To the hunter go the spoils? No evidence of nutritional benefit to being or marrying a well‐reputed Hadza hunter

Abstract

Objectives:
The incentives underlying men's hunting acquisition patterns among foragers are much debated. Some argue that hunters preferentially channel foods to their households, others maintain that foods are widely redistributed. Debates have focused on the redistribution of foods brought to camp, though the proper interpretation of results is contested. Here we instead address this question using two nutritional variables, employed as proxies for longer-term food access. We also report on broader patterns in nutritional status.

Materials and methods:
We measured male hunting success, hemoglobin concentration and body fatness among bush-living Hadza. Hunting success was measured using an aggregated reputation score. Hemoglobin concentration, a proxy for dietary red meat, was measured from fingerprick capillary blood. Body fatness, a proxy for energy balance, was measured using BMI and bioelectrical impedance.

Results:
We find no statistically significant relationship between a hunter's success and any measure of his nutritional status or that of his spouse. We further find that: women are, as elsewhere, at greater risk of iron-deficiency anemia than men; men had slightly lower BMIs than women; men but not women had significantly lower hemoglobin levels than in the 1960s.

Discussion:
The absence of an association between hunting reputation and nutritional status is consistent with generalized food sharing. Null results are difficult to interpret and findings could potentially be a consequence of insufficient signal in the study measures or some confounding effect. In any event, our results add to a substantial corpus of existing research that identifies few nutritional advantages to being or marrying a well-reputed Hadza hunter.

Full text

RESEARCH ARTICLE
To the hunter go the spoils? No evidence of nutritional benefit
to being or marrying a well-reputed Hadza hunter
Duncan N. E. Stibbard-Hawkes
1
| Robert D. Attenborough
2,3
|
Ibrahim A. Mabulla
4
| Frank W. Marlowe
2†
1
Department of Anthropology, Durham
University, Durham, UK
2
Department of Archaeology and
Anthropology, Cambridge University,
Cambridge, UK
3
School of Archaeology and Anthropology,
The Australian National University, Canberra,
Australia
4
Institute of Resource Assessment, University
of Dar es Salaam, Dar es Salaam, Tanzania
Correspondence
Duncan N. E. Stibbard-Hawkes, Department of
Anthropology, Durham University, Dawson
Building, South Road, Durham, DH13LE, UK.
Email: duncanstibs@cantab.net
Funding information
Robinson College Cambridge; the Anthony
Wilkin Fund and the Ridgeway-Venn Fund;
The Leakey Foundation, Grant/Award
Numbers: RG75255, JEAG/056; The Ruggles-
Gates Fund of the Royal Anthropological
Institute; The Smuts Memorial Fund
Abstract
Objectives: The incentives underlying men's hunting acquisition patterns
among foragers are much debated. Some argue that hunters preferentially
channel foods to their households, others maintain that foods are widely
redistributed. Debates have focused on the redistribution of foods brought to
camp, though the proper interpretation of results is contested. Here we instead
address this question using two nutritional variables, employed as proxies for
longer-term food access. We also report on broader patterns in nutritional
status.
Materials and Methods: We measured male hunting success, hemoglobin concen-
tration and body fatness among bush-living Hadza. Hunting success was mea-
sured using an aggregated reputation score. Hemoglobin concentration, a proxy
for dietary red meat, was measured from fingerprick capillary blood. Body fat-
ness, a proxy for energy balance, was measured using BMI and bioelectrical
impedance.
Results: We find no statistically significant relationship between a hunter's suc-
cess and any measure of his nutritional status or that of his spouse. We further
find that: women are, as elsewhere, at greater risk of iron-deficiency anemia
than men; men had slightly lower BMIs than women; men but not women had
significantly lower hemoglobin levels than in the 1960s.
Discussion: The absence of an association between hunting reputation and nutri-
tional status is consistent with generalized food sharing. Null results are difficult to
interpret and findings could potentially be a consequence of insufficient signal in the
study measures or some confounding effect. In any event, our results add to a sub-
stantial corpus of existing research that identifies few nutritional advantages to being
or marrying a well-reputed Hadza hunter.
KEYWORDS
BMI, food sharing, hemoglobin, hunter–gatherers, hunting reputation
†
Died September 25, 2019
Received: 9 October 2019 Revised: 16 January 2020 Accepted: 10 February 2020
DOI: 10.1002/ajpa.24027
Am J Phys Anthropol. 2020;1–19. wileyonlinelibrary.com/journal/ajpa © 2020 Wiley Periodicals, Inc. 1

1|INTRODUCTION
“Large game were shared not like common goods, but in
ways that significantly advantaged producers’house-
holds…Men channeled the foods they produced to their
wives, children, and their consanguineal and affinal kin
living in other households.”
—Wood and Marlowe, 2013, p. 280, on Hadza hunting.
“Men targeted big game to the near exclusion of other
prey even though they were rarely successful and most of
the meat went to others, at a significant opportunity cost
to their own families.”
—Hawkes, O'Connell, & Blurton Jones, 2014, p. 596, on
Hadza hunting.
Among many hunter–gatherer populations, there is a strong sex-
ual division of labor (Bird, 1999; Kelly, 2013; Marlowe, 2007). Men
tend to pursue high-variance, high packet size resources, often includ-
ing hunted game, while women typically pursue smaller packet size
resources that are more reliably attained, with lower daily variance in
success, and less widely shared (Codding, Bird, & Bird, 2011). Sex dif-
ferences in both Hadza labor specialization (Crittenden, Conklin-
Brittain, Zes, Schoeninger, & Marlowe, 2013; Froehle et al., 2019) and
proclivity for risk-seeking behavior (Apicella, Crittenden, & Tobolsky,
2017) both appear in childhood.
For >30 years in hunter–gatherer studies, food-sharing and the
adaptive goals of male hunting/foraging effort have been extensively
debated (Bird, 1999; Gurven, 2004; Gurven & Hill, 2009; Hawkes
et al., 2014; Hawkes & Bird, 2002; Hawkes, O'Connell, & Blurton
Jones, 1991; Hawkes, O'Connell, & Blurton Jones, 2018; Wood &
Marlowe, 2013) and some authors have argued that male resource
acquisition and sharing patterns represent a puzzle or collective action
problem (see Hawkes et al., 2018; Stibbard-Hawkes, 2019a, for dis-
cussion). There are two prevailing schools of thought about men's
hunting goals and the sexual division of labor. Some (Kelly, 2013; Mar-
lowe, 2000b, 2003; Wood & Marlowe, 2014) argue that hunters and
their wives cooperate in order to maximize overall household resource
acquisition. Others (Bird, 1999; Bliege Bird & Bird, 2008; Hawkes,
O'Connell, & Coxworth, 2010) argue that men act not primarily to
provide their immediate or even extended families, but instead in
large measure to acquire foods that can be shared widely (Hawkes
et al., 2014). It has been proposed that such food-sharing serves to
show off or signal to potential mates, allies and rivals (Hawkes
et al., 2018).
This disagreement is central to a recent debate about Hadza
hunting and food redistribution patterns (Hawkes et al., 2014; Wood &
Marlowe, 2013; Wood & Marlowe, 2014). All parties acknowledge
that Hadza hunters engage in some food-sharing beyond their imme-
diate families. However, Hawkes, O'Connell, and Blurton Jones
(2001a) have shown that for all large game but those species >180 kg,
hunters’households can expect no higher a proportion of any carcass
he acquires than can others’households. Obversely, Wood and Mar-
lowe (2013) demonstrated, in a more recent study, that hunters’wives
and children receive a substantial preferential share of the meat he
brings back (Wood & Marlowe, 2013), such that, for example, a
hunter's household retained on average 42% of the in-camp shares of
the meat he produced while nonproducers’households received 11%.
Based on these findings, Wood and Marlowe (2013) calculate that the
households of the best hunters can expect 4.2 times more male-
produced calories in the long term than can those of the worst. Ber-
besque, Wood, Crittenden, Mabulla, and Marlowe (2016) have also
shown that men eat much of the food they acquire, mostly honey but
also some meat, out of camp, without sharing.
Although Wood and Marlowe (2013) and Hawkes et al. (2014)
agree that “Hadza men treat their wives and children differently than
they do those of other men”(Hawkes et al., 2014, p.599), Hawkes
et al. (2014) have contested Wood and Marlowe's interpretation of
their data and the extent to which hunting is an optimal family provi-
sioning strategy. First, Hawkes et al. (2014) highlight that Wood and
Marlowe's data were derived from in-camp sharing only, and excluded
meat eaten or redistributed at the site of the kill. Second, Hawkes
et al. (2014) highlighted that Wood and Marlowe classified animal
skins as cuts of meat, even though skins provide no direct nutritional
advantage. Third, Hawkes et al. (2014) argued that many of Wood
and Marlowe's data came from a camp experiencing a period of
unusually high meat availability, where the pressure to share was
atypically low. Fourth, Hawkes et al. (2014) highlight that the relative
size of procurers’households were unreported, and may have been
larger than the average household, artificially inflating procurers’meat
share. Hawkes et al. (2014) concluded that Wood and Marlowe's data
were consistent with their own previous reports of extreme food-
sharing and calculated, for example, that 82% of large game meat
eaten by any Hadza was procured by someone outside their own
household. Unlike Wood and Marlowe (2013), Hawkes et al. (2014)
concluded that Hadza hunters do not act primarily to provision their
households and that men's foraging decisions represent a “significant
opportunity cost to their own families”(p. 576). The above pair of
contradictory quotes, published within a year of each other, con-
cerning the same study population and, in large part, the same data,
illustrate the extent of this disagreement.
It is apparent, in all datasets, that a greater proportion of meat is
redistributed than is retained by the procurer's household. It also
appears in at least some circumstances, either ordinarily (Wood &
Marlowe, 2013) or when there is a surplus of meat (Hawkes et al.,
2001a, 2014), that the procurer's household can expect some prefer-
ential access to the meat he acquires. However, the degree to which a
Hadza hunter and his wife and children receive preferential access to
his hunting returns, and whether the households of better hunters
receive substantially more food in the long term, both remain disputed
questions (Hawkes et al., 2014; Wood & Marlowe, 2013; Wood &
Marlowe, 2014). Due to the extreme variance in daily hunting success
typical of the Hadza and many other populations (Hill & Kintigh,
2009; Marlowe, 2010; Kelly, 2013; Stibbard-Hawkes et al., 2018), and
questions over the generalizability (Hawkes, O'Connell, & Blurton
2STIBBARD-HAWKES ET AL.

