent to double that found in an average regurgitation
(Wingham 1989), but it is unknown whether the adults’
nutrient requirements are similar to those of their chick.
This could be an important consideration when analysing
gannet diets from carcasses and regurgitations, in terms
of: (1) whether the adults were rearing chicks; (2) the time
of the last feed, considering that digestion of a complete
fish takes between 2 and 6 h (Davies 1956; Machovsky-
Capuska et al. 2011b); and (iii) whether the diets obtained
represent the adult or the chick’s meal (Richoux et al.
2010). This warrants further investigation.
In the field, there are complex sets of interacting varia-
bles that present several logistical challenges for collecting
reliable daily intake data for nutritional studies. However,
an estimate of the proportional composition of the diet can
be obtained using gut contents analysis (Petry et al. 2007;
Machovsky-Capuska et al. 2011b), regurgitations (Duffy
and Jackson 1986; Schuckard et al. 2012) faecal analy-
sis (Duffy and Jackson 1986; Lea et al. 2002), bite rates
analysis (Paddack et al. 2006), video footage (Machovsky-
Capuska et al. 2011a, 2012), stable isotopes (Cherel et al.
2000, 2005) and when possible a combination of these
techniques. Furthermore, there are other aspects relevant to
the accuracy of the methods use for estimating diet compo-
sition (Duffy and Jackson 1986; Votier et al. 2003). We rec-
ommend combining RMTs’ with pre-existing techniques
to enhance our accurate estimation of dietary intakes in
marine predators.
Interspecific nutritional variability
Do gannets have similar nutrient gains across environ-
ments? We have used published data on the three gannet
species from different populations to illustrate how RMTs
can be used to show how dietary nutrient balance can vary
between populations and that patterns for achieving the
observed nutrient gains differed between populations and
species (Figs. 3a–c, 4).
The diet of the PPB population was higher in lipid and
intermediate in protein compared to the New Zealand pop-
ulations. Although the nutrient space accessible to PPB
overlapped with that of HG, the average diet of PPB did
not fall within this area of overlap. It is possible that PPB
gannets have different nutrient requirements and/or experi-
ence greater fluctuations in prey availability leading them
to select more lipid-rich prey than their conspecifics in
New Zealand (Bunce 2001), though differences in labora-
tory methods for lipid measurement may also have con-
tributed to this finding. Northern gannets from FI and BR
showed a remarkably similar balance of nutrients in their
average diets, which both fell within the area of overlap
in their nutrient spaces. Cape gannets from MI and BI had
the same prey species in their diets thus sharing a region
in nutrient space. However, the diets of BI population were
higher in lipid and protein than their conspecifics from MI.
While gannets from BI consumed high-quality pelagic fish
(mainly sardine, Sardinops sagax and anchovy, Engraulis
encrasicolus), the MI population relied on discarded fish-
eries waste (hake, Merluccius paradoxus) (Pichegru et al.
Fig. 4 Right-angled mixture triangle showing (as a % of wet weight)
the average macronutrient composition of the diet consumed by dif-
ferent populations of Australasian, Northern and Cape gannets.
Marker colour represents gannet species (red Australasian, yellow
Northern, aqua Cape), and marker shape represents population loca-
tion (circle HG, triangle FS, diamond PPB, upside-down triangle FI,
square BR, bowtie MI, oval BI)
Author's personal copy

2799Mar Biol (2014) 161:2791–2801
1 3
2007). This difference in the macronutrient composition
of the diets related to prey availability and was suggested
to contribute to the decline of the MI population (Pichegru
et al. 2007). An overall comparison of the composition of
diets from the three gannet species showed interspecific
nutritional variability as a function of geographic location.
The lipid intakes of the MI colony and the New Zealand
colonies analysed were relatively similar, despite their
opposite population trends. Life-history responses to nutri-
tional state are likely to play an important role in nutritional
regulation in both species and require further examination.
