What's burning got to do with it? Primate foraging opportunities in fire-modified landscapes

Abstract

Objectives:
Anecdotal and formal evidence indicate that primates take advantage of burned landscapes. However, little work has been done to quantify the costs and benefits of this behavior. Using systematic behavioral observations from a population of South African vervet monkeys (Chlorocebus aethiops pygerythrus), we evaluate differences in food availability and energetics before and after controlled burns altered vegetation near their home range. We aim to determine whether burned habitats offer improved foraging opportunities.

Methods:
We collected feeding data from foraging individuals and analyzed common plant foods for their energetic content. We then used the feeding and energetic data to calculate postencounter profitabilities and encounter rates for food types. Using negative binomial and mixed linear regression models we compared data from burned and unburned habitats.

Results:
Our results show significantly improved encounter rates in burned landscapes for two prey items, invertebrates and grasses. However, postencounter profitabilities in burned areas were not significantly different than those achieved in unburned areas.

Conclusions:
Results suggest that improved encounters alone can motivate changes in foraging behavior. These foraging benefits enable the exploitation of burned savanna habitats, likely driving postburn range expansions observed among populations of vervet monkeys. Thus quantified, these results may serve as a foundation for hypotheses regarding the evolution of fire-use in our own lineage. Am J Phys Anthropol, 2015. © 2015 Wiley Periodicals, Inc.

Full text

What’s Burning got to do With it? Primate Foraging
Opportunities in Fire-Modified Landscapes
Nicole M. Herzog,
1,2,3
* Earl R. Keefe,
1
Christopher H. Parker,
1
and Kristen Hawkes
1
1
Department of Anthropology, University of Utah, Salt Lake City, UT 84112
2
Natural History Museum of Utah, Salt Lake City, UT 84112
3
Applied Behavioural Ecology and Ecosystems Research Unit (ABEERU), University of South Africa, FL,
South Africa
KEY WORDS optimal foraging models; behavioral ecology; vervets; feeding ecology
ABSTRACT
Objectives: Anecdotal and formal evidence indicate that primates take advantage of burned landscapes. However,
little work has been done to quantify the costs and benefits of this behavior. Using systematic behavioral observa-
tions from a population of South African vervet monkeys (Chlorocebus aethiops pygerythrus), we evaluate differences
in food availability and energetics before and after controlled burns altered vegetation near their home range. We
aim to determine whether burned habitats offer improved foraging opportunities.
Methods: We collected feeding data from foraging individuals and analyzed common plant foods for their ener-
getic content. We then used the feeding and energetic data to calculate postencounter profitabilities and encounter
rates for food types. Using negative binomial and mixed linear regression models we compared data from burned
and unburned habitats.
Results: Our results show significantly improved encounter rates in burned landscapes for two prey items, inver-
tebrates and grasses. However, postencounter profitabilities in burned areas were not significantly different than
those achieved in unburned areas.
Conclusions: Results suggest that improved encounters alone can motivate changes in foraging behavior. These
foraging benefits enable the exploitation of burned savanna habitats, likely driving postburn range expansions
observed among populations of vervet monkeys. Thus quantified, these results may serve as a foundation for hypoth-
eses regarding the evolution of fire-use in our own lineage. Am J Phys Anthropol 000:000–000, 2015. V
C2015 Wiley
Periodicals, Inc.
Like other animals inhabiting fire-prone environ-
ments, savanna-dwelling primates take advantage of
newly burned landscapes. Two recent reports detail
changes in primate ranging behavior following a fire
(Jaffe and Isbell, 2009; Herzog et al., 2014). In each, ver-
vet monkeys foraged in burned habitat not previously
exploited, and range size grew substantially to include
burned areas. Possible draws for primates in burns
include fluctuations in insect prey and recently cooked
foods (for a review see: Herzog et al., 2014). However,
relatively few studies synthesize the behavioral
responses of animals regularly exposed to burning (but
see: Green et al., 2014), and until now, none have made
primate behavior around fire the primary target of
investigation. As a result, no comprehensive picture has
yet developed of the broad range of interactions that
other members of our own order have with landscape
fires (Parr and Chown, 2003; De Ronde et al., 2004).
These data are important not only for understanding the
observed behaviors, but also because the effects of burn-
ing on foraging payoffs can contribute to the debate sur-
rounding the evolution of obligate fire-use in our own
lineage (Burton, 2009; Wrangham, 2009; Parker, 2014).
To assess the potential foraging improvements fire
may provide to savanna-dwelling primates, we rely on
the theoretical and mathematical framework of Optimal
Foraging Theory (Stephens and Krebs, 1986) to con-
struct hypotheses regarding burn use. Predictions gener-
ated from optimal foraging models help researchers
explain foraging variability within and across taxa (Pyke
et al., 1977; Stephens, 2008; Davies et al., 2012). Despite
the frequent use of optimal foraging theory as an heuris-
tic tool in human studies (for a review of archaeological
applications see: Bird and O’Connell, 2006; Codding and
Bird, 2015, for ethnographic applications: Kelly, 2013),
its use in primatological research is limited. As noted by
Sayers et al. (2010; p 338): “Although direct tests of OFT
models are scarce... a number of primatologists have
referenced foraging theory as an a posteriori tool to
explain observed behavior” (for a priori applications see:
Nakagawa, 1989, 1990; Grether et al., 1992; Altmann,
1998; Baritell et al., 2009; Sayers et al., 2010).
Optimal foraging models focus on the economics of
nutrient acquisition and the decisions foragers make
Grant sponsor: Leakey Foundation General Grant, “Vervet Forag-
ing Strategies in Fire-Altered Landscapes, Loskop, South Africa.”
*Correspondence to: Nicole Herzog, Department of Anthropology,
University of Utah, 270 South 1400 East, Room 102, Salt Lake City,
Utah 84112, USA. E-mail nicole.herzog@anthro.utah.edu
Received 9 March 2015; revised 29 September 2015; accepted 5
October 2015
DOI: 10.1002/ajpa.22885
Published online 00 Month 2015 in Wiley Online Library
(wileyonlinelibrary.com).
Ó2015 WILEY PERIODICALS, INC.
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 00:00–00 (2015)