Jones, 2001b; Wood & Marlowe, 2014) of published Hadza food shar-
ing data (e.g., due to methodological concerns or due to the impacts
on sharing of atypically high meat access), a definitive answer based
on return rates and patterns of sharing may yet require more data.
Certain measures of nutritional status may better reflect long-term
access to food than do currently available measures of daily hunting
returns and sharing patterns. An investigation into the relationship
between mid-to-long term nutritional status and hunting success,
therefore, may better address the question of whether a hunter and
his spouse benefit most from the food he produces. Such an investiga-
tion also allows us to address other questions, including whether
there are any sex differences in nutritional status and whether the
nutritional status has changed since previous studies.
Hadza nutritional status has been investigated several times
before. Bennett, Barnicot, Kagan, and Woodburn (1970) reported
hemoglobin concentrations during 1966 and 1967 among Hadza from
two sedentary settlements and six bush camps though were not con-
cerned with the significance of hunting and did not assess the impact
of hunting ability on hemoglobin levels. Since this period, no further
data on Hadza hemoglobin concentrations have been published. The
current study affords the ability to investigate whether any changes
have occurred since 1966–1967.
Sherry and Marlowe (2007) reported Hadza body mass index
[BMI] and Hadza body fatness measured using bioelectrical imped-
ance [BI], both cited as proxies for energy balance and energy avail-
ability. They found general homogeneity in BMI across ages and sexes
but showed that men between the ages of 30 and 45 had a “signifi-
cantly higher body fat level compared to other age groups”which they
hypothesized might be related to “better foraging skills in the bush, or
greater foraging efforts compared to younger and older males”
(p. 114). Other authors have also reported Hadza BMI data (Blurton
Jones, 2016; Pontzer et al., 2012) or data from which BMIs could be
calculated (Hiernaux & Hartono, 1980). However, with the exception
of Blurton Jones (2016), these studies did not explicitly aim to investi-
gate associations between nutritional status and hunting ability.
Several previous studies among the Hadza have looked for a rela-
tionship between hunting ability and some measure of children's and
spouse's nutritional status.
Hawkes, O'Connell, and Blurton Jones (2001b) related children's
growth rates to father's hunting return rate and found “no relationship
between father's hunting success [return rate] and children's weight
changes”(p.687). Hawkes et al. (2001b) did find such a relationship
within two subsets of their data: First, findings from 1985 and 1986
showed “more positive weight changes for the weaned children of
better hunters”, although they argued that this result was a conse-
quence of “small sample sizes and strong seasonal variation in weight
changes”(p.687). Second, when excluding the late dry season,
throughout which meat was abundant, from their data, Hawkes et al.
(2001b) did observe a relationship between father's hunting success
and child's weight gain. However, as seasonal weight changes were
not related, during this period, to father's seasonal hunting return rate,
they concluded the finding was not because children were eating
more meat from their father's kills. Instead, they concluded that this
pattern was due to the fact that more productive hunters had more
productive wives (though Wood & Marlowe, 2013, have disputed this
explanation).
Blurton Jones (2016) reported that the presence of a father had a
statistically significant positive effect on child weight between the
ages 5–15 and the presence of a non-expert hunter father had a very
small positive effect on children's survival. However, Blurton Jones
(2016) also found that “the [hunting] reputations of the father or step-
father do not predict [children's] higher weight”(p. 423). Moreover,
Blurton Jones (2016) found that fathers frequently nominated as good
hunters were associated with lower child survival (p.426). Blurton
Jones (2016) also found that “good hunters were no bigger than
other men”(p.330) and that having an expert hunter as a husband
had no effect on women's weight, BMI, upper arm circumferences,
or triceps skinfolds (SI16.2). Blurton Jones (2016) concluded that
Hadza marriage is “not an exchange for paternal care”(p.308) and
that a hunter's returns are “so widely and evenly distributed that
we can see no effect on individual women or their children due to
the individual men responsible for the acquisition of this
meat”(p.445).
Of further relevance are two studies by Marlowe (2000b, 2003).
The first briefly reported, in a sample of 55, that paternal hunting rep-
utation appears positively related to total number of living children, a
fact which Marlowe (2000b) took to indicate the importance of pater-
nal provisioning. The second reported that men provide more calories
to offset a drop in maternal productivity when their wives are nursing.
Marlowe did not, however, provide a direct assessment of nutritional
status and its relationship to hunting reputation or hunting
return rate.
Thus, while “father effects”on nutrition and survivorship have
previously been investigated among the Hadza (Blurton Jones, 2016;
Hawkes et al., 2001b), the relationship between a man's success and
ability as a hunter, and his nutritional status and that of his wife merits
further research. In the current study, we employ two measures of
nutritional status: hemoglobin concentration (in situ finger prick mea-
sure) and body fatness (BMI and bioelectrical impedance derived
estimates).
Hemoglobin cannot be synthesized within the body without
access to dietary iron. In the absence of pregnancy, disease or injury,
hemoglobin concentration is, within each sex, strongly related to iron
intake (Clark, 2008). Unlike daily hunting return measures, which may
fluctuate dramatically (Hill & Kintigh, 2009; Marlowe, 2010), blood
hemoglobin is buffered by iron stores and depends on iron intake over
a period of several months (Clark, 2008). The main source of iron in
the Hadza diet is meat from wild game (see Supporting Information
Material S2 for full discussion), almost all of which is collected by men
(Marlowe, 2010). If, as Wood and Marlowe (2013) have reported,
hunters and their wives and children receive a substantially higher rel-
ative share of the meat he acquires than do others in camp, then,
ceteris paribus, over a period of several months, hunters and wives of
hunters who acquire more meat might be expected to have eaten
more meat in total than less successful hunters and their wives.
Indeed, by Wood and Marlowe's (2013) estimates, better hunters and
STIBBARD-HAWKES ET AL.3

