Journal of Human Evolution xxx (2014) 1e7
Contents lists available at ScienceDirect
Journal of Human Evolution
journal hom epage: www.elsevier.com /locate /jhevol
Macronutrient contributions of insects to the diets of
hunteregatherers: A geometric analysis
David Raubenheimer a, *,1, Jessica M. Rothman b, c, Herman Pontzer b, c, Stephen J. Simpson d
a
Charles Perkins Centre and Faculty of Veterinary Science and School of Biological Sciences, University of Sydney, Sydney, Australia
Department of Anthropology, Hunter College of the City University of New York, New York, USA
c
New York Consortium in Evolutionary Primatology, New York, USA
d
School of Biological Sciences and the Charles Perkins Centre, University of Sydney, Sydney, NSW, Australia
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 11 November 2012
Accepted 16 February 2014
Available online xxx
We present a geometric model for examining the macronutrient contributions of insects in the diets of
pre-agricultural humans, and relate the findings to some contemporary societies that regularly eat insects. The model integrates published data on the macronutrient composition of insects and other foods
in the diets of humans, recommended human macronutrient intakes, and estimated macronutrient intakes to examine the assumption that insects provided to pre-agricultural humans an invertebrate
equivalent of vertebrate-derived meats, serving primarily as a source of protein. Our analysis suggests
that insects vary more widely in their macronutrient content than is likely to be the case for most wild
vertebrate meats, spanning a broad range of protein, fat and carbohydrate concentrations. Potentially,
therefore, in terms of their proportional macronutrient composition, insects could serve as equivalents
not only of wild meat, but of a range of other foods including some shellfish, nuts, pulses, vegetables and
even fruits. Furthermore, humans might systematically manipulate the composition of edible insects to
meet specific needs through pre-ingestive processing, such as cooking and selective removal of body
parts. We present data suggesting that in modern societies for which protein is the more limiting
macronutrient, pre-ingestive processing of edible insects might serve to concentrate protein. It is likely,
however, that the dietary significance of insects was different for Paleolithic hunteregatherers who were
more limited in non-protein energy. Our conclusions are constrained by available data, but highlight the
need for further studies, and suggest that our model provides an integrative framework for conceiving
these studies.
Ó 2014 Published by Elsevier Ltd.
Keywords:
Human insectivory
Human entomophagy
Nutritional ecology
Human diet
Nutritional geometry
Protein
Pre-ingestive processing
Introduction
That insects play a significant, but variable, role in the diets of
many contemporary human populations is unquestionable
(Raubenheimer and Rothman, 2013), and the same is likely true in
the early human diet (McGrew, 2001). Indeed, a review of huntere
gatherer diets by Kaplan et al. (2000) indicates that invertebrates
such as insects account for up to 20% of ingested calories in some
traditional cultures. But what nutritional roles do insects and other
invertebrates play in the human diet? It is commonly assumed that
they fulfill a role equivalent to the more conventional vertebratederived meats, and compositional analyses of insects seem to
* Corresponding author.
E-mail address: david.raubenheimer@sydney.edu.au (D. Raubenheimer).
1
Faculty of Veterinary Sciences, 308 JD Stewart Building B01, University of
Sydney, Sydney, NSW 2006, Australia.
support this. Studies have reported similarities between insects and
conventional vertebrate-derived meats in energy and protein
levels, fatty acid and amino acid profiles, as well as vitamin and
mineral composition (Bukkens, 1997, 2005; McGrew, 2001;
O’Malley and Power, 2012; Raubenheimer and Rothman, 2013).
However, several factors lead us to believe that the nutritional
role of insects in the human diet warrants deeper analysis. First, the
phenomenal diversity of insects raises the question of whether they
might play a variety of roles in the human diet. Insects are by a
respectable margin the most phylogenetically and probably also
ecologically diverse of the multi-cellular organisms (Price et al.,
2011), and studies have suggested that both taxonomy
(Raubenheimer and Rothman, 2013) and ecology (Fagan et al.,
2002) are associated with compositional variance in insects.
Many insects also have complex life cycles comprising several
disparate developmental stages, which may differ significantly in
structural and chemical properties (Rothman et al., 2014).
http://dx.doi.org/10.1016/j.jhevol.2014.02.007
0047-2484/Ó 2014 Published by Elsevier Ltd.
Please cite this article in press as: Raubenheimer, D., et al., Macronutrient contributions of insects to the diets of hunteregatherers: A geometric
analysis, Journal of Human Evolution (2014), http://dx.doi.org/10.1016/j.jhevol.2014.02.007
2
D. Raubenheimer et al. / Journal of Human Evolution xxx (2014) 1e7
Additionally, the composition of insects may be systematically
altered during pre-ingestive processing by humans (Raubenheimer
and Rothman, 2013), further increasing the potential for nutritional
diversity in this category of foods. Second, the decision of whether
to eat a specific food is determined not only by the composition of
that food, but a multitude of causal factors and their interactions
including the relationships of the food with the nutritional requirements of the consumer (Oftedal, 1991), the consumer’s
nutritional state, the compositions of alternative foods (Simpson
and Raubenheimer, 2012) and, in some species, cultural norms
(Turner and Thompson, 2013). Given the very large number of food
types that comprise the human diet, and the variability of the human diet both within and across populations, it seems likely that
insects may play a variety of nutritional roles, depending on the diet
of the human group in question and the specific foods the insects
are combined with in a particular instance. These and other complexities suggest that human dietary choices are best understood
using integrative models that incorporate the multiple causal factors and illuminate their interactions (Story et al., 2008; Simpson
and Raubenheimer, 2012).