about resource exploitation given local ecological con-
straints (Stephens and Krebs, 1986; Stephens, 2008). In
perhaps the simplest of the optimal foraging models, the
classic prey model also known as the “optimal diet” or
“diet breadth” model (Emlen, 1966; MacArthur and
Pianka, 1966; Charnov, 1976a; Stephens and Krebs,
1986), an individual forager’s goal is to maximize net
rate of energetic return per unit foraging time (kcal/hr)
by exploiting the optimal set of resources. The model
partitions foraging into two mutually exclusive aspects,
searching and handling, where the later includes all
postencounter activities such as pursuit, preparation,
and consumption. Upon encountering a prey item, the
forager must choose either to handle the resource or
bypass it in search for more profitable prey. The forager
should handle the item when its postencounter return
rate or profitability (kcal acquired/time spent in pursuit
and consumption) is equal to or greater than the
expected overall foraging return rate. The highest
ranked resource is always included in the optimal diet
with lower-ranked resources added or dropped in order
of profitability. The model assumes that resources are
encountered randomly relative to their abundance.
Where prey are patchily distributed, a derivation of the
prey model, the patch choice model (Charnov and
Orians, 1973; Stephens and Krebs, 1986) treats patches
as the prey choice model treats individual prey items. In
the patch choice model, a forager’s goal is to maximize
the net rate of return per unit foraging time by exploit-
ing the optimal array of patches. Decision variables in
the model are: enter a patch or bypass it to travel to
(search for) another with a higher rate.
Here we use a variant of the prey model to investigate
changes in encounters and postencounter return rates
for a population of savanna-dwelling vervet monkeys
(Chlorocebus aethiops pygerythrus) following a fire. We
expect that vervet food items will be unevenly distrib-
uted across the landscape and therefore take advantage
of both prey and patch choice simplifications. We follow
others (Schoener, 1974, 1987; Stephens and Krebs, 1986;
Sayers et al., 2010) in operationally defining all food
resources as individual prey whether they occur as
aggregates/patches, like berries from a bush, or as singu-
lar items. We define an encounter as initial contact with
a prey item/aggregate. Postencounter, we use the inclu-
sive term ‘handling event’ to describe the combined
amount of time spent in pursuit and consumption. Here,
pursuit, whether failed or successful, involves active
engagement with a particular resource (for mobile
resources this includes chasing, for sessile resources
such as tubers this may include digging). We calculate
the rate of energy gained by a forager for each handling
event based on the total energy of the prey item/s con-
sumed and the amount of time spent in handling
(expressed in our analyses as kcal/min, see below).
We test two hypotheses about the net foraging benefits
underpinning documented range expansions postfire.
First, that burning reduces search costs by increasing
encounter rates for food resources that remain postburn.
Observations among contemporary hunter-gatherers
exemplify fire as a tool to increase encounter rates with
profitable prey (Bird et al., 2005; Bliege Bird et al.,
2008; Codding et al., 2014). Fire can improve encounters
with sessile resources by decreasing groundcover and
increasing visibility, exposing resources otherwise con-
cealed by plant growth (Gowlett, 2010). For mobile prey,
fire may improve encounters both by improving visibility
and by altering the prey’s spatial patterns. These
improvements may occur in two temporally distinct
phases. Immediately post-fire, vertebrate and inverte-
brate prey may flee to unburned or sheltered refuges
within the broader landscape of the burn. Aggregated
here, they may be easy targets for foragers. For exam-
ple, Jaffe and Isbell (2009) suspected that vervets at
Segara Ranch were attracted to exposed populations of
Crematogaster spp. which had evacuated acacia thorn
domatia for the safety of deeper crevices at the bases of
host-trees during a fire. As the invertebrates emerged
from tree bases to repopulate their domatia, they
became susceptible to predation. A second influx of
mobile prey may come several weeks after a fire, when
invertebrates return to burned areas to feed on abun-
dant seed-fall and nitrogen-rich new shoots (Andersen,
1988; Swengel, 2001).
Second, that fire will significantly increase the posten-
counter return rates (i.e. the profitability) of certain
resources. By altering the amount of time dedicated to
handling a resource fire can shift profitabilities in several
ways. First, fire may stun, immobilize or otherwise alter
the behavior of mobile prey making them easier to cap-
ture postencounter. Changes to postencounter pursuit are
largely expected to occur among invertebrate prey. Sec-
ond, fire/cooking may alter the mechanical and physiolog-
ical properties of some plant and animal foods making
them easier to process. Mechanical and physiological
changes are expected to occur across all prey types, but
will be especially apparent in two resource categories:
seeds/nuts and underground storage organs (USOs). For
seeds and legumes, cooking or parching can reduce the
amount of work spent on extricating seeds from woody
chaff by either prompting the release of seeds (seed fall)
or by charring already exposed organs. For USOs, fire
can gelatinize starchy tissues and reduce the work of
fracture (Wandsnider, 1997; Dominy et al., 2008; Zink
et al., 2014). Therefore, when found, cooked or charred
USOs and seeds may be consumed at a faster rate.
Finally, fire may alter the energetic availability of certain
plants by contributing nitrogen to soil and promoting
nutrient-rich new growth (Van de Vijver et al., 1999).
The results of this study are consistent with the first
hypothesis; some prey were encountered significantly
more often in burned areas. In contrast, significant
changes in postencounter profitabilities were not detected.
Results suggest that improved encounters alone can moti-
vate changes in foraging behavior. Because primates in
this study calmly avoided the path of the fire (Herzog
et al., 2014), and capitalized on the foraging opportunities
fire created, we take a position similar to Pruetz and
LaDuke (2010) in assuming a deep phylogenetic founda-
tion for passive fire exploitation in the primate clade. Con-
tinuing investigation of foraging benefits for primates in
burned landscapes can provide an analog for those bene-
fits available to early hominins.
METHODS
Study site and subjects
To assess pre- and postburn foraging behaviors field
observers (NH, EK, CP) spent a minimum of five days
per week from February 2012 to July 2012 (with the
exception of 2 week-long breaks, one in April and another
in June) collecting data among a troop of habituated
vervet monkeys (Chlorocebus aethiops pygerythrus)at
the Loskop Dam Nature Reserve (LDNR), Mpumalanga
2N.M. HERZOG ET AL.
American Journal of Physical Anthropology

Province, South Africa. This observational window
allowed researchers to collect data on foraging behavior
for over 2 months before burning and nearly 4 months
postburn. Controlled burns were conducted in early April
2012.
Over the course of the study, the troop consisted of 23
to 25 individuals (four to six adult males, seven adult
females, five subadults, four juveniles, and three
infants). The study subjects’ home range (Fig. 1) is com-
prised of three primary vegetation communities (from
Barrett 2009): Olea europaea subsp. africana-Rhus lepto-
dictya woodland (Olea), Lippia javanica-Loudetia sim-
plex shrubland (Lippia), and Acacia nilotica-Acacia
caffra woodland (Acacia). Olea vegetation is primarily
located within and along dongas (steep-edged drainages)
and is characterized by tall (20 m) closed canopy vege-
tation, a thick bushy understory, and sparse ground-
level plants. Bordering Olea habitat is a dense shrub-
woodland, Acacia, which contains little or no closed can-
opy. In Acacia habitats, short grasses and small shrubs
create significant ground cover and limit ground-level
visibility. Lippia shrublands are situated at the edges of
Acacia woodlands. These shrublands contain few large
trees. Between trees, broad swaths of tall (up to 2 m)
grass are common. Within tall grasses, visibility is
extremely limited.
In 2012, Park fire ecologists and staff ignited several
small fires on April 3rd and 4th. These fires primarily
impacted Lippia shrubland and Acacia woodland habi-
tats located to the south and southwest of the subjects’
preburn home range (Fig. 1). Approximately 343 hec-
tares of land were burned. Of the plant communities in
and surrounding the vervets’ preburn home range, the
plant community in which they spent most time, Olea,
was only minimally affected; a bigger but relatively lim-
ited area of Acacia woodland was also impacted; and
large swaths of Lippia shrubland south of the troop’s
preburn range were burned. Postburning, the vervets
expanded their home range to include much of this
newly burned, but previously uninhabited, Lippia habi-
tat (see: Herzog et al., 2014). The 2012 burns were part
of an annual fire management plan for the LDNR (Eks-
teen, 2003). As per the plan, certain areas are targeted
each year for burning. Burn site selection is made by
park fire ecologists, and controlled burns are carried out
by fire maintenance staff. Several criteria determine
which areas burn. First, management assesses the over-
all occurrence and location of natural fires. If the impact
of natural fires has been limited, controlled burns are
implemented to suppress plant growth in vegetation-
dense areas. Second, to facilitate game viewing, plants
obscuring roadside overlooks are targeted for burning
and burns are conducted such that they promote and
stimulate the growth of new grass, which attracts graz-
ers to areas of higher visibility. Finally, when necessary,
burning is used as a tool for reducing the presence/abun-
dance of animal parasites including various species of
tick (Barrett, 2009; Filmalter, 2010). To reduce the risk
of runaway fires and to promote quick regrowth, burning
typically occurs within the last few weeks of the wet
season.
Behavioral data collection
In order assess changes in forging decisions related to
burning, we collected behavioral data via focal follows
(Altmann, 1974) conducted continuously throughout
each observational day. Follows began in the AM at first
troop encounter and ended in the evening after subjects
had settled at a sleeping site. The first focal subject of
the day was chosen randomly. Subsequent focal animals
were selected in an alternating male-female pattern
without replacement, when possible. Observers con-
ducted focal follows once each hour for a duration of 20
min. During each follow, the behavior of a single vervet
was documented. Each observed behavior was timed and
the following data recorded: habitat, subhabitat, and
burn status (for postburn observations). For tests of our
predictions about changes in foraging return rates,
researchers focused explicitly on feeding behaviors. We
distinguished pursuit (individual has identified a food
and is actively working to obtain or extract it) and con-
sumption (individual has obtained a food and is consum-
ing edible portions). Within the consumption category,
food items were classified by the edible portion con-
sumed. For plants, these categories included leaf, seed,
pod, flower, fruit, fungi, and USO (tubers, rhizomes,
corms, and taproots). Nonplant categories included ani-
mal (vertebrate or invertebrate), water, and other/
unknown. Additionally, where consumption was
observed, the number of bites per handling event was
recorded. Behavioral sequences were digitally recorded
using Trimble GPS units with Cyber Tracker software
(www.cybertracker.org). Over the course of the study,
694 focal follows were recorded (totaling 232 h of obser-
vation). However, in order to compare spatially but not
temporally distinct foraging efforts in burned versus
unburned areas, we restrict our analyses to postburn
focal follows and include only data collected from adult
and subadult subjects. These data include 469 focal fol-
lows, 111 of which document activity in burned habitats.
Because no handling events were observed in burned
Olea, the data we examine are limited to comparisons
between burned and unburned Acacia and Lippia plant
communities.
Fig. 1. Map shows the distribution of plant communities
(Olea europea woodland, Lippia javanica shrubland, and Acacia
spp. woodland) in and around Donga troop pre (outlined in
black) and postburn (outlined in dark gray) home ranges. The
2012 fire scar is also depicted (dashed red line) to highlight the
extent of Lippia and Acacia burned. [Color figure can be viewed
in the online issue, which is available at wileyonlinelibrary.
com.]
FIRE AND PRIMATE FORAGING 3
American Journal of Physical Anthropology