their households should receive 4.2 times more male-produced calo-
ries than the families of worse hunters.
In a resource-limited environment, hunted foods are also impor-
tant because they provide metabolic fuel. When energy intake
exceeds expenditure (positive energy balance), there is increased stor-
age of surplus metabolic fuel, largely as adipose tissue (Bender, 2014).
For this reason, body fatness is a good measure of long-term energy
balance. Assuming energy expenditure is constant, therefore, if better
hunters and their spouses have greater access to meat and other
foods, as predicted by Wood and Marlowe (2014), they should have a
more positive energy balance and, thus, all else held constant, greater
body fatness. Conversely, as with hemoglobin concentration, if meat
is widely shared and not preferentially channeled to a hunter's house-
hold, there should be no clear relationship between hunting reputa-
tion and body fatness.
Hunting reputation is often employed as a proxy for a hunter's
ability (e.g., Blurton Jones, 2016; Marlowe, 2000b; Stibbard-Hawkes,
2019b). However, previous methods of assessing hunting reputation
have relied on peer nomination, which only allows comparisons to be
made between a minority of frequently nominated hunters versus
the rest. Stibbard-Hawkes et al. (2018) introduced a novel measure
of hunting reputation which, unlike previous measures (Blurton
Jones & Marlowe, 2002; Marlowe, 2000b), allows granular distinc-
tions to be made between hunters at all levels of perceived ability.
Furthermore, this reputation measure has been shown to act as a
good proxy for several hunting skills and, it is inferred, generalized
hunting ability (Stibbard-Hawkes et al., 2018; Supporting Informa-
tion Material S1).
In order to assess the nutritional benefits of being or marrying a
good hunter, we here relate hunting reputation to hemoglobin con-
centration and two measures of body fatness; BMI and body fatness
estimated by bioelectrical impedance. In comparing results across sea-
sons, in reviewing the iron content of the Hadza diet and by dis-
cussing previous measures of the prevalence of anemia-causing
conditions both in text and in the associated Supporting Information
Material S2, we also comment on the likely impact of a number of
other potential confounding effects. Finally, we consider sex differ-
ences in both measures, as well as differences in reported hemoglobin
concentration between current results and measures reported by
Bennett et al. (1970).
2|MATERIALS AND METHODS
2.1 |Study population and field trips
Research was conducted among the Hadza, an ethnolinguistic group
living in Northern Tanzania in the area around Lake Eyasi. Roughly
1,000 people speak the Hadza language and, of these, some 150–250
still rely on hunting and gathering for the majority of their diet
(Blurton Jones, 2016; Marlowe, 2010). In the foraged Hadza diet,
roughly 32% of calories come from meat, 19% from berries, 19% from
tubers 15% from honey and 14% from baobab fruit (Marlowe, 2010).
The Hadza hunt more than 880 animal species, the majority birds
(Marlowe, 2010). There is a strong sexual division of labor where
men do the majority of hunting and women the majority of gathering
(Marlowe, 2010), although there is little to no further labor specializa-
tion reported within sexes (Apicella et al., 2017). Hadza hunters have
been characterized as large game specialists (Hawkes et al., 2014),
although this designation has recently been disputed (Berbesque
et al., 2016; Wood & Marlowe, 2014). Most Hadza hunting trips are
conducted solo (Berbesque et al., 2016), although men sometimes
cooperate in ambushes during the dry season when prey are more
clustered.
In recent years, this traditional diet has been increasingly aug-
mented with cultigens, bought using cash from engagement with the
tourist industry. Such dietary changes are especially prevalent in the
Mangola region, though such change has been happening to a lesser
degree across the entire Eyasi region. With the exception of one camp
which we believed may have been receiving grain handouts from a
nearby mission (detailed further in Section 3.2.1) we observed no evi-
dence of grain consumption in these camps. However, it is possible
that even in remote bush camps, people were consuming domesti-
cated foods at low frequencies in the months preceding our fieldwork.
This study was conducted over three field trips of circa 1 month
each. We visited a total of 17 camps. Trips were between 17th
August–17th September 2013 (mid-dry season); 7th December
2013–6th January 2014 (early wet season); 19th October 2014–27th
of November 2014 (late dry season, concluding early wet season).
Using season estimates from Marlowe (2010) in tandem with
observed rainfall in the area during the study periods, we designated
data collected in the first and third trips as “dry season”and the sec-
ond trip as “wet season.”Participants were all 17 years of age or older.
The final dataset used in this study included 130 individuals across all
nutritional measures, 59 women and 71 men. Thirty-two pairs of men
and women were married. Reputation rank data from 67 individuals
of both sexes were included in the final study. Nutritional measures
were exclusively collected outside Mangola in remote bush camps.
Here, participants were given gifts: shoes, blankets, soap, petroleum
jelly, plates, hammers, cold chisels, and other useful items. Some inter-
view data were collected in the Mangola region where Hadza increas-
ingly rely on cash-bought grain to augment their foraged diet. Here
local laws require remuneration with money and so, instead of provid-
ing gifts, we gave the equivalent of £12 per camp visit, to be shared
between all camp members (including non-participants). In all camps,
we assured people that they were free not to participate in data col-
lection and to drop out of the study at any time. We also made it clear
that those who decided not to participate in some/all measurements
would still receive remuneration.
2.2 |Demographic data and age estimation
During the first visit to each camp, demographic and anthropometric
data were collected from all participants. Demographic data were:
age, place of birth, parent's names, current spouse's name, number of
4STIBBARD-HAWKES ET AL.

children born, number of children still living, and the names of children
living in camp.
Although ages could not be verified by official documents, many
Hadza under the age of 45 years, especially those who have attended
school, are able to report (or approximate) their ages. Many older
Hadza cannot give exact ages, and some cannot provide approxima-
tions, a recurring problem in hunter–gatherer populations (Hill &
Hurtado, 1996). Frank Marlowe has been working with the Hadza for
over 20 years and Nicholas Blurton Jones, for more than 35 years.
Where data were available, we compared ages in our dataset to those
in the Marlowe–Blurton Jones dataset for those participants. Where a
participant did not know their age, or where there were substantial
discrepancies, we used those ages from Marlowe–Blurton Jones’
dataset. Where a participant was not found in previous demographic
data and did not know their age, we estimated that individual's age
visually.
2.3 |Height, weight, BMI, and % body fat
measurement
Height was measured using a portable stadiometer. Participants were
asked to remove their shoes, stand erect with their feet together and
flat, straighten their backs against the vertical ruler, and look directly
forward. Height was measured to the nearest millimeter from the
crown of the head.
Weight was measured using electronic scales. Participants were
asked to put aside any items they were carrying such as quivers, bows,
or digging sticks and to remove their shoes and heavy items of cloth-
ing. Most Hadza women wear two kanga (lightweight Tanzanian blan-
kets), while men wear shorts and occasionally t-shirts or a shuka
(larger Tanzanian blankets). These items of clothing are usually light. A
shuka is no more than 430 g and kangas are under half this weight.
For reasons of modesty, participants of both sexes were not asked to
remove their shorts, t-shirts, kanga or shuka.
The electronic scales included a battery-powered foot-to-foot
bioelectrical impedance device, as used by Sherry and Marlowe
(2007). Participants brushed clean their feet and placed them in posi-
tion on the metal pads of the device. The device reported body fat
percentages estimated via bioelectrical impedance. The manufac-
turer's documentation does not report the algorithm used to make
this estimation.
2.4 |Hemoglobin measurement
Hemoglobin concentration was measured in the field using a Hemo-
Cue Hb 301. The HemoCue is a portable, battery-operated device
that can be easily transported to the field and, if used with finger-prick
capillary blood, provides consistent results in a range of climatic con-
ditions (Morris, Osei-Bimpong, McKeown, Roper, & Lewis, 2007) with
minimal discomfort and a high degree of accuracy compared to labo-
ratory references (e.g., Jaggernath et al., 2016; Lardi, Hirst,
Mortimer, & McCollum, 1998; Morris et al., 2007). The measurement
procedure was conducted in accordance with the protocols set out in
the HemoCue Hb 301 online instructional videos.
1
Samples were
taken wearing sterile surgical latex gloves. The procedure was first
demonstrated on the finger of a researcher. Participants were advised
they may withdraw at any time. Blood samples were drawn from a fin-
ger on the participant's nondominant hand using a sterile
“Hemolance”spring-loaded lancet. “Normal flow”lancets were used
for women and young men. Many Hadza men over the age of 25 have
calloused hands. Where normal flow lancets were ineffective, “max
flow”lancets were used. The first two drops of blood were dabbed
away with a sterile disposable towel and the third drop collected in a
“Microcuvette”. Results were analyzed immediately. Individuals with
iron levels <130 g/L were provided a 3-week supply of iron supple-
ments. Provision was in place to convey any participant with danger-
ously low hemoglobin levels (<80) to the hospital although this was
never necessary. Spent materials were safely destroyed.
Visibly pregnant women were asked not to participate in this
measurement as were women who believed they were pregnant. As
Hadza women may have irregular menses (Marlowe, 2010), it can be
difficult for women to identify the early stages of pregnancy. It is
therefore possible that data include a small number of women in the
first trimester of pregnancy. During two of three field trips, no
researchers or research assistants were female. The Hadza treat both
nudity and bodily function with a degree of modesty. For this reason,
women were not asked when they had last menstruated.
2.5 |Hemoglobin adjustments and anemia
prevalence
Differences in altitude, tobacco consumption, and pregnancy may lead
to a higher risk of symptomatic anemia for identical hemoglobin con-
centrations. When inferring anemia from hemoglobin levels, it is stan-
dard practice to adjust raw hemoglobin data based on these
considerations (Sullivan, Mei, Grummer-Strawn, & Parvanta, 2008;
WHO, 2011). Although this article is not primarily concerned with
anemia prevalence—instead using hemoglobin concentration as a
proxy for dietary iron—formal cut-offs of the determination of iron-
deficiency anemia have some utility in distinguishing between those
who are iron-replete and those who are lacking in iron. For this rea-
son, minor adjustments were made to the reported hemoglobin as per
WHO guidelines. We report adjusted hemoglobin concentrations
throughout.
Camps in this study are at elevations of between 1,086 and
1,380 m above sea level. In accordance with World Health Organiza-
tion guidelines (WHO, 2011), we adjusted all hemoglobin concentra-
tions by −2 g/L to account for altitude. While some individuals may
occasionally smoke the equivalent of >10 cigarettes per day when
tobacco is available, access to tobacco is inconsistent, especially in
more remote camps. For this reason, all individuals were classified as
infrequent smokers (<10 cigarettes per day) and no adjustments were
made for tobacco consumption (see Sullivan et al., 2008). Similarly, as
STIBBARD-HAWKES ET AL.5