These results are, however, subject to the caveat that we
were unable to obtain proximate composition values from
the literature that were contemporary and geographically
similar to the dietary studies available for the species stud-
ied. This is likely to influence the shapes of the nutritional
niche landscapes as well as the macronutrient compositions
of the estimated diets.
Macronutrients are, however, clearly not the only func-
tionally important nutritional components of foods: the
constituent molecules in macronutrients (amino acids
and fatty acids, for example) and micronutrients such as
vitamins and minerals also play a critical role in an ani-
mal’s nutritional strategies and physiology (Simpson and
Raubenheimer 2012). Micronutrients, such as essential
minerals, vitamin E and carotenoids, must be obtained
through the diet (Evans and Halliwell 2001). Deficiencies
in dietary micronutrients have been linked to an increased
risk of many diseases (Hegseth et al. 2011; Lucas et al.
2014).
Herbivores and omnivores, including species of insects,
birds and mammals, have been shown using geometrical
analysis to regulate their intake of macronutrients and some
micronutrients and to make post-ingestive adjustments to
help attain the optimal balance of nutrients to meet their vari-
ous requirements (Raubenheimer and Jones 1996; Rauben-
heimer et al. 2007; Simpson and Raubenheimer 2012).
Nutritional geometry has been used to model the interactive
effects that nutrients have on different animals and humans
(Raubenheimer et al. 2014), dietary problems in the critically
endangered herbivorous parrot, the kakapo (Strigops habrop-
tila, Raubenheimer and Simpson 2006), and to explain the
patterns of annual migration in giant pandas (Ailuropoda
melanoleuca, Nie et al. 2014). We view similar studies in
seabirds and other marine predators as a priority.
Confirmation of significant nutritional variability at dif-
ferent scales in the diets of these marine predators high-
lights the important question for future field-based stud-
ies as to what extent it provides nutritional opportunity
vs. constraint. A comparison of methods used for macronu-
trient analysis is a priority in order to validate the large dif-
ferences in diet observed between species and studies. The
challenge ahead is to integrate this nutritional modelling
framework for dietary assessments with long-term sys-
tematic collection of samples for proximate analysis. Our
approach allows us to continue to gain a better understand-
ing of the mechanisms governing dietary choices in wild
carnivores, which is of central importance to understand-
ing the evolution and adaptations of predators and their
influence on their food webs (Simpson and Raubenheimer
2012).
Acknowledgments We thank Sarah Dwyer, Rob Schuckard, Willie
Cook, David Melville, Danny Boulton, Karen and Sabrina Macho-
vsky and Sonja Clements for their assistance with sample collection.
We also thank the anonymous reviewers for helpful comments on
early versions of the manuscript. Aspects of this work were funded
by Massey University Research Fund (MURF) and Faculty of Vet-
erinary Science (The University of Sydney). Samples were collected
under Department of Conservation permits NM-32772-FAU and
AK-26359-FAU.
References
Adams N, Abrams R, Siegrfried W, Nagy K, Kaplan I (1991) Energy
expenditure and food consumption by breeding Cape gannets
Morus capensis. Mar Ecol Prog Ser 70:1–9
Annett CA, Pierotti R (1999) Long-term reproductive output in west-
ern gulls: consequences of alternate tactics in diet choice. Ecol-
ogy 80:288–297
AOAC (2005) Official methods of analysis of AOAC International,
18th edn. AOAC International, Arlington
Bunce A (2001) Prey consumption of Australasian gannets (Morus
serrator) breeding in Port Phillip Bay, southeast Australia, and
potential overlap with commercial fisheries. ICES J Mar Sci
58:904–915
Chambers PG, Simpson SJ, Raubenheimer D (1995) Behavioural
mechanisms of nutrient balancing in Locusta migratoria nymphs.