Encounters. To generate a list of all foods encountered
in each of the two habitats impacted by the fire (Acacia
and Lippia), we combed through behavioral data and
identified all prey items that were both available and in
the diet (foods that were pursued or ingested during at
least one handling event in both burned and unburned
areas). Because some prey species were infrequently
taken, we lumped resources into six categories (fruit
[Bridelia molis,Combretum zeyheri, and Grewia flaves-
cens], gum, invertebrate, leaf [Clover spp, Crassula spp.
various grasses, new shoots], seed [Acacia caffra,Acacia
karoo,Acacia nilotica, and Acacia robusta], and USO).
To justify the use of lumped categories, we compared the
Ei(see Eicalculations below and in Tables 1 and 2) of
each species within each type using a Tukey’s Honest
Differences test. No significant differences (P<0.05)
were detected for the fruit or leaf prey sets. For the seed
prey set, one species, Acacia nilotica was significantly
different from the others. However, the number of obser-
vations for this species is very small (n56 handling
events) so we include it here despite its variance.
We calculated encounter rates for each prey type by
summing the total duration (in seconds) of each focal fol-
low spent in search (arboreal or terrestrial travel or
bipedal scanning) and using their sum as the denomina-
tor for the number of discrete handling events observed
for each species. We include time in bipedal scans as
search time unless the scan was in response to a threat
(visually confirmed by field observer), was directly pre-
ceded or followed by an alarm call, or was in response to
the presence of interspecific neighbors. Some follows
included time in both burned and unburned habitat;
where this occurred, both search time and number of
encounters were split according to the burn status of the
habitat. We could only confirm an encounter by a sub-
ject’s response. Encounters with food items that the
monkey did not pursue are not included. We may thus
underestimate encounter rates. Failure to pursue a prey
item identified during vegetation surveys as present in
the habitat where the follow took place was quantified
as a “zero encounter.”
Profitabilities. The rate of energy gained by a forager
from pursuing any given prey is calculated based on the
total digestible energy per item and the time spent in
pursuit and handling (for mathematical equations see
Table 2). We calculated profitabilities in several steps.
First, to estimate the total digestible energy per item,
we collected 23 commonly consumed items and meas-
ured the nutrient and non-nutrient content of each
(Table 1). To obtain plant samples, we gathered 200 g
of each plant part during monthly phenological surveys.
After collection, wet samples of each were weighed
whole. To obtain average weights, we measured approxi-
mately 50 individual seeds, fruits, or fruit pulps. Fruit
pulp weight was determined by subtracting the seed
weight from the fruit total weight. After wet weights for
TABLE 1. Prey species nutrient and energetic values
Prey species
Prey
type
Eaten
in burn ME
1
/dry g %OM %CP %NDF %Fat %Ash g per bite ME
1
/bite
Acacia caffra Seed 13.246 55.1 21.7 20.4 2.2 4.0
a
0.040 0.130
Acacia karoo Seed 13.227 58.8 18.5 20.2 1.8 4.2
a
0.033 0.106
Acacia nilotica Seed 13.180 68.4 8.4 21.0 1.5 4.2
a
0.125 0.399
Acacia robusta Seed 13.055 60.4 15.0 23.3 0.6 4.2
a
0.050 0.153
Acacia tortilis Seed 23.184 58.4 19.5 20.2 1.3 4.5
a
0.039 0.124
Bridelia molis Fruit 12.784 55.2 4.3 37.8 2.1 4.0
b
0.258 0.718
Celtis africana Fruit 23.841 62.7 12.1 15.8 8.7 1.2
b
0.080 0.307
Chaetachme aristata Fruit 23.582 53.4 11.6 25.0 9.4 1.0 0.038 0.135
Combretum zeyheri Fruit 13.691 63.3 9.1 11.7 5.2 4.0
c
0.364 1.344
Euclea crispa Fruit 22.868 59.5 2.9 34.9 2.0 0.7
b
0.039 0.112
Ficus burkei Fruit 22.898 59.7 6.7 31.6 1.4 2.0
b
0.041 0.119
Grewia flavescens Fruit 12.639 57.5 4.3 36.8 0.8 3.2
b
0.225 0.594
Gum
f
Gum 13.535 86.0 1.8 10.8 0.4 2.8
d
0.375 1.325
Invertebrates
g
Inv. 13.566 43.3 27.7 20.2 8.3 3.8
d
0.250 0.892
Kalanchoe lanceolata Leaf 13.143 62.3 6.2 28.0 2.8 0.7 0.083 0.262
Mimosops zeyheri Fruit 22.818 62.9 2.6 33.2 0.6 1.6
b
0.405 1.141
Olea europaea Fruit 22.810 58.9 2.6 35.7 2.1 1.5
e
0.138 0.388
Pappaea capensis Fruit 23.852 50.2 6.9 26.4 15.9 0.7
b
0.079 0.304
Rhus leptodicta Fruit 22.802 57.6 6.0 34.1 1.7 2.6
b
0.027 0.075
Rhus pyroides Fruit 23.263 71.2 6.3 20.4 1.4 2.6
b
0.025 0.082
Sclerocarya birrea Fruit 23.311 73.6 4.4 18.9 2.3 3.8
b
0.438 1.450
Unknown grass
(postburn regrowth)
h
Leaf 12.552 53.0 8.2 37.3 0.9 5.0 0.120 0.306
USO (Romulea sp.)
i
USO 13.657 – – – – – 0.330 1.207
ME
1
5total metabolizable energy per 1 g prey item; OM 5organic matter; CP 5crude protein; NDF 5neutral detergent fiber.
a
Aganga et al. (n.d.).
b
Wehmeyer (1986).
c
Nakagawa (2003).
d
Isbell et al. (2013.
e
Zamora et al. (2001).
f
Composite of three species: Acacia karoo,Acacia tortilis, and Combretum zeyheri.
g
Composite of three species: Nephila senegalensis,Zonocerus elegans, and Crematogaster spp.
h
Research shows increased nutrients in postfire regrowth (Van de Vijver et al., 1999); we therefore estimated a relatively high ash
content for this sample.
i
Data from Bennett and Jarvis (1995).
4N.M. HERZOG ET AL.
American Journal of Physical Anthropology