Human foraging and food choices are commonly modeled
within the paradigm of optimal foraging theory (Belovsky, 1987;
Winterhalder and Smith, 2000; Dusseldorp, 2012), in which the
goal of foraging is assumed to be intake maximization subject to
constraints (Illius et al., 2002). This approach has been criticized for
circularity (Owen-Smith, 1996; Illius et al., 2002) and for uncritically assuming that one dietary component, usually energy, predominates in foraging over others (Raubenheimer et al., 2009). As
an alternative, a geometric framework has been developed for
modeling the complexities of nutrition (Raubenheimer and
Simpson, 1993; Simpson and Raubenheimer, 1993, 2012;
Raubenheimer et al., 2009). At the core of a geometric nutritional
model is a two- or higher-dimensional nutrient space, where each
dimension (axis) defines a nutrient of interest. The nutrient space
provides a platform on which the various facets of animal nutrition
(foods, diets, current nutritional state, nutrient requirements, etc.)
can be defined in common, multi-component terms, and interrelated. This approach has provided new insight into many questions
in nutritional ecology (reviewed in Simpson and Raubenheimer,
2012), as well as human nutrition (Simpson et al., 2003; Simpson
and Raubenheimer, 2005; Gosby et al., 2011).
Here we apply an approach from nutritional geometry, the
right-angled mixture triangle (RMT) (Raubenheimer, 2011), to
examine the possible roles of insects in the diets of human huntere
gatherers. We do so by constructing a model to integrate data from
the literature on the compositions of insects reported to be eaten by
humans, published estimates of the macronutrient compositions of
the diets of hunteregatherers, the compositions of foods other than
insects, and recommended nutrient intakes for modern humans. In
addition to assessing insects in hunteregatherer diets, we also
consider the role of pre-ingestive processing of insects, for example
by selectively removing body parts, in altering the balance of nutrients in a way that suggests a specific nutritional role in the diets
of humans (Raubenheimer and Rothman, 2013). The data for this
analysis come from three populations of contemporary human
societies with proportionately low dietary protein intake (in
Nigeria, Thailand and Mexico), and our hypothesis is that preingestive processing should serve to concentrate protein if insects
are eaten primarily as a source of protein to complement a low
protein diet. Given the limitations on available data, we have
restricted our model to the proportional contributions of the
macronutrients protein, lipids and available carbohydrates,
although absolute intakes are clearly also important and micronutrients might also play a role. Our aims are, firstly, to establish
how an integrative analysis of existing data might help to shed light
on the role of insects in human macro-nutrition, and secondly to
introduce a framework that might inspire the collection of data for
refining and extending such analyses on the roles of insects and
other foods in the diets of humans.
Methods
The right-angled mixture triangle
The right-angled mixture triangle (RMT) is a graphical framework for exploring proportional three- or higher-dimensional
compositional data and constructing integrative models in nutritional ecology (Fig. 1). Unlike conventional equilateral mixture
triangles, which have three equal angles of 60 (Emmans, 1991),
two of the axes in RMTs are orthogonal (i.e., at right-angles). This
facilitates a hierarchical approach in which the individual effects of
nutrients, the two-way interactions and three-way interactions are
visually distinguished (Raubenheimer, 2011). Some of the key
properties of these models are:
i. In the general case, model selection involves mathematically
isolating stipulated food components of interest so that the
relationships among these are not confounded by other
components of the mixture. For example, consider an
analysis in which the mass composition of an insect
was measured as 4% non-structural carbohydrate (Cho), 30%
fat (F), 40% protein (P) and 20% chitin (Chi), with the
residue
of
unidentified
components
contributing
6% (100% [4 þ 30 þ 40 þ 20] ¼ 6). If the model is formulated to explore the relationships among protein (P), fat (F),
Figure 1. Right-angled mixture triangles represent three components in twodimensional plots. In this example, each point depicts a mixture of protein (P), fat
(F) and chitin (Chi). By convention, the third, implicit, variable (in this case Chi) is
denoted in square brackets (Raubenheimer, 2011). The % P increases along the X-axis
and % F along the Y-axis. The P:F balance of a mixture is given by the slope of the radial
that connects the point to the origin. The % Chi decreases along each radial with distance from the origin. The negatively sloped diagonals are chitin isolines, such that any
point falling on the same isoline has the same % Chi (denoted by the numbers in square
brackets). For example, mixtures i. and ii. have the same P:F ratio, which is lower than
the P:F ratio of iii., and ii. has a higher concentration of Chi than i. and iii.
Please cite this article in press as: Raubenheimer, D., et al., Macronutrient contributions of insects to the diets of hunteregatherers: A geometric
analysis, Journal of Human Evolution (2014), http://dx.doi.org/10.1016/j.jhevol.2014.02.007
D. Raubenheimer et al. / Journal of Human Evolution xxx (2014) 1e7
and chitin (Chi) (e.g., Fig. 1), then each of these components
would be expressed as a mass percentage of the sum of the
three rather than expressing each as a proportion of the total.