Anim Behav 50:1513–1523
Cherel Y, Hobson KA, Weimerskirch H (2000) Using stable-isotope
analysis of feathers to distinguish moulting and breeding origins
of seabirds. Oecologia 122:155–162
Cherel Y, Hobson KA, Bailleul F, Groscolas R (2005) Nutrition, phys-
iology, and stable isotopes: new information from fasting and
molting penguins. Ecology 86:2881–2888
Davies DH (1956) The South African pilchard (Sardinops ocellata)
and maasbanker (Trachurus trachurus) bird predators, 1954–
1955. South African Department of Commerce and Industry,
Division of Fisheries Investigational Report 23
Duffy DC, Jackson S (1986) Diet studies of seabirds: a review of
methods. Colon Waterbird 9:1–17
Ellis HI (1984) Energetics of free-ranging seabirds. In: Whittow
GC, Rahn H (eds) Seabird energetics. Springer, New York, pp
203–234
Evans P, Halliwell B (2001) Micronutrients: oxidant/antioxidant sta-
tus. Br J Nutr 85:S67–S74
Fryxell JM, Lundberg P (1997) Individual behavior and community
dynamics. Chapman and Hall, New York
Furness RW (1978) Energy requirements of seabird communities: a
bioenergetics model. J Anim Ecol 47:39–53
Galef BG (1996) Food selection: problems in understanding how we
choose foods to eat. Neurosci Biobehav Rev 20:67–73
Garthe S, Montevecchi WA, Davoren GK (2011) Inter-annual changes
in prey fields trigger different foraging tactics in a large marine
predator. Limnol Oceanogr 56:802–812
Author's personal copy

2800 Mar Biol (2014) 161:2791–2801
1 3
Grémillet D, Pichegru L, Kuntz G, Woakes AG, Wilkinson S, Craw-
ford RJM, Ryan PG (2008) A junk-food hypothesis for gannets
feeding on fishery waste. Proc R Soc Lond Biol 275:1149–1156
Hamer KC, Phillips RA, Wanless S, Harris MP, Wood AG (2000)
Foraging ranges, diets and feeding locations of gannets (Morus
bassanus) in the North Sea: evidence from satellite telemetry.
Mar Ecol Prog Ser 200:257–264
Hamer KC, Humphreys EM, Wanless S, Garthe S, Hennicke J, Peters
G, Phillips RA, Harris MP (2007) Annual variation in diets, feeding
locations and foraging behaviour of gannets in the North Sea: flex-
ibility, consistency and constraint. Mar Ecol Prog Ser 338:295–305
Hawkins JM (1988) The Farewell Spit gannetry—a new sea level col-
ony. Notornis 35:249–260
Hegseth MN, Camus L, Helgason LB, Bocchetti R, Gabrielsen GW,
Regoli F (2011) Hepatic antioxidant responses related to levels of
PCBs and metals in chicks of three Arctic seabird species. Comp
Biochem Physiol C Toxicol Pharmacol 154:28–35
Jarvis MJF (1970) Interactions between man and the South African
Gannet Sula capensis. Ostrich 40(Suppl. 8):497–513
Lea MA, Cherel Y, Guinet C, Nichols PD (2002) Antarctic fur seals
foraging in the polar frontal zone: inter-annual shifts in diet as
shown from fecal and fatty acid analyses. Mar Ecol Prog Ser
245:281–297
Lenky C, Eisert R, Oftedal OT, Metcalf V (2012) Proximate com-
position and energy density of nototheniid and myctophid fish
in McMurdo Sound and the Ross Sea, Antarctica. Polar Biol
35:717–724
Lucas A, Morales J, Velando A (2014) Differential effects of specific
carotenoids on oxidative damage and immune response of gull
chicks. J Exp Biol 217:1253–1262