edible parts were obtained, we dried the samples at 408
C for 48 to 72 h.
Once samples were dried we ground them and shipped
100 g of each to the University of Free State Animal
Nutrition Laboratory for organic matter, crude protein,
fat, and neutral-detergent fiber (NDF) analyses. We
used these results to estimate the metabolizable energy
(ME) in each prey item following the methods outlined
in Conklin-Brittain et al. (2006). We calculated ME
using a low-fermentation assumption (ME
1
) (Table 2)
because data suggest that vervets are able to effectively
ferment some fiber (Conklin-Brittain et al., 1998; Isbell
et al., 2013). In the low-fermentation formulation, sour-
ces of metabolizable energy include fat, protein, carbohy-
drates, and a small amount of fermented NDF. Because
we were unable to conduct ash analyses on our plant
samples, we used estimates from published data to sup-
plement our own (Wehmeyer, 1986; Aganga et al., 1997;
Zamora et al., 2001; Nakagawa, 2003; Isbell et al.,
2013). Where direct measures were unavailable we used
estimates from closely related species, and where no
close analog was found we used conservatively low esti-
mates (see notes to Table 1). For species in which seeds
or fruits enclosed in an inedible pod were targeted, we
extracted the edible portions for analysis. When possible,
we removed seeds and endocarps from fleshy fruits and
included only pulp. For some fleshy fruits, we were
unable to extract the seeds and endocarps. In these
cases (Celtis africana,Olea europaea,Grewia flavescens),
entire fruits were sent for analysis. We report whole
fruit nutrient data here (Table 1), but because seeds of
these species were often either spat out or passed undi-
gested by vervets these values are clearly overinflated.
To be certain that inclusion of overinflated ME values
did not impact profitability results, we calculated three
offset ME conditions (total ME offset by 15%, 20%, and
30%) for the overinflated species that occurred in burned
settings (Grewia flavescens) and included these values in
analyses of profitability. Handling events for this species
made up only a small proportion of the foraging data
(n521 out of 1,116 handling events), and changes in
ME for this species did not affect model outcomes (see
statistical analyses and results below). Due to insuffi-
cient sample sizes of invertebrates and gum, ME for
these prey types were calculated using a composite of
several species. For gum, composite species included:
Acacia karoo,Acacia tortilis, and Combretum zeyheri.
The invertebrate composite included: Nephila senegalen-
sis,Zonocerus elegans, and Crematogaster spp. Field
researchers were unable to collect any samples of USOs.
Energetic data presented here are from a semi-arid
adapted rhizome (Romulea sp.) similar to the suspected
target species taken during this study (Bennett and Jar-
vis, 1995).
Second, we calculated ingestion rates (IRs) for each
prey type by summing the amount of time spent han-
dling a resource and dividing by the number of bites of
that resource taken (Table 2). We used IR averages to
fill in missing data where bite quantities for a given
handling event were unknown. Handling events in
which the prey type was unknown were eliminated from
the data set. For handling events where species-specific
ME or IR were unavailable, the average ME or IR for
that prey type (i.e. fruits, seeds, etc.) was used in calcu-
lations. For the purpose of these analyses, we assume
that a handling bout ends after bites are no longer
taken. Although prey items continue to be processed
through both oral manipulation and digestion long after
the last observable bite, integration of internal handling
costs are beyond the scope of this study. Here we rely
only on the observable phenomena of bites taken and
time spent delivering bites to mouth. For reviews of
issues related to primate digestion see Lambert (1998)
and Milton (1984); for cercopithecine-specific digestive
retention times see Blaine and Lambert (2012).
Finally, we combined these ingestion rate data to cal-
culate the postencounter return rate, or profitability, for
each handling event (Ei). To do this, we multiplied the
metabolizable energy per bite by the total number of
bites taken during the handling event (Table 2). To calcu-
late the amount of ME per bite per species, we measured
the weight (g) of each bite (calculated from weight of food
unit divided by number of bites to full consumption) then
multiplied ME/g of each food by the number of bites
observed. For items consumed in many bites, we divided
the item into the average number of bites taken to eat it,
then measured the weight of one bite. Some food items
were small and many were consumed in a single bite.
For these species, we used the average bite weight for
that prey type in our calculations. Following first contact
with prey, any time spent in pursuit, extraction, or con-
sumption was considered part of the handling event.
After we calculated the profitability of each handling
event, we split the handling events into categories based
on the burn status of the habitat in which they occurred.
Statistical analyses
Prey encounters. Like many ecological and abundance
variables, our encounter counts include a large number
of zeros. Observed zero counts may be driven by two dif-
ferent processes: (1) foods were not detected/available, or
(2) available foods were detected by the forager but were
not pursued. Because we could not distinguish between
these two classes of zeros, we consider them both “zero
encounters” regardless of the underlying reason. Linear
regression models, such as Poisson, are insufficient for
handling this type of data because they do not accurately
TABLE 2. Formulas used to calculate energetic returns
Measure Equation Variables
ME1Total metabolizable energy per
100 g prey item (low fermentation)
ME1543%TNCðÞ143%CPðÞ193%LðÞ
10:5433%NDF
ðÞ
TNC, total nonstructural carbohydrates;
CP, crude protein; L lipids; NDF,
neutral detergent fiber
ME1b Metabolizable energy per bite ME1b5ME13gi
bi
 gi;grams per prey item; bi;
bites per prey item
IR Ingestion rate: mean number of
bites per handling event (seconds) IR5Pbp
Php
bp;total number of bites per handling
event; hp;time handle per event (s)
EiEnergy ingested per handling event Ei5ME1b3bp
FIRE AND PRIMATE FORAGING 5
American Journal of Physical Anthropology