care was taken to exclude pregnant women from the hemoglobin
measure, no adjustments were made to account for pregnancy.
We determined anemia prevalence from adjusted hemoglobin
values using WHO anemia cut-offs (WHO, 2011) of <130 g/L for men
and <120 g/L for women (Table 1). Distributions were unchanged and
so results, except those analyses that directly incorporated anemia
prevalence, were not impacted by these adjustments.
2.6 |Hunting reputation
Reputation data were collected via interviews with 67 adults of both
sexes. These interviews took place after nutritional data had been col-
lected. Interviews were conducted in Swahili and Hadzane by a
Hadza-speaking research assistant with supervision from Duncan
Stibbard Hawkes. Each interviewee was shown a high-resolution
face-on photograph of each of the 71 hunters (all male) in the sam-
ple. They were asked to provide that hunter's first name, father's
name, and the amount of time since they last lived in the same camp
as the interviewee. Interviewees were deemed to “know”that
hunter if they answered the first two questions correctly and had
lived together within 2 years. Photos of those hunters whom the
interviewee knew were laid out before them in a random order.
Interviewees were asked to remove the best hunter and their selec-
tion was noted. This procedure was repeated until there were no
hunters remaining, providing a ranked list. A mathematical formaliza-
tion of this procedure is given by Stibbard-Hawkes et al. (2018) and
in Supporting Information Material S1.
Ranked lists were collated by taking the proportional rank of each
hunter in each list (i.e., the fraction of the way up each list that each
hunter appeared). We then calculated the mean of these scores for
each list in which a particular hunter appeared. This provided an
aggregated hunting reputation score for each of the 71 hunters in our
nutritional status dataset. This reputation measure showed a high
degree of agreement between interviewees and was positively related
to success on three important hunting tasks (Stibbard-Hawkes et al.,
2018). For this reason, the measure appeared to a viable proxy for
hunting ability/success more broadly (See Stibbard-Hawkes, 2019b;
Stibbard-Hawkes et al., 2018, for discussion).
TABLE 1 WHO (2011) cut-offs for mild, moderate and severe
anemia among men and women
Mild (g/L) Moderate (g/L) Severe (g/L)
Male 129–110 109–80 <80
Female 119–110 109–80 <80
FIGURE 1 Histogram of altitude-adjusted
hemoglobin concentrations for men (top) and
women (bottom) overlaid with sex-specific WHO
cut-offs for mild (gray) and moderate (red) anemia
6STIBBARD-HAWKES ET AL.

2.7 |Ethics, anonymization, data security, and
availability
Research was approved by the Cambridge Human Biology Research
Ethics Committee and conducted with permission from the Tanzanian
Commission for Science and Technology (COSTECH Permits:
2013-271-ER-2000-80 & 2014317-ER-2000-80). Hemoglobin measures
were conducted with permission from the Tanzanian National Institute
for Medical Research (NIMR Permits: NIMR/HQ/R.8a/Vol.II/387 &
NIMR/HQ/R.8c/Vol.IX/1536). Dates, research procedures, anemia rates
and other descriptive statistics related to hemoglobin measures were
reported to NIMR. All data were stored on a password-protected hard-
drive encrypted using Apple Firevault and anonymized using an ID known
only to a select group of Hadza researchers. Certain participants are iden-
tifiable by their hunting reputations, especially when accompanied by
their ages. For this reason, to protect participant's anonymity, we have
not made these data freely available online.
3|RESULTS
3.1 |Results: Hemoglobin
3.1.1 |Hemoglobin: Distribution and anemia
prevalence
The dataset included hemoglobin concentration measures from
127 adults, 71 from men and 56 from nonpregnant women aged
between 17 and 75. The hemoglobin concentrations (Figure 1) of
TABLE 2 Quadratic models of hemoglobin concentration (g/L) by
age (years) for men and women
Sex B SE B βR
2
p
Men 0.01 .71
Age −0.51 0.67 −.52 .45
Age
2
0.01 0.01 .48 0.49
Women 0.02 .67
Age 0.24 0.59 .27 .69
Age
2
−0.002 0.01 −.16 .82
Note: ***p< .001; **p< .01; *p< .05;
.
<0.1.
FIGURE 2 Boxplot of altitude-adjusted hemoglobin
concentrations by sex and season with medians and first and third
quartiles
TABLE 3 Shapiro–Wilk results for hemoglobin concentration
distributions by sex and season
Sex Season Wp
Men Dry 0.98 .50
Wet 0.97 .68
Women Dry 0.96 .10
Wet 0.98 .98
TABLE 4 Simple linear model of hemoglobin concentration by
mean hunting score
BSEBβR
2
Sig
0.0005 −0.015 0.85
Mean Hunting Score −1.91 10.16 0.85
Note: ***p< .001; **p< .01; *p< .05;
.
< 0.1.
FIGURE 3 Scatterplot of hemoglobin concentration by mean
hunting score, overlaid with a simple regression line and 95%
confidence band
STIBBARD-HAWKES ET AL.7

TABLE 5 Simple linear models of
wife's hemoglobin by husband's mean
hunting score and hemoglobin
BSEBβR
2
Sig
1 0.03 0.36
Husband's Mean Hunting Score −11.18 12.08 −0.17 0.36
2 0.03 0.31
Husband's Hemoglobin 0.15 0.14 0.18 0.31
3 0.05 0.44
Husband's Mean Hunting Score −9.80 12.22 −0.14 0.37
Husband's Hemoglobin 0.13 0.14 0.16 0.42
Note: ***p< .001; **p< .01; *p< .05;
.
< 0.1.
FIGURE 4 Scatterplots of wife's
hemoglobin concentration by
husband's mean hunting score (top)
and husband's hemoglobin
concentration (bottom), both overlaid
with a simple regression line and 95%
confidence band
8STIBBARD-HAWKES ET AL.

neither sex departed significantly from the expected normal distribu-
tion in a Shapiro–Wilk test and so are suitable for parametric testing.
Altitude-adjusted hemoglobin concentrations for men ranged from
99 to 172 g/L (range = 73) with a mean of 142.50 and a standard
deviation of 13.95. Altitude-adjusted hemoglobin concentrations for
women ranged from 102 to 154 g/L (range = 52) with a mean of
125.40 and a standard deviation of 12.72.
Anemia prevalence among. Hadza women were higher than
among Hadza men. Ten of 71 men (14.04%) had anemia by these
criteria. Of these, 8/10 had mild anemia while 2/10 (2.82%) had mod-
erate anemia (Figure 1). Anemia prevalence among.
Hadza women were higher than among Hadza men. Twenty of
56 women (35.71%) had anemia (Figure 1). Of these, 14 had mild ane-
mia (25.00%) while six were moderately anemic (10.71%). No one of
either sex had severe anemia.
3.1.2 |Hemoglobin: Age and seasonal effects
Neither age nor age-squared significantly predicted altitude-adjusted
hemoglobin concentrations among men or women (Table 2).
Hemoglobin levels were similar across seasons. Mean male altitude-
adjusted hemoglobin levels were 143.3 and 142.2 g/L in the wet and
the dry seasons respectively, while mean female altitude-adjusted
hemoglobin levels were 127.1 and 125.9 g/L (Figure 2). None of the
distributions of hemoglobin concentration (subsetted by sex and sea-
son) departed significantly from a normal distribution (Table 3). Fur-
thermore, variance in hemoglobin concentrations was similar for
women between the wet and the dry season, F(1, 54) = 1.05, p= .31,
but significantly different for men, F(1, 69) = 7.82, p= .01. For these
reasons, we used the Welch independent t-test to compare means.
Mean hemoglobin concentrations showed no statistically significant
difference between the wet and the dry seasons for either men, t
(59.54) = −0.37, p= .72, or women, t(18.57) = −0.81, p= .43.
3.1.3 |Hemoglobin concentration and hunting
reputation
A simple linear regression (Table 4) found no statistically significant
relationship between adjusted hemoglobin and hunting success
among men (Figure 3). Mean hunting score for those 10 hunters with
FIGURE 5 Histograms of BMI for
Hadza adults (top), men (bottom left) and
women (bottom right), overlaid with
WHO (1995) cut-offs for mild (18.5) and
severe (16) thinness
STIBBARD-HAWKES ET AL.9