Machovsky-Capuska GE, Vaughn RL, Würsig B, Katzir G, Rauben-
heimer D (2011a) Dive strategies and foraging effort in the Aus-
tralasian gannet Morus serrator revealed by underwater videog-
raphy. Mar Ecol Prog Ser 442:255–261
Machovsky-Capuska GE, Dwyer SL, Alley MR, Stockin KA,
Raubenheimer D (2011b) Evidence for fatal collisions and klep-
toparasitism while plunge diving in gannets. Ibis 153:631–635
Machovsky-Capuska GE, Howland HC, Vaughn RL, Würsig B,
Raubenheimer D, Hauber ME, Katzir G (2012) Visual accommo-
dation and active pursuit of prey underwater in a plunge diving
bird: the Australasian gannet. Proc R Soc B 279:4118–4125
Machovsky-Capuska GE, Hauber ME, Libby E, Amiot C, Raubenhe-
imer D (2013) The contribution of private and public information
in foraging by Australasian gannets. Anim Cogn 17:849–858
Machovsky-Capuska GE, Hauber ME, Dassis M, Libby E, Wikel-
ski MC, Schuckard R, Melville D, Cook W, Houston M,
Raubenheimer D (2014) Foraging behaviour and habitat use of
chick-rearing Australasian Gannets in New Zealand. J Ornithol
155:379–387
Manighetti B, Carter L (1999) Across-shelf sediment dispersal, Hau-
raki Gulf, New Zealand. Mar Geol 160:271–300
Mayntz D, Raubenheimer D, Salomon M, Toft S, Simpson SJ (2005)
Nutrient-specific foraging in invertebrate predators. Science
307:111–113
Mayntz D, Nielsen VH, Sørensen A, Toft S, Raubenheimer D,
Hejlesen C, Simpson SJ (2009) Balancing of protein and lipid
intake by a mammalian carnivore, the mink Mustela vison. Anim
Behav 77:349–355
Meynier L, Stockin KA, Bando MKH, Duignan PJ (2008) Stomach
contents of common dolphin (Delphinus sp.) from New Zealand
waters. NZ J Mar Freshw 42:257–268
Montevecchi WA, Ricklefs RE, Kirkham IR, Gabaldon D (1984)
Growth energetics of nestling northern gannets (Sula bassanus).
Auk 101:334–341
Nelson JB (2005) Pelicans, cormorants and their relatives. Oxford
University Press, Oxford
Nie Y, Zhang Z, Raubenheimer D, Elser JJ, Wei W, Wei F (2014)
Obligate herbivory in an ancestrally carnivorous lineage: the
giant panda and bamboo from the perspective of nutritional
geometry. Funct Ecol. doi:10.1111/1365-2435.12302
Paddack MJ, Cowen RK, Sponaugle S (2006) Grazing pressure of
herbivorous coral reef fishes on low coral-cover reefs. Coral
Reefs 25:461–472
Paulin C, Stewart A, Roberts C, McMillan P (1989) New Zealand
Fish: a complete guide. In: National museum of New Zealand
miscellaneous series no. 19
Petry MV, Fonseca VSD, Scherer AL (2007) Analysis of stomach
contents from the black-browed albatross, Thalassarche melano-
phris, on the coast of Rio grande do sul, southern Brazil. Polar
Biol 30:321–325
Pichegru L, Ryan P, van der Lingen C, Coetzee J, Ropert-Coudert
Y, Grémillet D (2007) Foraging behaviour and energetics of
Cape gannets Morus capensis feeding on live prey and fishery
discards in the Benguela upwelling system. Mar Ecol Prog Ser
350:127–136
Polis GA, Holt RD (1992) Intraguild predation—the dynamics of
complex trophic interactions. Trends Ecol Evol 7:151–154
Polis GA, Myers CA, Holt RD (1989) The ecology and evolution of
intraguild predation: potential competitors that eat each other.