estimate overdispersed data. To determine the effect of
burning on encounters we compared two model classes,
each designed to manage zero-inflated or overdispersed
data. The first model class, zero-inflated negative bino-
mials (ZINB), are tailored to deal with a high frequency
of zero counts by implementing two linear models to esti-
mate the probability of obtaining a zero value: a logit
model accounts for the binary encounter-no encounter
interaction, and a Poisson model measures the magni-
tude of the interaction once an encounter has taken place
(Zuur et al., 2009; Cameron and Trivedi, 2013). ZINB
models were constructed for each prey type, and fitted
using the zeroinfl function of the pscl statistical package
in R (R Core Team, 2012). We also constructed Negative
Binomial Generalized Linear models (GLMNB) for each
prey type. GLMNB models were fitted using the glm.nb
function within the MASS package. We compared the
outcomes of the two model classes using Vuong’s non-
nested test. The GLMNB models provided the best fit,
and we report only their results below.
Postencounter return rates. To determine if burning
had an influence on postencounter returns (Ei), we con-
structed linear mixed-effects models using the lmer func-
tion of the lme4 package in R. The full model included
profitability as the dependent variable and burn status
(coded as 0 for unburned and 1 for burned), prey species,
date, and habitat as fixed effects, and focal ID as a ran-
dom effect (M1). We then used the ANOVA function in R
to compare the full model with a null model (M2) which
did not include the burn variable as a fixed effect. Each
model included 1,143 handling events across 21 unique
individual IDs (including all known adults and suba-
dults, and three focal observations from unidentified
adults). Before comparing the two models, we assessed
the possibility that an inflated ME for Grewia flavescens
might skew results. We constructed full models for each
of four ME conditions (whole ME, ME offset by 15%,
20%, and 30%), then compared the models using the
anova function in R. The inclusion of offset ME condi-
tions did not alter the outcomes of the statistical models
used to assess profitability. None of the models was stat-
istically different from any another (v
2
50, P51). There-
fore, rather than speculate about which offset may be
most accurate, we report only results for whole values.
RESULTS
Fire’s effect on search: Encounter rates
Fire appears to exert variable impacts on encounters
among the different prey types. Encounters for two types
Fig. 2. Scatterplots showing encounters against the amount of time spent in search for each prey type. Orange dots represent
encounters in burned habitats (Acacia and Lippia combined), encounters in unburned habitats (Acacia and Lippia combined) are
shown in gray. Lines depict smoothed loess curves (derived from models estimated separately for burned and unburned conditions)
for burned (orange) and unburned (gray) data. [Color figure can be viewed in the online issue, which is available at wileyonlineli-
brary.com.]
TABLE 3. Summary of negative binomial regression models
illustrating the effect of search time and burn on encountering
resources
Prey type Variable Estimate
SE of
estimate Pr (>|t|)
Ruit Intercept 22.703 0.209 <0.0001***
Search time 0.001 0.001 0.101
Burn 20.062 0.262 0.813
Gum Intercept 21.761 0.223 <0.0001***
Search time 0.002 0.001 0.000***
Burn 0.301 0.252 0.232
Invertebrate Intercept 21.058 0.134 <0.0001***
Search time 0.002 0.000 <0.0001***
Burn 0.369 0.145 0.011*
Leaf Intercept 23.549 0.376 <0.0001***
Search time 0.001 0.001 0.136
Burn 1.176 0.389 0.002**
Seed Intercept 23.070 0.242 <0.0001***
Search time 0.001 0.001 0.349
Burn 20.213 0.313 0.493
USO Intercept 22.607 0.323 <0.0001***
Search time 0.003 0.001 0.001***
Burn 0.029 0.372 0.938
*P<0.05; **P<0.01;
a
P<0.001. Positive statistically signifi-
cant results for burn variable shown in bold.
6N.M. HERZOG ET AL.
American Journal of Physical Anthropology

of prey were significantly positively impacted by burn-
ing, these include leaves and invertebrates (Fig. 2, Table
3). Encounters for other categories were not improved by
burning. However, most primate foods were located in
trees and shrubs above-ground and did not come in
direct contact with fire. Of the terrestrially available
prey types, two out of three (invertebrates and leaves)
showed significant improvements over time (Fig. 3).
Improved encounters with leaves were almost certainly
driven by the emergence of new regrowth approximately
4 weeks postburning. It is likely that the emergence of
the new shoots also strongly influenced the distribution
of insect prey as many of the invertebrate species tar-
geted by the vervets regularly consume shoots and
grasses. When total time spent in handling events is
compared across subjects, 20% was allocated to handling
invertebrates in burned areas compared to only 11% in
unburned areas. Unlike other studies where vervets con-
sumed invertebrates in burned areas within several
days of burning (Jaffe and Isbell, 2009), we saw the larg-
est improvements in insect encounters several weeks
postburning, roughly coincident with the emergence of
regrowth (Fig. 3).
Fire and handling: Prey profitabilities
We did not detect any fire-related changes in profitabil-
ities. Results from the comparison of the null (M2) and
full (M1) models indicate that the inclusion of the burn
variable does not significantly alter outcomes, suggesting
little difference in overall profitabilities between burned
and unburned conditions (v
2
51.7515, P50.1857). While
burn status did not affect the overall outcome of the
model, mean profitabilities (kcal/min) of several prey
types were higher in burned areas than in unburned
(fruit: unburned -M57.338 64.868 [SD], burned -M
57.537 63.676 [SD]; leaves: unburned -M52.545 61.573
[SD], burned -M52.815 62.000 [SD]; USOs: unburned
-M54.894 64.932 [SD], burned -M55.478 64.287 [SD]).
However, the interquartile ranges of each largely overlap
(Fig. 4). These results prompt us to reconsider the expec-
tation of improved postencounter returns for vervets
below.
DISCUSSION
Where natural and/or anthropogenic burning are com-
mon, primatologists have often observed their subjects
using burned areas in novel and unexpected ways (for a
summary see: Herzog et al., 2014). But without quanti-
tative data to compare the foraging behavior of primates
in burned and unburned areas, we are left to wonder,
“What’s fire got to do with it?” Data reported here show
that fire improves encounters with new grasses and
invertebrates. This observation is not entirely new,
human hunters have long used fire as a tool to attract
herbivores into favorable locales well aware of the
appeal of new shoots to grazers (Fisher, 1948; Hall,
1984; Lewis and Ferguson, 1988; Bowman et al., 2001).
Fire ecologists too have noted this relationship, and
termed the draw of herbivores to the nitrogen-rich new
shoots that emerge after a fire the “magnet effect” (Van
de Vijver et al., 1999; Sensenig et al., 2010). This
“magnet effect” draws grazers away from unburned
areas and into recently burned ones (Archibald and
Bond, 2004). Vervets would hardly be considered brows-
ers, but like grass-eating ungulates, they are also drawn
to nutrient-dense regrowth. In this study, and others
among the same population (Barrett, 2009), vervets only
infrequently foraged on grasses before burning, even
new shoots. However, their increased likelihood of pur-
suing leaves in the burn suggests that burning alters
the energetic value of the shoots (i.e. increased nitrogen
values) or decreases the costs of detection, or both.
The influx of invertebrates to areas of regrowth cre-
ates prey aggregates (Swengel, 2001; Moya-Raygoza and
Larsen, 2014), and likely facilitates increased encounters
between invertebrates and their predators. When readily
available, invertebrates supplement and even supplant
other primate foods, a pattern documented even among
primates that only occasionally consume insects (O’Mal-
ley and Power, 2012; Isbell et al., 2013; Rothman et al.,
2014). It should come as no surprise, then, that much of
the available literature detailing primate behavior in
burns describes opportunistic insect hunting (Harrison,
Fig. 3. Scatterplots showing individual encounters with
invertebrates (upper) and leaves (lower) through time (meas-
ured as number of days passed since burning occurred). Orange
dots represent encounters in burned areas and gray dots repre-
sent encounters in unburned settings. Lines show smoothed
loess curves (derived from separately estimated models for
burned and unburned encounters) for data from burned
(orange) and unburned (gray) areas. Dotted green lines repre-
sent emergence of regrowth. [Color figure can be viewed in the
online issue, which is available at wileyonlinelibrary.com.]
Fig. 4. Average postencounter prey profitability per prey
type in unburned (gray, left plot) and burned (orange, right
plot) conditions. Whiskers represent a 95% confidence interval
and circles represent outliers. [Color figure can be viewed in the
online issue, which is available at wileyonlinelibrary.com.]
FIRE AND PRIMATE FORAGING 7
American Journal of Physical Anthropology