mild or moderate anemia was higher (0.57) than for non-anemic
hunters (0.46), although an independent Welch two-sample t-test
found no statistically significant difference between the two groups, t
(11.7) = 1.9, p= .08. The sample size for anemic hunters is lower than
is recommended for the test and so these results should be inter-
preted cautiously.
Of the 59 non-pregnant women in our study, only 32 were mar-
ried to men in the current study. Hemoglobin values for these 32 mar-
ried women did not depart significantly from the expected normal
distribution in a Shapiro–Wilk test, W= 0.99, p= 1 and were suitable
for parametric testing. Women's altitude-adjusted hemoglobin con-
centrations showed no statistically significant relationship (Table 5) to
the hunting score or the hemoglobin concentration of their husband
(Figure 4).
3.1.4 |Current hemoglobin measures compared to
1966–1967 measures
Hadza hemoglobin concentrations have been reported once
before, by Bennett et al. (1970). We compare our results to their
hemoglobin figures from residentially mobile (reported as
“nomads”) Hadza from 1966 to 1967. Bennett et al.'s results were
unadjusted, so we use unadjusted results for comparison. We
could find no published research comparing the outputs of the
“Battery Operated Evans Electroselenium colorimeter”used by
Bennett et al. with those of the HemoCue Hb 301 used in the cur-
rent project or discussing the reliability of the device. However, as
Bennett et al. (1970) presented results in the same units and
within a similar range to those in the current study, the two
datasets appear comparable.
Hadza men in the 1966–1967 “nomadic”bush group over the age
of 20 (Bennett et al., 1970, p. 872) had a combined mean hemoglobin
concentration of 153.19 g/L (n= 21, SD = 12.85), while women over
the age of 20 had a combined mean hemoglobin concentration of
124.88 g/L (n= 17, SD = 6.62). Although the 1966–1967 dataset is
not available online, Bennett et al. provided standard errors which we
convert into standard deviations sufficient to perform a t-test. An
independent, two-sample t-test shows that mean hemoglobin concen-
trations for women from the 1966–1967 dataset were not signifi-
cantly different from those of women (mean = 127.4 g/L) in the
2013–2014 dataset, t(52.71) = 1.09, p= .28. Men in the 2013–2014
dataset had a significantly lower mean hemoglobin concentration
(144.50 g/L) than those in the 1966–1967 dataset, t(35.17) = −2.67,
p= .01, although the difference between the means was small
(<10 g/L).
Bennett et al. (1970) did not provide exact statements of ane-
mia prevalence but report that six of the roughly 150 individuals
over the age of 14 of both sexes in their dataset from either bush
camps or settlements had hemoglobin concentrations under
100 g/L. Bennett et al. (1970) provide no further information
about age distribution. There were no individuals under the age of
17 in our dataset although none of the combined men and women
in the 2013–2014 dataset aged ≥17 had unadjusted hemoglobin
concentrations <100 g/L.
FIGURE 6 Scatterplot of BMI by BI-estimated % body fat for
men and women, overlaid with a linear regression line and 95%
confidence band
TABLE 6 Regression models of men
and women's BMI and body fat by Age
and Age
2
Dependent variable B SE B βR
2
Adj R
2
Sig
Men's BMI 0.002 −0.01326 0.74
Age −0.004 0.01 −0.04 0.74
Men's BF 0.32 0.3124 0.00
Age*** 0.16439 0.03 0.57 0.00
Women's BMI 0.02 −0.0003 0.33
Age −0.02 0.02 0.33
Women's BF 0.05 0.03 0.14
Age 0.10 0.07 0.22 0.14
Note: ***p< .001; **p< .01; *p< .05;
.
< 0.1.
10 STIBBARD-HAWKES ET AL.

3.2 |Results: Body fatness
3.2.1 |BMI: Distribution, over/underweight, and
seasonal effects
Mean BMI in the current study was 20.3 (SD = 2.18): 19.6 for men
(SD = 1.5) and 21.4 for women (SD = 2.5). The difference between
sexes was statistically significant in a Welch two-sample t-test (t
[70.79] = −4.20, p< .000). Under and overweightedness were deter-
mined using WHO (Bailey 1995) cut-offs of <18.5 for mild thinness,
<16 for severe thinness, and >25 for overweightedness. The majority
of adult individuals (98/116) were not underweight (Figure 5). Eigh-
teen of 116 sampled, 3 women and 15 men, were underweight. Only
one participant had a BMI indicative of “severe thinness”. This individ-
ual was 17 years old, the youngest in the study. His BMI was probably
a result of his age, not chronic undernourishment. Four women had
BMI > 25, classified as “overweight”by the WHO. Two of these lived
in a camp near a small Christian mission and may have received grain
handouts for participating in prayer. Due to the presence of outliers,
both the combined BMI data and women's BMI results departed sig-
nificantly from a normal distribution (Figure 5), although a Shapiro–
Wilk test could not confirm that men's BMI data were non-normally
distributed. For this reason, non-parametric tests are used when relat-
ing women's BMI to other variables. As with hemoglobin, a Wilcoxon
rank-sum test could not reject the hypothesis that BMI measurements
in the wet (mean = 20.3, SD = 2.0) and the dry (mean = 20.4, SD = 2.3)
season are from identical populations, W= 1,274, p= .79.
3.2.2 |BI-estimated body fat percentages:
Distribution and seasonal effects
Mean BI-estimated % body fat was 16.0 for both sexes (SD = 7.2):
12.3 (SD = 4.2) for men and 21.5 (SD = 7.2) for women. The distribu-
tion of % body fat for men and women appears heavily skewed, and a
Shapiro–Wilk test indicated that the sample was from a non-normally
distributed population, W= 0.91, p= <.001, and, thus, not suitable for
parametric testing. Once again, a Wilcoxon rank-sum test could not
reject the hypothesis that wet season (mean = 15.3, SD = 6.5) and dry
season (mean = 16.29, SD = 7.4) % BF measurements were from iden-
tical populations, W= 1,409.5, p= .57. Among the men, there was
one clear outlier, an old man (age = 70 years) with 31.6% BF, 8.3 per-
centage points greater than the second “fattest”man in the study.
Although this man was living in the camp which may have received
grain handouts (see Section 3.2.1), he did not appear atypically fat and
we suspect equipment error in this case. This outlier was deemed
non-credible and excluded from the regression of BI-Estimated BF %
against hunting reputation.
FIGURE 7 Scatterplot of BMI and BI-
estimated % body fat by mean hunting
reputation or husbands mean hunting
reputation for men (top) and women
(bottom), overlaid with a linear regression
line and 95% confidence band
STIBBARD-HAWKES ET AL.11

3.2.3 |Relationship between BMI, BI-estimated %
body fat, and hemoglobin concentration
There was a clear positive correlation between BMI and BI-Estimated
% body fat (Figure 6); those individuals with higher body fat also had
a higher BMI, r
s
= 0.68, p< .001, n= 127. There were two noticeable
outliers (Figure 6), one woman with a high BMI relative to her body
fat and one man with a high body fat relative to his BMI, the same
individual identified in the previous paragraph. Removing these out-
liers did not greatly increase the Spearman's correlation coefficient
(r
s
= .69, p< .001, n= 125).
The relationship between hemoglobin concentration and BMI
was negative and statistically significant (r
s
=−.22, p= .02, n= 114).
The relationship between hemoglobin concentration and BF was also
negative and statistically significant (r
s
=−.36, p< .000, n= 114). It is
probable that these relationships were the product of sex effects and
disappeared when data were disaggregated. The relationship between
hemoglobin concentration and BMI was statistically non-significant
for both men (r
s
=−.05, p= .68, n= 67) and women (r
s
=−.02, p= .89,
n= 45). Similarly, the relationship between hemoglobin concentration
and BF was also statistically nonsignificant for both men (r
s
=−.17,
p= .16, n= 67) and women (r
s
= .04, p= .81, n= 45) when the sample
was divided by sex.
3.2.4 |BMI, BI-estimated % body fat and age
BMI showed little relationship to age among both men and women
and a multiple regression showed no statistically significant associa-
tion between the two variables (Table 6). BI-estimated % body fat was
positively related to age among both men and women although the
relationship was only statistically significant among men (Table 6).
3.2.5 |BMI, BI-estimated % body fat and hunting
reputation
The relationship between men's hunting reputation and BMI was posi-
tive (Figure 7), but did not approach significance (Table 7). The rela-
tionship between BI-estimated % body fat and hunting reputation
was positive and was stronger than the relationship between BMI and
hunting reputation. This relationship, although non-significant at the
0.05 level, approached significance (Table 7). When age and age
2
were
included, however, the relationship no longer approached significance
(Table 7). The relationship between women's BMI and their husband's
hunting reputation was negative (Figure 7) but did not approach sig-
nificance. This was also true of the relationship between women's BI-
estimated % body fat and husband's hunting reputation.
TABLE 7 Regression models men
and women's BMI and body fat by
hunting reputation/husband's hunting
reputation, with and without Age and
Age
2
included as a control
Dependent variable B SE B βR
2
Adj R
2
Sig
Men's BMI 0.01 −0.004 0.41
Mean Hunting Score 0.89 1.07 0.10 0.41
Men's BMI 0.05 0.01 0.30
Mean Hunting Score 0.73 1.11 0.08 0.52
Age 0.08 0.07 0.84 0.24
Age
2
−0.001 0.0008 −0.97 0.17
Men's % BF 0.06 0.04 0.05
Mean Hunting Score
.
4.50 2.29 0.23 0.05
Men's % BF 0.35 0.32 0.00
Mean Hunting Score 1.60 2.02 0.08 0.43
Age** 0.41 0.13 1.82 <0.00
Age
2
*−0.004 0.002 −1.32 0.03
Women's BMI 0.04 0.002 0.31
Mean Hunting Score −2.84 2.77 −0.19 0.31
Women's BMI 0.18 0.08 0.16
Mean Hunting Score −4.58 2.93 −0.31 0.13
Age
.
0.53 0.26 2.85 0.05
Age
2
*−0.006 0.003 −2.90 0.05
Women's % BF 0.02 −0.01 0.42
Mean Hunting Score −6.03 7.32 −0.15 0.42
Women's % BF 0.16 0.06 0.21
Mean Hunting Score −12.43 7.76 −0.32 0.12
Age
.
1.34 0.69 2.78 0.06
Age
2.
−0.02 0.008 2.61 0.08
Note:*** p< .001; ** p< .01; * p< .05;
.
< 0.1.
12 STIBBARD-HAWKES ET AL.