Thus, P in the sample would be expressed as P/
(Chi þ F þ P) 100 ¼ 40/(20 þ 30þ40) ¼ 44.4%, F would be
expressed as 30/(20 þ 30 þ 40) ¼ 33.3%, and Chi would be
expressed as 20/(20 þ 30 þ 40) ¼ 22.2%. This excludes from
the model Cho and R, ensuring that the relationships among
the focal components are not confounded. In the specific case
where macronutrients (P, F and Cho) are modeled and
expressed in energy units, these model components are
automatically isolated because they are usually the only
contributors to dietary energy (i.e., P þ F þ Cho ¼ 100%). An
exception would be if alcohol comprised a significant proportion of energy intake (Yeomans, 2010), in which case P, F
and Cho could not be expressed as a percentage of total energy intake but as a percentage of energy contributed by
P þ F þ Cho.
ii. A powerful property of RMTs is that they enable n-dimensional mixtures to be represented in n 1 dimensional plots
(Fig. 1). This is because in a model investigating, say, three
components, once the first two are stipulated, the value of
the third component is implicit as the difference between
100% and the sum of the other two. Thus, in the above
example,
the
model
is
constructed
such
that
Chi þ F þ P ¼ 100%, therefore Chi ¼ 100% (P þ F) ¼ 100% (44.4 þ 33.3) ¼ 22.3%. Geometrically, this
means that if P is plotted against F then the position of the
point will reflect not only P and F, but implicitly also Chi
(Fig. 1). In general, the values of P and F increase along their
respective axes, and the third component (here, Chi) increases as the city block distance (i.e., the sum of X þ Y) from
the origin decreases. The Chi axis can thus be graphically
expressed as a series of negatively sloped diagonal lines
representing different values for the sum of P and F (i.e., with
slope of 1), where each diagonal intersects the X and Y axes
at the point Chi ¼ 100% (X þ Y). For example, in Fig. 1,
mixtures i and iii both fall on the negatively sloped line that
intersects the X and Y axes at 80%, indicating that they have a
Chi concentration of 100 80% ¼ 20%; by the same
reasoning, point ii. has a Chi concentration of 70%.
iii. RMTs provide a framework within which key facets of
nutritional biology can be integrated (Raubenheimer, 2011).
These might include foods, diets (i.e., the nutrient gains
resulting from eating combinations of foods), and nutrient
needs, expressed either as a single point or as a multivariate
range (Fig. 2).
Dietary macronutrient compositions
To increase the generality of our model, we sought from the
literature estimates of the likely range of diets eaten by huntere
gatherers in general rather than the diet of any particular huntere
gatherer group. For this we used the empirically-parameterized
models of Eaton et al. (1997), Cordain et al. (2000) and Kuipers
et al. (2010). Eaton et al. (1997) reconstructed an ‘average’ Paleolithic diet, a description they qualify as being analogous to the
‘average’ American diet, based on nutritional data from an extensive list of 321 game and wild vegetable items. Their model
assumed a savannah-type diet in which animals were consumed
non-selectively (i.e., implying that all of the available organs were
eaten). Cordain et al. (2000) refined this model, assuming a
savannah diet with a range of prey body compositions (% body fat
ranging from 2.5 to 20.0% by energy) and multiple plant-animal
3
Figure 2. Illustration of the use RMTs to integrate various facets of human nutrition.
The macronutrient ratios recommended in the human diet for reducing the risk of
chronic disease is commonly expressed as Acceptable Macronutrient Distribution
Ranges (AMDR) for protein (10e35% of energy), fat (20e35%) and carbohydrate (45e
65%) (Food and Nutrition Board, 2005). In RMTs, each of these recommendations can
be delineated using constraint lines (the gray dotted lines), where the conjunction of
the three constraints yields an area representing the integrated recommendation for
all three macronutrients (the shaded region). Also plotted are the proportional
macronutrient compositions of four foods. One of these, Chinese broccoli, falls within
the integrated AMDR region and is thus balanced with respect to the recommended
macronutrient intake of humans. In contrast, cashew nuts are too high in lipids, oysters
too high in protein and wild rice too high in carbohydrates. However, by combining
these three imbalanced foods, a diet can be composed with a composition that falls
anywhere within the triangle connecting the foods, the exact position depending on
the relative proportions of the three eaten. In this way a balanced diet can be obtained
through the mixing of foods that are individually imbalanced, such foods being
nutritionally complementary with respect to each other. If two foods are combined, the
composition of the resulting diet will fall on the line connecting the foods. Thus,
humans can obtain a macronutrient-balanced diet (i.e., compose a diet that falls within
the AMDR region) by mixing wild rice and cashew nuts; these are nutritionally
complementary. By contrast, the line connecting cashew nuts and oysters does not
lead to the AMDR region, showing that these on their own are not a complementary
combination. Food compositions from U.S. Department of Agriculture (2012).
subsistence ratios. The likely range of plant-animal subsistence
ratios was obtained from data on the diets of 229 historical huntere
gatherer societies (Gray, 1999). The most comprehensive estimate
of Paleolithic diets was that provided by Kuipers et al. (2010). These
authors combined information from databases of the compositions
of predominately East African plants and animals (including both
mammalian meat and fish) with a likely range of plant/animal
subsistence ratios of 70/30e30/70 and two pathophysiological
constraints to estimate the medians and ranges of likely macronutrient intakes under four Paleolithic foraging strategies. The
pathophysiological constraints were a presumed maximal protein
intake of 35% energy, to avoid exceeding the maximum capacity of
the liver to convert excess nitrogen into urea, and a minimal linoleic
acid intake of above 1% energy to prevent linoleic acid deficiency.