Annu Rev Ecol Syst 20:297–330
Raubenheimer D (2011) Toward a quantitative nutritional ecology:
the right-angled mixture triangle. Ecol Monogr 81:407–427
Raubenheimer D, Jones SA (1996) Nutritional imbalance in an
extreme generalist omnivore: tolerance and recovery through
complementary food selection. Anim Behav 71:1253–1262
Raubenheimer D, Simpson SJ (2006) The challenge of supplementary
feeding: can geometric analysis help save the kakapo? Notornis
53:100–111
Raubenheimer D, Mayntz D, Simpson SJ, Toft S (2007) Nutrient-spe-
cific compensation following diapause in a predator: implications
for intraguild predation. Ecology 88:2598–2608
Raubenheimer D, Machovsky-Capuska GE, Gosby AK, Simpson S
(2014) The nutritional ecology of obesity: from humans to com-
panion animals. Br J Nutr. doi:10.1017/S0007114514002323
Richoux NB, Jaquemet S, Bonnevie BT, Cherel Y, McQuaid CD
(2010) Trophic ecology of grey-headed albatrosses from Marion
Island, Southern Ocean: insights from stomach contents and diet
tracers. Mar Biol 157:1755–1766
Robertson D (1992) Diet of the Australasian gannet Morus serrator
(G.R. Gray) around New Zealand. N Z J Ecol 16:77–81
Ropert-Coudert Y, Grémillet D, Kato A, Ryan P, Naito Y, Le Maho
Y (2004) A fine-scale time budget of Cape gannets provides
insights into the foraging strategies of coastal seabirds. Anim
Behav 67:985–992
Rothman JM, Raubenheimer D, Chapman CA (2011) Nutritional
geometry: gorillas prioritize non-protein energy while consuming
surplus protein. Biol Lett 7:847–849
Rubio V, Sanchez F, Zamora S, Madrid J (2008) Endogenous modifi-
cation of macronutrient selection pattern in sea bass (Dicentrar-
chus labrax). Physiol Behav 95:32–35
Rubio VC, Navarro DB, Madrid JA, Sanchez-Vazquez FJ (2009)
Macronutrient self-selection in Solea senegalensis fed macronu-
trient diets and challenged with dietary protein dilutions. Aqua-
culture 291:95–100
Schuckard R, Melville DS, Cook W, Machovsky-Capuska GE (2012)
Diet of the Australasian gannet (Morus serrator) at Farewell Spit,
New Zealand. Notornis 59:66–70
Simpson SJ, Raubenheimer D (2012) The nature of nutrition: an inte-
grative framework from animal adaptation to human obesity.
Princeton University Press, Princeton
Stephens DW, Krebs JR (1986) Foraging theory. Princeton University
Press, Princeton
Author's personal copy

2801Mar Biol (2014) 161:2791–2801
1 3
Votier SC, Bearhop S, MacCormick A, Ratcliffe N, Furness RW
(2003) Assessing the diet of great skuas, Catharacta skua, using
five different techniques. Polar Biol 26:20–26
Votier SC, Bearhop S, Witt MJ, Inger R, Thompson D, Newton J
(2010) Individual responses of seabirds to commercial fisheries
revealed using GPS tracking, stable isotopes and vessel monitor-
ing systems. J Appl Ecol 47:487–497
Wanless S, Harris MP, Redman P, Speakman JR (2005) Low energy
values of fish as a probable cause of a major seabird breeding
failure in the North Sea. Mar Ecol Prog Ser 294:1–8
Westoby M (1974) An analysis of diet selection by large generalist
herbivores. Am Nat 108:290–304
Westoby M (1978) What are biological bases of varied diets? Am Nat
112:627–631
Wilder SM, Eubanks MD (2010) Might nitrogen limitation promote
omnivory among carnivorous arthropods? Comment. Ecology
91:3114–3117
Wingham EJ (1985) Food and feeding range of the Australasian gan-
net Morus serrator (gray). Emu 85:231–239
Wingham E (1989) Energy requirements of Australasian gannets
Morus serrator (Gray) at a breeding colony. Emu 89:65–70
Work TM (2000) Avian necropsy manual for biologists in remote ref-
uges. US Geological Survey, National Wildlife Health Centre,
Hawaii Field Station, Honolulu
Author's personal copy