4|DISCUSSION
4.1 |Discussion: Hemoglobin concentrations
4.1.1 |Factors potentially confounding the
relationship between hemoglobin concentration and
red meat access
The human body normally contains 3–5 g of iron, which is lost,
through various processes, at a rate of 1–2 mg per day (Pantopoulos,
Porwal, Tartakoff, & Devireddy, 2012). Although iron may be present
in the myoglobin of the muscle tissue (300 mg) and may be stored in
the spleen (30 mg) and liver (1 g), the majority of iron present in the
human body takes the form of heme in the hemoglobin of erythro-
cytes (>2 g; Pantopoulos et al., 2012). Red blood cells die and are rep-
laced daily, surviving for an average of 120 days (Clark, 2008). When
insufficient dietary iron is available, iron reserves are gradually
depleted as iron is lost. In conditions where dietary iron is lacking, and
in the absence of other confounding effects, therefore, hemoglobin
concentration provides an estimate of iron intake averaged over the
previous 6 months (Clark, 2008). As, in this study, hemoglobin con-
centration is used as a proxy for dietary iron and specifically access to
red meat, it is important to account for those factors which may con-
found this relationship. These are (a) iron repletion, (b) disease (espe-
cially helminthic infection and malaria), and (c) other sources of iron
(or micronutrients which modulate the absorption of iron) in the
Hadza diet. Because of the high variation in hemoglobin levels
between participants, generalised iron repletion is unlikely to have
confounded results. Similarly, diseases causing iron leaching have his-
torically been infrequent among the Hadza (Bennett et al., 1970) and
so probably have not introduced significant confounding. There are
very few major sources of iron in the Hadza diet beyond the easily
absorbed heme iron from hunted meat. Grewia spp. berries are a nota-
ble exception. However, these are ubiquitously available when in sea-
son and consumed by everyone (Marlowe, 2010), so there are
probably few significant inequalities in access and consumption. These
factors are discussed further in Supporting Information Material S2.
4.1.2 |Hemoglobin and hunting success
Proponents of the provisioning hypothesis (e.g., Wood & Marlowe,
2013, 2014) argue men hunt primarily to provision themselves and
their families. Conversely, Hawkes et al. (2014) have argued that food
is widely shared as a means of male status competition resulting in “col-
lective benefits [i.e. wide and generous food-sharing with other camp
members] that would not be supplied if men foraged mainly to provi-
sion their own households”(p. 596). In the current dataset, a hunter's
hemoglobin concentration was unrelated to his hunting reputation. Fur-
thermore, there was no statistically significant relationship between
women's hemoglobin concentration and husbands’hunting reputation.
To investigate this question further, we split the data into two
categories—those without anemia (iron-replete) and those with
anemia. Among men, we found no statistically significant difference in
hunting reputation between the two groups. Although the sample size
for anemic men, at 10, was small, this result is consistent with Hawkes
et al.'s conclusion that greater hunting ability does not disproportion-
ately advantage the producer's household and that food is widely
shared. Due to small sample size, we did not conduct a similar mean
comparison for anemic/nonanemic women married to hunters in the
current study.
The current dataset provides no evidence that hunters with bet-
ter reputations and/or their spouses had greater access to meat. Thus,
results matched the findings of Blurton Jones (2016) which found lit-
tle impact of hunting reputation (measured via nominations, rather
than rankings and see Stibbard-Hawkes et al., 2018 for discussion) on
wife's nutritional status in an earlier dataset. Present results did not
match the finding of Wood and Marlowe (2013) that wives of better
hunters systematically received substantially more meat than did
women outside their households. However, unlike some other studies
(Blurton Jones, 2016; Hawkes et al., 2001b) which find no unequivo-
cal positive impact of fathers return rate (Hawkes et al., 2001b) or
hunting reputation (Blurton Jones, 2016) on children's nutrition, the
current data provide no information concerning the nutritional status
of the hunter's children. Although very large kills may be shared with
individuals in other camps, most food sharing takes place within
camps. Hadza move between camps regularly, between 6.5 and
9 times a year (Marlowe, 2010), which probably dilutes any camp-level
effects. However, recent research among the Hadza has found that
unlike Agta hunter–gatherers (Smith et al., 2019) and despite high levels
of residential mobility (Marlowe, 2010), individuals assort based on pro-
pensity for cooperation (Smith, Larroucau, Mabulla, & Apicella, 2018).
It would therefore be interesting to assess whether hemoglobin
concentrations were more similar within camps than between them.
This assessment was not possible in the current study as many of the
camps had fewer than 10 adult members. This question may, how-
ever, be fertile ground for further a study drawing on a larger sample
of camps, and focusing primarily on larger camps. Results are, in gen-
eral, more consistent with the findings of Hawkes et al. (1991, 2014),
that hunters widely shared meat beyond their households and the
finding of Blurton Jones (2016) that men with better hunting reputa-
tions appeared to provide little extra benefit to their families. How-
ever, as discussed in Section 5, this interpretation is not definitive.
4.1.3 |Sex differences in hemoglobin
concentration and anemia risk
In the current study, we found notable sex differences in anemia prev-
alence. More than a third (35.71%) of women in the sample had some
form of anemia, 25.00% mild and 10.71% moderate. In contrast,
14.04% of men had some form of anemia, and only two men (2.82%)
were moderately anemic. These figures are close to global anemia
prevalence rates which, in the most recent World Health Organization
world population survey, were 30.20% for nonpregnant women and
12.70% for men (WHO, 2005).
STIBBARD-HAWKES ET AL.13

These sex differences are possibly a product of greater blood loss
experienced by women through menstruation, which accounts for
much of the global disparity also. However, there is some evidence
that women in natural fertility populations have a substantially lower
number of lifetime menses than contraceptive populations
(e.g., Bentley, 1985; Howell, 1979; Strassmann, 1999 and reviewed by
Fitzpatrick, 2018, p. 113–118). This is multi-causal and due in part to
energetic stress via regular activity (Bentley, 1985) and in part to a
higher proportion of the lifespan spent in pregnancy or lactational
amenorrhoea (Strassmann, 1999). However, high fidelity data on the
topic are rare. There are anecdotal reports that Hadza women have
irregular menses (Marlowe, 2010, p. 152). Interviews conducted in
2015 with Hadza women did not confirm these reports, though did
show that Hadza women typically had short menses, with a mean
length of 2.3 days (Fitzpatrick, 2018). As a result, it is possible, that
women in the study population experience less severe menstrual iron
loss than is the global norm.
The observed sex differences in anemia risk are also concordant
with previous work showing sex differences in Hadza dietary compo-
sition. Berbesque, Marlowe, and Crittenden (2011) found that Hadza
men typically eat more meat as a proportion of their total in-camp diet
than women, while women eat more tubers, an observation in line
with related sex differences in dental micro-wear (Berbesque et al.,
2012) and food preference (Berbesque & Marlowe, 2009). Moreover,
out of camp, women mainly eat tubers and eat very little meat
(Berbesque et al., 2011; Marlowe, 2010) while meat, usually from
small game, comprises 11% of men's out of camp diet (Berbesque
et al., 2016). Finally, several iron-rich cuts of meat including the heart,
kidneys, back, lungs, and tongue of most larger game animals
(Marlowe, 2004, p.689) are generally reserved for adult epeme men
(those men over the age of 30 or who have killed a large animal), fur-
ther compounding sex disparities in access to heme iron.
It should be further noted that while anemia rates are higher for
Hadza women than Hadza men, both are substantially lower than
those typically reported elsewhere in the country. Tanzania has a
country-wide anemia prevalence of 46.90% for nonpregnant women.
No country-wide male anemia rate is listed in the WHO global data-
base on anemia although those male anemia rates in most regions of
Tanzania from which data were available exceeded 40.00%
(WHO, 2006).
4.1.4 |Changes in dietary iron since 1966–1967
Although current accounts of Hadza lifeways bear great similarity to
historic accounts (Marlowe, 2010), there have been many recent
changes in the region. The utilization of Hadza foraging territory by
herders from several neighboring populations, although noted in the
1950s–1960s (Woodburn, 1964), has increased in recent years. The
clearing of thorn bushes to build cattle enclosures (“boma”) and to
facilitate grazing has probably impacted game populations. This is
implied by (a) a decrease in game numbers observed during aerial sur-
veys both in the Eyasi region and throughout the country between
1977 and 2001 (reviewed by Blurton Jones, 2016, p.30) and (b) an
observed dip in frequency of large game kills from 3% of days
(Hawkes, 1991) to 0.97% of days (Wood & Marlowe, 2013) and an
associated drop from 4.89 kg (Hawkes, 1991) of meat per day to
1.08 kg (Wood & Marlowe, 2013) of meat per day (estimates from
Blurton Jones, 2016 based on data collected by Hawkes, 1991 and
Wood & Marlowe, 2013). Furthermore, as noted elsewhere, although
ethnotourism is more prevalent in Mangola than in more remote
areas, it happens throughout the whole Lake Eyasi region. It is proba-
ble that even in the study areas some cash-bought domesticated culti-
gens were coming into camp though at low frequency.
One might, therefore, expect a drop in hemoglobin levels since
the 1966/1967 period of fieldwork (Bennett et al., 1970) when Hadza
hemoglobin concentrations were last measured. Contrary to expecta-
tions, women's hemoglobin concentrations showed no statistically sig-
nificant difference from their 1966/1967 levels. Men's hemoglobin
concentrations were, however, a mean 8.69 g/L lower. This difference
was small in absolute terms, and it should be stressed that people are
still getting much meat. While we do not report returns measures, we
observed recently killed large animal carcasses in 6/17 camps visited
during the study period: One was a lesser kudu (Tragelaphus imberbis),
one a common eland (Taurotragus oryx), one an impala (Aepyceros
melampus), and the remaining three were antelope that had been too
substantially butchered to identify. Nevertheless, findings are consis-
tent with other evidence showing a decline in dietary quality and,
although 8.69 g/L hgb is small when considered within the normal
range of human hemoglobin concentrations, at the population level
this mean difference probably corresponds to a real increase in ane-
mia risk. However, as distributions were not reported by Bennett
et al. (1970) it is not possible to say with certainty how much anemia
has increased. The fact that changes have impacted men's hemoglobin
concentrations more than women's is notable. This finding is concor-
dant with the previously discussed fact that men eat more and more
iron-rich meat than women (Marlowe, 2010) both in (Berbesque et al.,
2011) and out (Berbesque et al., 2016) of camp.
4.2 |Discussion: BMI and body fatness
4.2.1 |BMI related to body fatness
The BMI and bioelectrical impedance variables are both imperfect
measures of body fatness (Kyle et al., 2004; Prentice & Jebb, 2001). In
the case of BMI, individuals with a high bone or muscle mass may be
misleadingly diagnosed as overweight (Prentice & Jebb, 2001). Fur-
thermore, associations between BMI and % body fat may be different
for certain populations and ethnicities (Mascie-Taylor & Goto, 2007;
Stevens, 2003).
BI-estimated body fat percentages have different issues. First,
commercial BI devices are not standardized and differ between manu-
facturers. Furthermore, the specific method used for calculating body
fat %, as well as the actual impedance measure itself, is often, as in
the present case, hidden from the end user.
2
For this reason,
14 STIBBARD-HAWKES ET AL.