The four foraging strategies were:
Model 1. Selective hunteregatherer/scavenger savannah diet.
This is composed of plant and animals foods widely available on
Please cite this article in press as: Raubenheimer, D., et al., Macronutrient contributions of insects to the diets of hunteregatherers: A geometric
analysis, Journal of Human Evolution (2014), http://dx.doi.org/10.1016/j.jhevol.2014.02.007
4
D. Raubenheimer et al. / Journal of Human Evolution xxx (2014) 1e7
the savannah, excluding aquatic foods. In this strategy, body
components of prey animals are selectively eaten, including
skeletal muscle, marrow and brain but not the liver or adipose
tissue. The proportions of muscle, marrow and brain in the diet
were varied across values considered to be realistic.
Model 2. Non-selective hunteregatherer/scavenger savannah/aquatic diet including plant and animal foods available in
the savannah and aquatic environments. This model assumed
that all edible components of animals were non-selectively
consumed, with mammal/fish intakes varying between 100/
0 and 0/100 of the animal component of the diet.
Model 3. Selective hunteregatherer/scavenger savannah/
aquatic diet. This model is equivalent to Model 2, except that
muscle, bone marrow and brain are selectively scavenged from
savannah animals. The proportion of meat intake contributed by
muscle was varied between 0 and 50%, bone marrow 40e80%
and brain 10e30% (by energy).
Model 4. Non-selective hunteregatherer/scavenger aquatic diet
composed of plants and fish, but not mammalian meat, with the
fat content of fish varied between the range of 2.5 and 10.0%
energy (determined from a database of compositions for East
African fish species).
For comparison, we also plotted the estimated macronutrient
compositions of the national diet in the year 2000 of a selected
modern insect-eating country from the Americas (Mexico), Africa
(Nigeria) and Asia (Thailand) (data from the food balance sheets of
FAOSTAT, 2012). As discussed in the Introduction and further below,
these examples were selected because they represent countries for
which data were available both on the gross composition of edible
insects and their ‘as eaten’ composition (i.e., following preingestive processing), allowing us to test whether processing systematically alters their composition in a way that is consistent with
a role in balancing the macronutrient ratio of the diet. The data also
serve to illustrate the contrast in the composition of the diets of
hunteregatherers and some modern human populations affected
by the nutrition transition (Drewnowski and Popkin, 1997).
Macronutrient compositions of representative foods other than
insects were obtained from the USDA National Nutrient Database
for Standard Reference (U.S. Department of Agriculture, Agricultural Research Service, 2012). Selection was designed to cover a
range of dietary categories (e.g., eggs, seeds, honey, pulses, tubers,
leafy vegetables, nuts, mammalian meats, bird meat, fish, shellfish;
SOM Table 2). Since our interest is primarily in the pre-agricultural
human diet, we attempted where possible to include in the analysis
foods that would have been unaltered or little altered by agriculture
and industrialization; thus, only wild meats and where possible
wild fruits and vegetables were included. However, in some cases
wild equivalents were not available, and we therefore included also
examples of farmed fruits (e.g., banana) and vegetables (e.g.,
carrots).
Effect of pre-ingestive processing on insect compositions
Raubenheimer and Rothman (2013) presented data suggesting
that protein and fat concentration were higher in insect samples
reported as representing the form in which the insect is actually
eaten (e.g., following the selective removal of parts considered
inedible, or heating) than those where the insect was analyzed
whole and raw. That analysis suggests that a function of preingestive processing of insects by humans might be to increase
the macronutrient concentrations. Here we used the data on the
protein-lipid-carbohydrate composition of insects to examine
whether such pre-ingestive processing by modern humans might
also systematically alter the macronutrient balance in edible insects. We excluded from the analysis any instances where processing involved combining insects with exogenous sources of
macronutrients (e.g., oil from frying, Yhoung-aree et al., 1997). For
our comparison, we identified 19 instances of processed samples,
and compared these with 63 that were measured as raw, whole
insects.
Results and discussion
Diet macronutrient compositions
Compositions of insects and other foods
We collated data from the literature on the macronutrient
compositions of 164 insects reported as being eaten by humans, all
of which had values presented for the protein and fat content of the
samples. However, because of the importance of carbohydrates in
the human diet, and the fact that insects might contain significant
concentrations of carbohydrates (Raubenheimer and Rothman,
2013), we selected for analysis only that subset for which nonstructural carbohydrates were also reported (n ¼ 82;
Supplementary Online Material [SOM] Table 1) either as direct
measures or calculated by difference (Raubenheimer and Rothman,
2013; Rothman et al., 2014). We decided not to treat insect species
as a primary unit in the analysis, because the composition of insects
can vary substantially not only with phylogeny but also with such
factors such as life stage and feeding history (Bukkens, 1997;
Raubenheimer et al., 2007), and our goal was to evaluate this diversity in relation to the human diet. Omitting independent measures of the same species, or averaging them, would artificially
reduce this variation. Additionally, the information presented in the
source literature on the history, life stage, and even the species was
often incomplete, precluding the possibility of including these as
factors in an analysis. We therefore regarded each reported
composition as a nominally independent instance of edible insects.
However, where these were presented in the parent publication as
replicate samples of the same insect, we used the mean of the
replicates in our analysis.