1984; Berenstain, 1986; Jaffe and Isbell, 2009). Simi-
larly, vervets in the present study shifted to an
invertebrate-centric foraging strategy in burned
savanna. Subjects spent nearly twice as much time han-
dling invertebrates in burned areas than they did in
unburned habitats. This increase is strongly suggestive
of the pull that increased encounters with invertebrates
had in influencing foraging decisions in the burn. Previ-
ous research suggests that an immediate influx of inver-
tebrates postburn can act as a draw for vervet foragers
(Jaffe and Isbell, 2009). Immediately post-fire, immobile
invertebrate prey such as grubs and larvae that have
been cooked in the flames remain, their carcasses are
easily retrieved by foragers quickly moving into the
freshly burned landscape (Bouwman and Hoffman,
2007). Alternately, mobile prey fleeing from fire may be
stunned or immobilized by the smoke and flames making
their capture easier (bee smoking is a common honey
procurement strategy among human foragers worldwide
[Crane, 2013]). Within the first week of fire, invertebrate
prey may flee to unburned refuges within the broader
landscape of the burn. Aggregated here, they may be
easy targets for foragers (Jaffe and Isbell, 2009). How-
ever, immediate use of burned areas was not observed in
the present study. Rather, this population of vervets
appears to have been drawn into burned areas only after
they became repopulated by invertebrates also feeding
on nitrogen-rich new shoots (Swengel, 2001; Moya-
Raygoza and Larsen, 2014). These results bolster the
hypothesis that improvements in invertebrate search
efficiency, regardless of timing, motivated expansions
into burned areas previously described among this ver-
vet population and others (Jaffe and Isbell, 2009; Herzog
et al., 2014).
While the research presented here identified fire-
driven alterations in resource abundance, it failed to
detect shifts in resource profitability postburn. Several
interrelated issues limit our ability to accurately mea-
sure changes in the post-fire profitabilities of vervet
foods. First, we should only expect changes in profitabil-
ity (decreased time spent in handling events) among
resources directly impacted by flames. For vervets, this
subset of foods may be rather small, including only ses-
sile terrestrial resources such as USOs, slow and/or
immobile invertebrate prey, and fallen fruits and seeds
which may have become cooked on the ground surface
during the fires. However, these cooked resources are
quickly depleted by competitors. For example, avian spe-
cies, which are typically the first to arrive at a newly
burned patch, are attracted to the cooked and charred
fruits, seeds, and insects exposed after the removal of
ground cover. Fork-tailed drongos, lilac-breasted rollers,
and gray hornbills are known to exploit burned areas
until the abundance of recently deceased and fleeing
insects is exhausted (Bouwman and Hoffman, 2007).
Quick depletion of cooked foods precludes collection and
prevents any comparison of the profitabilities of cooked
resources to their uncooked counterparts. With access to
naturally cooked resources an issue, investigations of
the specific changes to nutrient availability via cooking
carried out in controlled laboratory settings (see: Kataria
et al., 1989; Clemente et al., 1998; Boback et al., 2007;
Carmody and Wrangham, 2009; Carmody et al., 2011;
Ee et al., 2011; Groopman et al., 2014; Zink et al., 2014)
may be one of the only ways to quantify changes in
cooked food profitabilities.
Finally, the preference for burned habitats may be
multidimensional. Jaffe and Isbell (2009) suggest that
burned areas offer primates an additional benefit in the
form of reduced predation. They note that primates for-
aging in burned areas have an increased ability to detect
predatory threats where reduced ground cover improves
overall visibility. At present, only limited work has been
done on the response of predators to fire. The outcomes
indicate that lions (Panthera leo) avoid hunting in
burned areas despite the fact that herbivore prey are
drawn to them (Eby et al., 2013). Given this supportive
data, future research involving the behavioral responses
of a larger suite of primary primate predators to fire is
warranted.
CONCLUSION
From chimpanzees to macaques and vervets, primates
spanning the Old World clades have been observed for-
aging in burned habitats, often extracting freshly cooked
resources (Harrison, 1983, 1984; Berenstain, 1986;
Armelagos, 2010). However, the effect may not be so
favorable for species dwelling in settings not adapted to
regular fire or for primates with highly constrained diets
(Berenstain, 1986; O’Brien et al., 2003). Vervets inhabit
a wide range of environments, and it is possible that
advantages for exploiting novel resources in varying con-
texts underlie the behavioral flexibility which enables
them to take advantage of foraging opportunities post-
fire (Jaffe and Isbell, 2009; Herzog et al., 2014).
These results and others suggest a likely precursor to
and foundation for the obligate pyrophilia that evolved
in our own lineage (Pruetz and LaDuke, 2010; Parker,
2014). Direct information about hominin plant use
(cooked or raw) is almost nonexistent in the archaeologi-
cal record, so hypotheses regarding types of foods con-
sumed and processing strategies cannot be easily tested
by that line of evidence. Because optimal foraging mod-
els track a simple currency directly related to evolution-
ary fitness, they provide a means to explore questions
about diet in our deep past (Kurland and Beckerman,
1985; Hawkes and O’Connell, 1992, 1995, 2006; O’Con-
nell et al., 1999, 2002; Griffith et al., 2010; O’Connell
and Allen, 2012; Sayers and Lovejoy, 2014). More sys-
tematic observations of our evolutionary cousins around
fire and burned landscapes can prompt investigators
interested in the question of hominin fire use to consider
the many and varied ways in which fire may have
impacted the foraging opportunities of hominins as they
moved into expanding, fire-prone savannas.
ACKNOWLEDGMENTS
The authors thank Alan Barrett, Jannie Coatzee, and
the Applied Behavioural Ecological & Ecosystem Research
Unit (ABEERU) at the University of South Africa, Pretoria
for their support. This manuscript was improved by com-
ments from Brian Codding, Jim O’Connell, Jill Pruetz,
Ken Sayers, Polly Wiessner, and one anonymous reviewer.
Research was approved by the South African Parks Board,
and the University of Utah IACUC Committee (protocol
#10-08007).
LITERATURE CITED
Aganga AA, Tsopito CM, Yeboah SO, Mokgoko C, Manne SN.
1997. Evaluation of the chemical composition of some locally
available Acacia seeds as animal feed in Botswana.
8N.M. HERZOG ET AL.
American Journal of Physical Anthropology