commercial BI-devices may not be fit for use in clinical settings.
Although some have raised questions about their accuracy (Kyle et al.,
2004), studies of consumer-grade foot-to-foot BI devices (Bosy-
Westphal et al., 2008; Peterson, Repovich, Eash, Notrica, & Hill, 2007;
Wang & Hui, 2015) show that devices from different brands typically
produce highly correlated results (Pearson's correlation coefficients
0.82–0.96) and agree with other methods of assessing body fat.
Here the strength of the relationship between BMI and BI-estimated
body fatness, although not 1:1, was comparable to that reported in larger
clinical studies (e.g., Flegal et al., 2009). This strong positive relationship
gives confidence that, despite the concerns of some authors over the reli-
ability of consumer BI devices (e.g., Buchholz, Bartok, & Schoeller, 2004)
and of the universal applicability of BMI (Deurenberg, Deurenberg-Yap, &
Guricci, 2002), both measures assessed broadly the same phenotype and
the equipment gave a good representation of individual differences in
body fatness. As this study is not a diagnostic exercise, and is primarily
concerned with population-level patterns, both measures were deemed
sufficiently reliable for current purposes.
4.2.2 |Body fatness and hunting reputation
The current study found no evidence that the wives of more well-reputed
hunters had higher body fatness or, therefore, a more positive energy bal-
ance. Although our method of assessing hunting reputation was different
(see Stibbard-Hawkes et al., 2018) and our data were more recent, this
finding is identical to that reported by Blurton Jones (2016) in data from
the 1980–1990s. In fact, the wives of well-reputed hunters in the present
study tended to have lower body fatness, although this relationship was
statistically nonsignificant. Men's body fatness results, conversely, show a
slight positive increase with hunting reputation and, in the case of BI-
estimated % body fat, this relationship approaches statistical significance,
consistent with the notion that better hunters benefit from their own
self-provisioning. However, when age and age
2
are included in the model
as controls, the relationship between % body fat and hunting reputation
no longer approaches statistical significance. Present results are thus more
consistent with the conclusions of Hawkes et al. (2014) and Blurton Jones
(2016) that hunted foods are widely shared and that hunting success does
not substantially benefit the producer's household. Results are also con-
cordant with the findings of Hawkes et al. (2001b) and Blurton Jones
(2016) that father's hunting return rate and reputation have no unequivo-
cal (or even negative) impact on child growth, although as we possess no
data on body fatness among individuals aged 16 and under we cannot
comment further. However, there are several other factors that might
confound results, namely, the influence of women's contributions to the
diet as well as unmeasured differences in energy expenditure, discussed
in Section 5.
4.2.3 |Sex differences in body fatness and BMI
Unlike three previous studies (Hiernaux & Hartono, 1980; Pontzer
et al., 2012; Sherry & Marlowe, 2007) which find general similarity
between sexes, we find that Hadza men had lower BMI than women
(19.6 vs. 21.4). Our results were very similar to the findings of Blurton
Jones (2016) who also reported that women had marginally but signif-
icantly higher BMI than men (20.25 vs. 19.86, p= .011, two-sample t-
test; figure 16.1). Although the difference in means was statistically
significant, as with data from Blurton Jones (2016), the difference was
small in relative terms and close to means reported elsewhere. Men in
the study had lower body fat percentages than women (12.3 vs. 21.5),
a normal pattern globally (e.g., Karastergiou, Smith, Greenberg, &
Fried, 2012). Although there were some underweight individuals, only
one person in the study was severely underweight, suggesting that
most people in the study of both sexes have healthy energy balances.
Berbesque et al. (2011) have previously observed that, within
camp, women eat less meat but eat more food overall than do men by
eating frequency. Given that Hadza men both have a higher physical
activity level (e.g., men travel 12.2 km per day and women travel
6.2 km per day) and higher total energy expenditure (2,649 kcal
vs. 1,877 kcal per day), the small difference in BMI here and the
homogeneity of BMI observed in other studies (Pontzer et al., 2012;
Sherry & Marlowe, 2007) suggest that if men get less food than
women within camp (Berbesque et al., 2011), they must acquire other
sources of food to achieve parity. Results are therefore concordant
with findings that men acquire a substantial proportion (mean daily
average = 2,405 kcal) of their caloric intake out of camp (Berbesque
et al., 2016) a pattern which begins to emerge in childhood
(Crittenden et al., 2013).
It is surprising, given the evidence of male reliance on out-of-
camp self-provisioning, that hunting success has so little impact on
male nutrition both here and in data reported by Blurton Jones
(2016). However, game comprised a relatively small proportion of
men's out of camp energy intake, 85% of which came from honey
(Berbesque et al., 2016), so high levels of energy intake via self-
provisioning are still consistent with the finding that hunting skill has
little impact on men's energy balance.
5|STUDY LIMITATIONS, COMPLICATING
FACTORS, AND DIRECTIONS FOR FURTHER
RESEARCH
An important limitation of the study is the use of hunting reputation
as a proxy for hunting success. The efficacy of reputation measures
has historically been questioned due to the potential error introduced
for example by halo effects (Kelly, 2013), by general uncertainty over
the “true”hunting ability of raters’peers (e.g., Hill & Kintigh, 2009),
and by the lag between decline in actual foraging efficacy and decline
in reputation.
3
These concerns were central to our development of a
novel method for assessing hunting reputation (Stibbard-Hawkes
et al., 2018). We concluded that this measure exhibited a good level
of internal validity (agreement between interviewees) and external
validity (predicting independently measured hunting skills). No reputa-
tion measure is free from noise and it remains possible that the mea-
sure is not powerful enough to detect small but real relationships
STIBBARD-HAWKES ET AL.15