The composition of the ‘average’ Paleolithic diet estimated by
the model of Eaton et al. (1997) is shown in Fig. 3, together with
estimated diets in the year 2000 for Mexico, Nigeria and Thailand.
Also plotted is the range of Paleolithic diets estimated by the model
of Cordain et al. (2000), and the ranges and median diets for
foraging strategies 1e4 of Kuipers et al. (2010). For reference we
have also plotted the macronutrient intakes recommended for
humans to reduce the risk of chronic diseases including cancers,
diabetes and atherosclerosis (Acceptable Macronutrient Distribution Ranges ¼ AMDR, Food and Nutrition Board, 2005). The areas
for recommended and estimated nutrient intakes are derived as the
conjunction of the ranges for separate nutrients, as shown in Fig. 2.
Also shown is the region within the nutrient space corresponding
with a diet that exceeds the presumed maximum capacity of the
liver to convert excess nitrogen into urea (35% of energy from
protein). This constraint defines what is assumed to be the upper
limit to estimates of the protein content of the diets of human
foragers (Kuipers et al., 2010).
Food macronutrient compositions
In Fig. 4, we have added to the model the protein:fat:carbohydrate composition of the 82 samples of insects, together with a
range of foods selected to be approximately equivalent to other
components in the diets of foraging humans. Rather than show the
diet models separately as in Fig. 3, for simplicity we show in Fig. 4
Please cite this article in press as: Raubenheimer, D., et al., Macronutrient contributions of insects to the diets of hunteregatherers: A geometric
analysis, Journal of Human Evolution (2014), http://dx.doi.org/10.1016/j.jhevol.2014.02.007
D. Raubenheimer et al. / Journal of Human Evolution xxx (2014) 1e7
Figure 3. Estimated proportional macronutrient intakes (%protein:fat:carbohydrate,
by energy) of Palaeolithic hunteregatherers compared with the recommended
macronutrient distributions (AMDR) for modern humans (yellow polygon) as derived
in Figure 2. The blue, red and green polygons show estimated ranges of intakes under
four different foraging models, and the color-coded ‘X’s the estimated median under
each model (Kuipers et al., 2010 see text for description of the models). The brokenlined black polygon shows the estimated range of intakes from the model of Cordain
et al. (2000), and the black ‘x’ the estimated ‘average Palaeolithic diet’ of Eaton et al.
(1997). The pink region shows a range of dietary compositions that are not possible
for humans, owing to constraints on the maximum rate at which protein can be
physiologically processed. Also shown are the average diet compositions in the year
2000 for Mexico (M), Nigeria (N) and Thailand (T) (data from the food balance tables of
FAO, 2012). (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)
only the combined range of diets estimated by Cordain et al. (2000)
and all four models of Kuipers et al. (2010), as well as the range
spanned by the median diets from the four models of Kuipers et al.
(2010) (the dark polygon in the area labeled HG). Also shown is the
area in composition space spanned by the mammal-derived meats
in our sample, including skeletal muscle, brain, marrow and liver,
and that spanned by the insect samples.
An important point to note from Fig. 4 is the comparatively large
range of insect compositions compared with mammal-derived
meat. The distribution is concentrated in the band between 0%
and 20% carbohydrate, the latter being the lower limit of carbohydrate intake predicted by Kuipers et al. (2010) for the diets of
hunteregatherers. However, some insect samples had carbohydrate content that fell within the predicted macronutrient distribution range for hunteregatherers, and others extended beyond it
into the AMDR region. One sample, the honey ant (on the lower left
hand side of the plot), had a carbohydrate concentration that was
considerably higher (81%) and protein and fat that were lower than
recommended for maintaining human health (5% and 14%,
respectively). By contrast, the range of protein:fat ratios spanned by
insects and mammal-derived meats was similar, but only if the
high-fat marrow is included in the latter. Compared with skeletal
muscle, brain, too, contained relatively high proportions of fat.
However, considering that brain constitutes only approximately
0.14% and marrow 0.30% of the mass of typical prey species (Kuipers
et al., 2010; SOM Table 2), the contribution of these tissues to the
overall composition of mammalian prey is minor.
This compositional variation suggests that insects are potentially a more versatile component of the human diet than are red
meats, fish or shellfish: some are high in protein, some in lipids, and
5
Figure 4. Geometric model integrating the protein:fat:carbohydrate composition (%
energy) of insects with examples of other components of the human diet, estimated
macronutrient intakes of hunteregatherers, and recommended macronutrient intake
ranges for modern humans (AMDR). The green area labeled HG shows the estimated
range of macronutrient intakes spanned by all four models of Kuipers et al. (2010), and
the blue polygon shows the range delineated by the estimated median intakes from
the four models. M, N and T represent the average diet compositions in the year 2000
for Mexico (M), Nigeria (N) and Thailand (T) (data from the food balance tables of FAO,
2012). Also shown are the range of dietary compositions that are attainable by eating
insects (solid black-lined polygon) and vertebrate-derived meats (red polygon),
including muscle, brain, liver and marrow. (For interpretation of the references to color
in this figure legend, the reader is referred to the web version of this article.)
carbohydrates span a range of values. Consequently, insects could
potentially play a variety of roles in the human diet, including
samples with macronutrient ratios similar to fish or red meat,
shellfish, nuts, vegetables, and pulses. One sample, honey pot ant,
even fell among the fruits, and could establish complementary
relationships, as illustrated in Fig. 2, with a wide range of other
foods in the human diet. Furthermore, some insects on their own
had compositions that fell within the estimated macronutrient
range of the diets of hunteregatherers, and others fell within the
AMDR region and are by this definition balanced with respect to
human macronutrient requirements.