Proceedings of the XVIII International Grassland Conference.
Winnepeg, Manitoba. Vol. 2. p 23–27.
Altmann J. 1974. Observational study of behavior: sampling
methods. Behavior 49:227–267.
Altmann SA. 1998. Foraging for survival: yearling baboons in
Africa. Chicago: University of Chicago Press.
Andersen AN. 1988. Immediate and longer-term effects of fire
on seed predation by ants in sclerophyllous vegetation in
south-eastern Australia. Aust J Ecol 13:285–293.
Armelagos G. 2010. Creatures of the flame: light and heat in
human evolution. Evol Anthropol 19:158–160.
Archibald S, Bond WJ. 2004. Grazer movements: spatial and
temporal responses to burning in a tall-grass African
savanna. Int J Wildland Fire 13:377–385.
Baritell KV, van Schaik CP, Vogel E. 2009. Foraging efficiency
in wild orangutans (Pongo pygmaeus wurmbii) at the Tuanan
field station in Central Kalimantan, Indonesia. Am J Phys
Anthropol 138S:82–83.
Barrett AS. 2009. Spatial and temporal patterns in resource
dispersion and the structure of range use and co-existence in
a social omnivore (Chlorocebus aethiops) [Ph.D. dissertation].
Pretoria: University of South Africa.
Bennett NC, Jarvis JUM. 1995. Coefficients of digestibility and
nutritional values of geophytes and tubers eaten by southern
African mole-rats (Rodentia:Bathyergidae). J Zool 236:
189–198.
Berenstain L. 1986. Responses of long-tailed macaques to
drought and fire in Eastern Borneo: a preliminary report.
Biotropica 18:257–262.
Bird DW, Bliege Bird R, Parker CH. 2005. Aboriginal burning
regimes and hunting strategies in Australia’s Western desert.
Hum Ecol 33:443–464.
Bird DW, O’Connell JF. 2006. Behavioral ecology and archaeol-
ogy. J Archaeol Res 14:143–188.
Blaine KP, Lambert JE. 2012. Digestive retention times for
Allen’s swamp monkey and L’Hoest’s monkey: data with
implications for the evolution of cercopithecine digestive
strategy. Integr Zool 7:183–191.
Bliege Bird R, Bird DW, Codding BF, Parker CH, Jones JH.
2008. The “fire stick farming” hypothesis: Australian aborigi-
nal foraging strategies, biodiversity, in anthropogenic fire
mosaics. Proc Natl Acad Sci USA 105:14796–14801.
Boback SM, Cox CL, Ott BD, Carmody R, Wrangham RW, Secor
SM. 2007. Cooking and grinding reduces the cost of meat
digestion. Comp Biochem Physiol A Mol Integr Physiol 148:
651–656.
Bouwman H, Hoffman R. 2007. The effects of fire on grassland
bird communities of Barberspan, North West Province, South
Africa. Ostrich 78:591–608.
Bowman D, Garde M, Saulwick A. 2001. Kunj-ken makka man-
wurrk–fire is for kangaroos: interpreting Aboriginal accounts
of landscape burning in Central Arnhem Land. In: Anderson
A, O’Connor S, editors. Central Arnhem Land, histories of old
ages. Essays in honour of Rhys Jones. Canberra: Pandanus
Books, Research School of Pacific and Asian Studies, Austra-
lian National University. p 61–78.
Burton FD. 2009. Fire - the spark that ignited human evolu-
tion. Albuquerque: University of New Mexico Press.
Cameron AC, Trivedi PK. 2013. Regression analysis of count
data. Cambridge University Press.
Carmody RN, Weintraub GS, Wrangham RW. 2011. Energetic
consequences of thermal and nonthermal food processing.
Proc Natl Acad Sci USA 108:19199–19203.
Carmody RN, Wrangham RW. 2009. The energetic significance
of cooking. J Hum Evol 57:379–391.
Charnov EL. 1976a. Optimal foraging: attack strategy of a man-
tid. Am Nat 110:141–151.
Charnov EL, Orians GH. 1973. Optimal foraging: some theoreti-
cal explorations. Salt Lake City: University of Utah Press.
Clemente A, S
anchez-Vioque R, Vioque J, Bautista J, Mill
an F.
1998. Effect of cooking on protein quality of chickpea (Cicer
arietinum) seeds. Food Chem 62:1–6.
Codding BF, Bird DW. 2015. Behavioral ecology and the future
of archaeological science. J Archaeol Sci 56:9–20.
Codding BF, Bird RB, Kauhanen PG, Bird DW. 2014. Conserva-
tion or co-evolution? Intermediate levels of Aboriginal burn-
ing and hunting have positive effects on kangaroo
populations in Western Australia. Hum Ecol 42:1–11.
Conklin-Brittain NL, Knott CD, Wrangham RW. 2006. Energy
intake by wild chimpanzees and orangutans: methodological
considerations and a preliminary comparison. In: Feeding
ecology in apes and other primates. New York: Cambridge
Press. p 445–471.
Conklin-Brittain NL, Wrangham RW, Hunt KD. 1998. Dietary
response of chimpanzees and cercopithecines to seasonal vari-
ation in fruit abundance. II. Macronutrients. Int J Primatol
19:971–998.
Crane EE. 2013. The world history of bee keeping and honey
hunting. New York, NY: Routledge.
Davies NB, Krebs JR, West SA. 2012. An introduction to behav-
ioural ecology. Oxford, UK: Wiley.
De Ronde C, Trollope WSW, Parr CL, Brockett B, Geldenhuys
CJ. 2004. Fire effects of flora and fauna. In: Wildland fire
management handbook for Sub-Sahara Africa. Global Fire
Monitoring Center. p 60–87, Germany: Freiburg University.
Dominy NJ, Vogel ER, Yeakel JD, Constantino P, Lucas PW.
2008. Mechanical properties of plant underground storage
organs and implications for dietary models of early hominins.
Evol Biol 35:159–175.
Eby S, Mosser A, Swanson A, Packer C, Ritchie M. 2013. The
impact of burning on lion Panthera leo habitat choice in an
African savanna. Curr Zool 59:335–339.
Ee KY, Agboola S, Rehman A, Zhao J. 2011. Characterisation of
phenolic components present in raw and roasted wattle (Aca-
cia victoriae Bentham) seeds. Food Chem 129:816–821.
Eksteen JJ. 2003. Condensed management plan for Loskop Dam
Nature Reserve. Nelspruit, SA: Mpumalanga Parks Board.
Emlen J. 1966. The role of time and energy in food preference.
World Arch 34:103–130.
Filmalter N. 2010. A vegetation classification and management
plan for the Hondekraal section of the Loskopdam Nature
Reserve. Pretoria: University of South Africa.
Fisher WS. 1948. Burning the bush for game. Afr Stud 7:36–38.
Gowlett JAJ. 2010. Firing up the social brain. In: Dunbar RIM,
Gamble C, Gowlett JAJ, editors. Social brain, distributed
mind. London: British Academy. p 345–370.
Green DS, Roloff GJ, Heath BR, Holekamp KE. 2014. Temporal
dynamics of the responses by African mammals to prescribed
fire. J Wildlife Manage 79:235–242.
Grether GF, Palombit RA, Rodman PS. 1992. Gibbon foraging
decisions and the marginal value model. Int J Primatol 13:1–17.
Griffith CS, Long BL, Sept JM. 2010. HOMINIDS: An agent-
based spatial simulation model to evaluate behavioral
patterns of early Pleistocene hominids. Ecol Model 221:
738–760.
Groopman EE, Carmody RN, Wrangham RW. 2014. Cooking
increases net energy gain from a lipid-rich food. Am J Phys
Anthropol 156:11–18.
Hall M. 1984. Man’s historical and traditional use of fire in
southern Africa. In: Booysen PV de , Tainton NM, editors.
Ecological effects of fire in South African ecosystems (ecologi-
cal studies). Heidelberg: Springer Berlin. p 39–52.
Harrison MJ. 1983. Patterns of range use by the green monkey,
Cercopithecus sabaeus, at Mt. Assirik, Senegal. Folia Primatol
(Basel) 41:157–179.
Harrison MJ. 1984. Optimal foraging strategies in the diet of
the green monkey, Cercopithecus sabaeus, at Mt. Assirik, Sen-
egal. Int J Primatol 5:435–471.
Hawkes K, O’Connell JF. 1992. On optimal foraging models and
subsistence transitions. Curr Anthropol 33:63–65.
Herzog NM, Parker CH, Keefe ER, Coxworth J, Barrett A,
Hawkes K. 2014. Fire and home range expansion: a behav-
ioral response to burning among savanna dwelling vervet
monkeys (Chlorocebus aethiops). Am J Phys Anthropol 154:
554–560.
Isbell LA, Rothman JM, Young PJ, Rudolph K. 2013. Nutri-
tional benefits of Crematogaster mimosae ants and Acacia
FIRE AND PRIMATE FORAGING 9
American Journal of Physical Anthropology