We note, however, that these conclusions rest on the assumption that macronutrient values for insects reported in the literature
are reliable, whereas the inconsistency in analytical methods suggests that there is room for improvement (Raubenheimer and
Rothman, 2013). On the other hand, considering that our analysis
included only 82 samples of the 1681 species of insects known to be
eaten by humans (and the true value might exceed 2000) (Paoletti
and Dreon, 2005), we can justifiably assume that our analysis has
under-estimated the diversity of insect compositions. It is quite
likely that the variety of insects available to human foragers in
some pre-agricultural habitats would have been greater yet.
Pre-ingestive alteration of insect compositions
Fig. 5 shows our examination of whether pre-ingestive processing (e.g., selective removal of body parts) by modern humans
systematically alters the composition of edible insects in a way that
is consistent with the hypothesis that processing plays a role in diet
balancing. The analysis indicated that the processed samples had
significantly higher proportional protein (t(1,80) ¼ 2.781, P ¼ 0.007)
Please cite this article in press as: Raubenheimer, D., et al., Macronutrient contributions of insects to the diets of hunteregatherers: A geometric
analysis, Journal of Human Evolution (2014), http://dx.doi.org/10.1016/j.jhevol.2014.02.007
6
D. Raubenheimer et al. / Journal of Human Evolution xxx (2014) 1e7
proportional protein intakes of Paleolithic hunteregatherers,
which exceed human AMDRs (Fig. 3), indicate, however, that fats
and carbohydrates, rather than protein, were limiting in those
populations. This predicts that Paleolithic hunteregatherers would
have valued most highly insects with a proportionally high fat and
carbohydrate content, and pre-ingestive processing involved
methods that specifically retained those nutrients or decreased
protein.
The analysis of Raubenheimer and Rothman (2013) indicates
that pre-ingestive processing might not only alter the macronutrient ratios, but also increase the overall concentration of fat and
protein. An interesting question is whether pre-ingestive processing of insects might also increase the digestibility of nutrients.
Studies have shown that both thermal and nonthermal methods of
processing foods can alter the availability of both energy and specific macronutrients (Carmody et al., 2011).
Conclusions
Figure 5. Comparison of the proportional protein:fat:carbohydrate composition (%
energy) of insects eaten by humans that are specifically reported in the literature as
having been analyzed in the form that they are actually eaten (e.g., following selective
removal of parts considered inedible) (solid squares) and those analyzed as the whole,
raw state (hollow circles).
and lower proportional fat content (t(1,44.4) ¼ 3.33, P ¼ 0.002, equal
variances not assumed), with a higher protein:fat ratio
(t(1,80) ¼ 2.193, P ¼ 0.031). Processing had no effect on the concentration of non-structural carbohydrates (t(1,80) ¼ 0.009,
P ¼ 0.993). This analysis suggests that pre-ingestive processing of
insects by modern humans in Mexico, Nigeria and Thailand might
increase the density of protein and decrease the fat content. One
likely mechanism for this is that fat is lost when insects are cooked
using dry methods such as roasting, smoking, baking or barbecuing
(Yhoung-aree et al., 1997). However, if this were the case then both
protein and carbohydrate would be concentrated relative to fat,
whereas our analysis showed no change in the concentration of
carbohydrate thus suggesting that other factors are likely involved.
The data are clearly inadequate to generalize, and it is probable
that a variety of processing methods are used to achieve different
nutritional outcomes. This analysis has, nonetheless, suggested that
pre-ingestive processing of insects is associated with macronutrient balance, and that processing by modern humans in Mexico,
Thailand and Nigeria might specifically concentrate the protein
content of edible insects. This is consistent with the observation
that the market price of protein is higher than fat or carbohydrate
(Brooks et al., 2010), and thus protein is in general the most limiting
macronutrient in many countries, especially with the rise in availability of cheap cooking oils (Popkin et al., 2012). Fig. 3 shows that
all three countries in our analysis did, indeed, have proportionately
low dietary protein. Other indications that insects are valued for
their protein content in such countries are that a high fat content
might lower the market value of insects (Agbidye et al., 2009), and
insects may be specifically pierced to reduce their fat content
before cooking (Dué et al., 2009).
It is, of course, more difficult to evaluate the effect of preingestive processing of insects in Paleolithic hunteregatherers,
but the above analysis provides a precedent suggesting that
hunteregatherers might, likewise, pre-process edible insects
specifically to alter their nutritional composition. The high
Our analysis suggests that it is an oversimplification to
consider insects as equivalent to vertebrate-derived meats in the
diets of humans, even if the different edible components (skeletal
muscle, liver, brain) are considered separately. Insects vary
widely in their macronutrient composition, and might play a
variety of roles in complementary associations with other components of the human diet. It is widely accepted that in modern
societies insects are valued as a source of protein, and our comparison of pre-ingestive processing of insects is consistent with
this. This is also consistent with the low proportional dietary
protein content in the diets of many contemporary human societies that eat insects, and the relatively high cost of protein.