drepanolobium gum for patas monkeys and vervets in Laiki-
pia, Kenya. Am J Phys Anthropol 150:286–300.
Jaffe K, Isbell LA. 2009. After the fire: benefits of reduced
ground cover for vervet monkeys (Cercopithecus aethiops).
Am J Primatol 71:252–260.
Kataria A, Chauhan BM, Punia D. 1989. Antinutrients and pro-
tein digestibility (in vitro) of mungbean as affected by domes-
tic processing and cooking. Food Chem 32:9–17.
Kelly RL. 2013. The lifeways of hunter-gatherers: the foraging
spectrum. Cambridge, UK: Cambridge University Press.
Kurland JA, Beckerman SJ. 1985. Optimal foraging and hominid
evolution: labor and reciprocity. Am Anthropol 87:73–93.
Lambert JE. 1998. Primate digestion: interactions among anat-
omy, physiology, and feeding ecology. Evol Anthropol 7:8–20.
Lewis HT, Ferguson TA. 1988. Yards, corridors, and mosaics:
how to burn a boreal forest. Hum Ecol 16:57–77.
MacArthur RH, Pianka ER. 1966. On optimal use of a patchy
environment. Am Nat 100:603–609.
Milton K. 1984. The role of food-processing factors in primate
food choice. In: Rodman PS, Cant J, editors. Adaptations for
foraging in nonhuman primates. New York: Columbia Univer-
sity Press. p 249–279.
Moya-Raygoza G, Larsen KJ. 2014. Response of ants to the leaf-
hopper Dalbulus quinquenotatus DeLong &Nault (Hemi-
ptera:Cicadellidae) and extrafloral nectaries following fire.
Sociobiology 61:136–144.
Nakagawa N. 1989. Foraging behavior of Japanese monkeys: a
viewpoint of optimal foraging strategy. Primate Res 5:1–13.
Nakagawa N. 1990. Choice of food patches by Japanese mon-
keys (Macaca fuscata). Am J Primatol 21:17–29.
Nakagawa N. 2003. Difference in food selection between patas
monkeys (Erythrocebus patas) and tantalus monkeys (Cerco-
pithecus aethiops tantalus) in Kala Maloue National Park,
Cameroon, in relation to nutrient content. Primates 44:3–11.
O’Brien T, Kinnaird M, Nurcahyo A, Prasetyaningrum M, Iqbal
M. 2003. Fire, demography and the persistence of siamang
(Symphalangus syndactylus: Hylobatidae) in a Sumatran
rainforest. Animal conserv 6:115–121.
O’Connell JF. 1995. Ethnoarchaeology needs a general theory of
behavior. J Archaeol Res 3:205–255.
O’Connell JF. 2006. How did modern humans displace Neander-
thals? Insights from hunter-gatherer ethnography and
archaeology. In: Conrad N, editor. Neanderthals and modern
humans meet? Tubingen: Kerns Verlag. p 43–64.
O’Connell JF, Allen J. 2012. The restaurant at the end of the
Universe: modelling the colonisation of Sahul. Aust Archaeol
74:5–16.
O’Connell JF, Hawkes K, Blurton Jones NG. 1999. Grandmo-
thering and the evolution of Homo erectus. J Hum Evol 36:
461–485.
O’Connell JF, Hawkes K, Lupo KD, Blurton Jones NG. 2002.
Male strategies and Plio-Pleistocene archaeology. J Hum Evol
43:831–872.
O’Malley RC, Power ML. 2012. Nutritional composition of
actual and potential insect prey for the Kasekela chimpanzees
of Gombe National Park, Tanzania. Am J Phys Anthropol
149:493–503.
Parker CH. 2014. On the evolution of human fire use [Ph.D.
dissertation]. Salt Lake City: University of Utah.
Parr CL, Chown S. 2003. Burning issues for conservation: a cri-
tique of faunal fire research in Southern Africa. Aust Ecol 28:
384–395.
Pruetz JD, LaDuke T. 2010. Brief communication: reaction to
fire by savanna chimpanzees (Pan troglodytes verus) at Fon-
goli, Senegal: conceptualization of fire behavior’’ and the case
for a chimpanzee model. Am J Primatol 141:646–650.
Pyke GH, Pulliam HR, Charnov EL. 1977. Optimal foraging: a
selective review of theory and tests. Q Rev Biol 52:137–154.
R Core Team. 2012. R: A language and environment for statisti-
cal computing. R Foundation for Statistical Computing:
Vienna, Austria.
Rothman JM, Raubenheimer D, Bryer MAH, Takahashi M,
Gilbert CC. 2014. Nutritional contributions of insects to pri-
mate diets: implications for primate evolution. J Hum Evol
71:59–69.
Sayers K, Lovejoy CO. 2014. Blood, bulbs, and Bunodonts: on
evolutionary ecology and the diets of Ardipithecus,Australo-
pithecus, and early Homo. Q Rev Biol 89:319–357.
Sayers K, Norconk MA, Conklin-Brittain NL. 2010. Optimal for-
aging on the roof of the world: Himalayan langurs and the
classical prey model. Am J Phys Anthropol 141:337–357.
Schoener TW. 1974. The compression hypothesis and temporal
resource partitioning. Proc Natl Acad Sci USA 71:4169–4172.
Schoener TW. 1987. A brief history of optimal foraging ecology.
In: Kamil AC, Krebs JR, Pulliam HR, editors. Foraging
behavior. New York, NY: Plenum Press. p 5–67.
Sensenig RL, Demment M, Laca EA. 2010. Allometric scaling
predicts preferences for burned patches in a guild of East
African grazers. Ecology 91:2298–2907.
Stephens DW. 2008. Decision ecology: foraging and the ecology
of animal decision making. Cogn Affect Behav Neurosci 8:
475–484.
Stephens DW, Krebs JR. 1986. Foraging theory. Princeton:
Princeton University Press.
Swengel AB. 2001. A literature review of insect responses to
fire, compared to other conservation managements of open
habitat. Biodivers Conserv 10:1141–1169.
Van de Vijver C, Poot P, Prins HHT. 1999. Causes of increased
nutrient concentrations in post-fire regrowth in an East Afri-
can savanna. Plant Soil 214:173–185.
Wandsnider L. 1997. The roasted and the boiled: food composi-
tion and heat treatment with special emphasis on pit-hearth
cooking. J Anthropol Archaeol 16:1–48.
Wehmeyer AS. 1986. Edible wild plants of southern Africa: data
on the nutrient contents of over 300 species. CSIR: National
Food Research Institute.
Wrangham R. 2009. Catching fire: how cooking made us
human. New York: Basic Books.
Zamora R, Alaiz M, Hidalgo FJ. 2001. Influence of cultivar and
fruit ripening on olive (Olea europaea) fruit protein content,
composition, and antioxidant activity. J Agric Food Chem 49:
4267–4270.
Zink KD, Lieberman DE, Lucas PW. 2014. Food material prop-
erties and early hominin processing techniques. J Hum Evol
77: 155–166.
Zuur A, Ieno EN, Walker N, Saveliev AA, Smith GM. 2009.
Mixed effects models and extensions in ecology with R. New
York, NY: Springer Science & Business Media.
10 N.M. HERZOG ET AL.
American Journal of Physical Anthropology