However, our predictions differ for pre-agricultural societies, for
which non-protein energy was more limiting than protein. For
them, insects could equally have been targeted as a source of fat
and carbohydrate to complement the high protein content of
many conventional meats, giving rise to the prediction that
methods of pre-ingestive processing might have been chosen to
reduce the proportional protein content of insects. Our analysis
is, however, based on limited data. Of the 164 compositions of
human-eaten insects that we recovered from the literature, only
82 reported data for non-structural carbohydrates. The estimates
were variable, and based on inconsistent methodologies
(Raubenheimer and Rothman, 2013), but the data clearly indicate
that it is worth analyzing carbohydrates as a nutritional component of insects. Even if all reported values are true, this sample
size is too small to support robust statistical analyses in which
important predictor variables such as phylogeny, developmental
stage and rearing history are taken into account, and furthermore
such details are incompletely reported in the literature. Further,
micronutrients might also play a role in the selection of edible
insects (Bukkens, 1997), but this is difficult to test given the
extreme paucity of data on their micronutrient content. Finally,
our analysis is based on proportional macronutrient compositions
of human foods, from which predictions can be derived of the
composition of the diet that could potentially result from
combining these foods (as illustrated in Fig. 2). However, the
composition of the actual diets composed by combining various
foods would also be influenced by the relative availability of
different foods. An important goal is thus to seek information on
the quantitative contribution of insects to the diets of huntere
gatherers, of the sort presented by Kaplan et al. (2000). We
believe that such data, combined with multi-dimensional analyses of the nutrient content of foods and diets, will help to
progress the understanding the roles of insects and other invertebrates in the diets of humans.
Please cite this article in press as: Raubenheimer, D., et al., Macronutrient contributions of insects to the diets of hunteregatherers: A geometric
analysis, Journal of Human Evolution (2014), http://dx.doi.org/10.1016/j.jhevol.2014.02.007
D. Raubenheimer et al. / Journal of Human Evolution xxx (2014) 1e7
Acknowledgments
We thank Rob O’Malley and Bill McGrew for inviting us to
contribute to this special issue, and for organizing an insightful
symposium on insectivory. Thanks also to the participants and attendees of the symposium for their helpful discussions. This work
was part-funded by Gravida: The National Research Centre for
Growth and Development, New Zealand.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jhevol.2014.02.007
References
Agbidye, F.S., Ofuya, T.I., Akindele, S.O., 2009. Marketability and nutritional
qualities of some edible forest insects in Benue state, Nigeria. Pakistan J. Nutr. 8,
917e922.
Belovsky, G.E., 1987. Hunter-gatherer foraging: a linear programming approach.
J. Anthropol. Archaeol. 6, 29e76.
Brooks, R.C., Simpson, S.J., Raubenheimer, D., 2010. The price of protein: Combining
evolutionary and economic analysis to understand excessive energy consumption. Obes. Rev. 11, 887e894.
Bukkens, S.G.F., 1997. The nutritional value of edible insects. Ecol. Food Nutr. 36,
287e319.
Bukkens, S.G.F., 2005. Insects in the human diet. In: Paoletti, M.G. (Ed.), Ecological
Implications of Minilivestock. Potential of Insects, Rodents, Frogs and Snails.
Science Publishers, Enfield, pp. 545e577.
Carmody, R.N., Weintraub, G.S., Wrangham, R.W., 2011. Energetic consequences
of thermal and nonthermal food processing. Proc. Natl. Acad. Sci. 108 (48),
19199e19203.
Cordain, L., Miller, J.B., Eaton, S.B., Mann, N., Holt, S.H., Speth, J.D., 2000. Plante
animal subsistence ratios and macronutrient energy estimations in worldwide
hunteregatherer diets. Am. J. Clin. Nutr. 71, 682e692.
Drewnowski, A., Popkin, B.M., 1997. The nutrition transition: new trends in the
global diet. Nutr. Rev. 55, 31e43.
Dué, E.A., Zabri, H.C.B.L., Kouadio, J.P.E.N., Kouamé, L.P., 2009. Fatty acid composition
and properties of skin and digestive fat content oils from Rhynchophorus palmarum L. larva. Afr. J. Biochem. Res. 3, 89e94.
Dusseldorp, G.L., 2012. Studying prehistoric hunting proficiency: applying optimal
foraging theory to the Middle Palaeolithic and Middle Stone Age. Quatern. Int.
252, 3e15.
Eaton, S.B., Eaton III, S.B., Konner, M.J., 1997. Paleolithic nutrition revisited: A
twelve-year retrospective on its nature and implications. Eur. J. Clin. Nutr. 51,
207e216.
Emmans, G.C., 1991. Diet selection by animals: theory and experimental design.
Proc. Nutr. Soc. 50, 59e64.
Fagan, W.F., Siemann, E., Denno, R.F., Mitter, C., Huberty, A.F., Woods, H.A., Elser, J.J.,
2002. Nitrogen in insects: implications for trophic complexity and species
diversification. Am. Nat 160, 784e802.
FAOSTAT database, 2012. URL: http://faostat.fao.org.
Food and Nutrition Board, 2005. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. The National
Academies Press, Washington D.C.
Gosby, A.K., Conigrave, A.D., Lau, N.S., Iglesias, M.A., Hall, R.M., Jebb, S.A., Brande
Miller, J., Caterson, I.D., Raubenheimer, D., Simpson, S.J., 2011. Testing protein
leverage in lean humans: A randomised controlled experimental study. PLoS
One 6, 25929.
Gray, J.P., 1999. A corrected ethnographic atlas. World Cultures J. 10, 24e85.
7
Illius, A.W., Tolkamp, B.J., Yearsley, J., 2002. The evolution of the control of food
intake. Proc. Nutr. Soc. 61, 465e472.
Kaplan, H., Hill, K., Lancaster, J., Hurtado, A.M., 2000. A theory of human life history
evolution: Diet, intelligence, and longevity. Evol. Anthropol. 9, 156e185.
Kuipers, R.S., Luxwolda, M.F., DijckeBrouwer, D.A.J., Eaton, S.B., Crawford, M.A.,
Cordain, L., Muskiet, F.A.J., 2010. Estimated macronutrient and fatty acid intakes
from an east African Paleolithic diet. Brit. J. Nutr. 104, 1666e1687.
McGrew, W.C., 2001. The other faunivory: primate insectivory and early human
diet. In: Stanford, C.B., Bunn, H.T. (Eds.), MeateEating and Human Evolution.
Oxford University Press, Oxford, pp. 160e178.
Oftedal, O.T., 1991. The nutritional consequences of foraging in primates e the
relationship of nutrient intakes to nutrienterequirements. Phil. Trans. R. Soc. B
334, 161e170.
O’Malley, R.C., Power, M.L., 2012. Nutritional composition of actual and potential
insect prey for the Kasekela chimpanzees of Gombe National Park, Tanzania.
Am. J. Phys. Anthropol. 149, 493e503.
Owen-Smith, N., 1996. Circularity in linear programming models of optimal diet.
Oecologia 108, 259e261.
Paoletti, M.G., Dreon, A.L., 2005. Minilivestock, environment, sustainability, and
local knowledge disappearance. In: Paoletti, M.G. (Ed.), Ecological Implications
of Minilivestock. Potential of Insects, Rodents, Frogs and Snails. Science Publishers, Enfield, pp. 1e18.
Popkin, B.M., Adair, L.S., Ng, S.W., 2012. Global nutrition transition and the
pandemic of obesity in developing countries. Nutr. Rev. 70, 3e21.
Price, P.W., Denno, R.F., Eubanks, M.D., Finke, D.L., Kaplan, I., 2011. Insect Ecology:
Behavior, Populations and Communities. Cambridge University Press,
Cambridge.
Raubenheimer, D., 2011. Toward a quantitative nutritional ecology: The righte
angled mixture triangle. Ecol. Monogr. 81, 407e427.
Raubenheimer, D., Rothman, J.M., 2013. The nutritional ecology of entomophagy in
humans and other primates. A. Rev. Entomol. 58, 141e160.
Raubenheimer, D., Simpson, S.J., 1993. The geometry of compensatory feeding in the
locust. Anim. Behav. 45, 953e964.
Raubenheimer, D., Mayntz, D., Simpson, S.J., Toft, S., 2007. Nutrientespecific
compensation following overwintering diapause in a generalist predatory
invertebrate: implications for intraguild predation. Ecology 88, 2598e2608.
Raubenheimer, D., Simpson, S.J., Mayntz, D., 2009. Nutrition, ecology and nutritional
ecology: toward an integrated framework. Funct. Ecol. 23, 4e16.
Rothman, J.M., Raubenheimer, D., Bryer, M.A.H., Takahashi, M., Gilbert, C.C., 2014.
Nutritional contributions of insects to primate diets: implications for primate
evolution. J. Hum. Evol. (in this issue).
Simpson, S.J., Raubenheimer, D., 1993. A multielevel analysis of feeding behaviour:
the geometry of nutritional decisions. Phil. Trans. R. Soc. B 342, 381e402.
Simpson, S.J., Raubenheimer, D., 2005. Obesity: the protein leverage hypothesis.
Obes. Rev. 6, 133e142.
Simpson, S.J., Raubenheimer, D., 2012. The Nature of Nutrition: A Unifying Framework from Animal Adaptation to Human Obesity. Princeton University Press,
Princeton.
Simpson, S.J., Batley, R., Raubenheimer, D., 2003. Geometric analysis of macronutrient intake in humans: The power of protein? Appetite 41, 123e140.
Story, M., Kaphingst, K.M., Robinson-O’Brien, R., Glanz, K., 2008. Creating healthy
food and eating environments: policy and environmental approaches. A. Rev.
Publ. Health 29, 253e272.
Turner, B.L., Thompson, A.L., 2013. Beyond the Paleolithic prescription: Incorporating diversity and flexibility in the study of human diet evolution. Nutr. Rev.
71 (8), 501e510.
U.S. Department of Agriculture, Agricultural Research Service, 2012. USDA National
Nutrient Database for Standard Reference. Release 25. Nutrient Data Laboratory
Home Page. http://www.ars.usda.gov/ba/bhnrc/ndl.
Winterhalder, B., Smith, E.A., 2000. Analyzing adaptive strategies: human behavioral ecology at twenty-five. Evol. Anthropol. 9, 51e72.
Yeomans, M.R., 2010. Alcohol, appetite and energy balance: is alcohol intake a risk
factor for obesity? Physiol. Behav. 100, 82e89.
Yhoung-aree, J., Puwastien, P., Attig, G.A., 1997. Edible insects in Thailand: An unconventional protein source? Ecol. Food Nutr 36, 133e149.
Please cite this article in press as: Raubenheimer, D., et al., Macronutrient contributions of insects to the diets of hunteregatherers: A geometric
analysis, Journal of Human Evolution (2014), http://dx.doi.org/10.1016/j.jhevol.2014.02.007