between the variables under study. Due to the difficulties in assessing
hunting skill using returns measures without prohibitively large sample
sizes (e.g., see Hill & Kintigh, 2009, for analysis of !Kung and Aché
returns data), direct measurements of <600 days per individual may
not be more reliable than reputation scores (Stibbard-Hawkes et al.,
2018) and we have no further suggestions concerning improvements
that could be made by further research in this regard. Moreover,
where previous studies have employed male returns measures, they
have had similarly little unequivocal impact on measures of familial
nutritional status (Hawkes et al., 2001b but see Wood & Marlowe,
2013, 2014).
There are some further factors that might complicate the relation-
ship between a hunter's ability and his nutritional status. Among them
is the fact that, if better hunters have more offspring (e.g., Blurton
Jones, 2016), any additional nutritional benefit gained might be
diluted by a greater investment in reproduction (e.g., Marlowe, 2003)
and into provisioning a greater number of dependent offspring, per-
haps to the extent that nutritional benefits to self and spouse become
invisible. The impact of hunting reputation on male RS is “significant
but modest”(e.g., Blurton Jones, 2016, p. 274) and so it appears to us
unlikely that the provisioning efforts of better hunters are diluted
greatly by extra children. Moreover, as elsewhere stated, previous
research has found little clear evidence of father's hunting success
effects on child nutrition (e.g., Blurton Jones, 2016; Hawkes et al.,
2001b). However, it is yet possible that male provisioning could alter-
natively be channeled into greater spousal fecundity, rather than child
nutrition (e.g., see Marlowe, 2000a).
Furthermore, it is possible that if better hunters do get more food,
additional energy is not uniformly channeled into fat deposition. Extra
energy garnered through self-provisioning could be channeled either
into other physiological processes (e.g., immune regulation, somatic
repair) or other behaviors (see discussion in Pontzer et al., 2015). It is
notable, for example, that better Hadza hunters had greater upper-
body strength in two independent studies (Apicella, 2014; Stibbard-
Hawkes et al., 2018), a relationship which could potentially be bidirec-
tional. Further research, in investigating the relationship between
hunting reputations and other markers of health, stress, and somatic
maintenance, might better address this question. These concerns do
not extend to hemoglobin measures.
Since we did not collect female foraging success data in the pre-
sent study, it is further possible that individual differences in body fat-
ness, though probably not hemoglobin concentrations, have more to
do with differences in women's contributions to the diet than with
men's hunting returns. Among the Hadza, Marlowe (2010, p. 214)
found that married women acquire slightly more energy per foraging
hour (764 kcal) than do married men (666 kcal). Berbesque et al.
(2011) showed that 55% of men's in-camp diet consists of tubers,
berries, and baobab and that 88% of these three food types were
acquired by women. It is therefore possible that the wives of less able
hunters offset their lack of productivity. Given evidence for
assortativity in partner choice and the finding that better hunters
marry “harder-working wives”(Hawkes et al., 2001b, p. 689) this
appears unlikely. Further, as Hadza women typically forage together,
it is likely that women's contribution to the diet is less variable than is
mens’(Berbesque et al., 2011) and so women's contributions are not
expected to introduce substantial confounding.
Finally, we collected no data concerning energy expenditure or
physical activity levels. We cannot rule out the possibility that better
hunters expend more energy in hunting and therefore have a similar
energy balance to worse hunters even though they get more food
from hunted meat. Hadza daily total energy expenditure and physical
activity levels have been reported elsewhere (Pontzer et al., 2012,
2015). Surprisingly, Pontzer et al. (2015) found daily total energy
expenditure to be unrelated either to daily physical activity level (mea-
sured via walking speed and total daily distance traveled) or to forag-
ing success, a fact they argued may be due to compensation for
increased foraging activity in other areas either through behavior
(e.g., resting, reducing activity in other areas of life) or physiological
mechanisms. The extent to which differences in total energy expendi-
ture or physical activity may confound results is therefore unclear.
Detailed data concerning energy expenditure and activity level, along-
side hunting ability measures and over a large sample, would be
enlightening.
6|CONCLUSIONS
Hunting reputation in the current study was not found to have any
statistically significant relationship with body fatness or hemoglobin
levels. These results are consistent with previous reports that a Hadza
hunter's food is widely redistributed and not substantially monopo-
lized by the hunter or his wife (Hawkes et al., 2001a, 2014). Current
results are further concordant with previous research showing that
fathers with high hunting return rates (e.g., Hawkes et al., 2001b) and
hunting reputations (Blurton Jones, 2016) provide few palpable nutri-
tional benefits (and perhaps even some minor disadvantages, see
Blurton Jones, 2016) to their children. Our findings replicate those of
Blurton Jones (2016), who reported that better hunters were not
larger than worse hunters and that marrying a better hunter had no
observable impact on women's BMI. Blurton Jones's results were
derived from data collected in the 1980s and 1990s, and using a dif-
ferent measure of hunting reputation. The fact that we find similar
results in the present study gives us some confidence in the long-term
validity of these findings.
For both nutritional status variables used here there were several
potential confounding effects: disease, iron repletion and iron from
Grewia in the case of hemoglobin measures, (discussed in Supporting
Information Material S2); wife's contribution to diet and variation in
energy expenditure in the case of body fatness. Moreover, variation
in hemoglobin concentrations does suggest that there are inequalities
in meat access among men even though these here seem unrelated to
hunting reputation. Furthermore, although the hunting reputation
measure used here improved upon previous reputation measures in
several key respects, showed high internal validity and showed a good
correlation with several independent measures of hunting skill
(Stibbard-Hawkes et al., 2018), it is a proxy rather than a direct
16 STIBBARD-HAWKES ET AL.

measure. It is difficult to definitively determine whether results repre-
sent a real absence of relationship or a type 2 error. Finally, it is possi-
ble that better hunters and their wives do eat more but channel the
energy into other activities.
Indeed, many primary data on Hadza food-sharing show that pro-
ducers' households do receive at least some more food than do other
households in at least some circumstances. For example, Hawkes
et al. (2001a) reported that producers' households receive twice as
much as others for quarry weighing >180 kg and found that, when
considering in-camp redistribution only, producers households could
expect 25% of meat brought back to camp versus 15% for the house-
holds of non-producers (See reanalysis by Hawkes et al., 2014). For
this reason, the absence of a clear relationship between hunting repu-
tation and nutritional status both here and elsewhere (Blurton Jones,
2016) is in some ways surprising. However, despite the potential
sources of noise in the study measures, if the households of the very
best hunters were indeed receiving 4.2 times more male-produced
calories than those of the worst, as calculated by Wood and Marlowe
(2013), we might at least expect to see some signal of this pattern in
the present data. Instead, current results add to a substantial body of
research that is unable to identify any clear nutritional benefit of
being or marrying a successful hunter.
Evidence of widespread and apparently costly food redistribution,
in some previous studies (Hawkes, 1991; Hawkes et al., 1991, 2014,
2018) been taken as support for the idea that hunting might act as a
costly signal of qualities related to hunting prowess (Hawkes et al.,
2018; Hawkes & Bird, 2002). However, demonstrations of apparently
realized costs are not diagnostic of costly signaling (see Higham, 2014;
Stibbard-Hawkes, 2019a; Számadó, 2011). Therefore, although the evi-
dence presented here is consistent with costly signaling, current findings
provide no positive evidence that hunting acts as a costly signal. Here, as
elsewhere (Blurton Jones, 2016; Hawkes et al., 2018), evidence for gener-
ous and widespread food-sharing may raise questions about the fitness
benefits of hunting, but it does not answer them.
Current findings are concordant with existing research which
shows that men eat a higher quality diet than women (Berbesque
et al., 2011), partly through self-provisioning outside of camp
(Berbesque et al., 2016). Our study also shows a drop in men's iron
levels when compared to results from 50 years ago. In this period
researchers have anecdotally observed an increase in ethnotourism
(the proceeds from which are used to augment a foraged diet with
grain) and increasing exploitation by farmers and pastoralists of areas
traditionally used by Hadza for hunting. The drop in hemoglobin levels
is small in relative terms and consistent with a diet (especially a male
diet) that still has much meat. However, an investigation of changes in
diet and activity budget brought about by changing subsistence prac-
tices would be timely.
ACKNOWLEDGMENTS
We would like to thank all those who have made this research possi-
ble, including Enrico Crema, Trevor Hawkes, Coren Apicella, Karen
Kempton, Bill Nolan, Theresa Tiffert, Charles Endeko, and Audax
Mabulla. Thanks to Jonathan Kingdon and to the late Colin Groves for
their help in identifying some of the butchered animals observed dur-
ing the study. Finally, thanks to our two anonymous reviewers for their
valuable comments and suggestions which have substantially improved
the manuscript. This research was funded by Robinson College Cam-
bridge, the Leakey Foundation, the Smuts Memorial Fund, the
Ruggles-Gates Fund of the Royal Anthropological Institute, the Cam-
bridge Department of Archeology and Anthropology, the Cambridge
Centre for African Studies, the Anthony Wilkin Fund and the
Ridgeway-Venn Fund.
AUTHOR CONTRIBUTIONS
The study was designed, and data collection conducted, by Duncan
Stibbard Hawkes and Frank Marlowe. Ibrahim Mabulla was also substan-
tially involved in project data collection. Duncan Stibbard Hawkes wrote
the article and conducted analyses with substantial support, input and
feedback from Robert Attenborough. Dr Marlowe retired in 2014 due to
serious illness and passed away in September, 2019. Due to his illness,
he was unable to contribute to the analysis and interpretation of data or
writing the subsequent article. Tributes have been published in the
journals Human Nature (Gray et al., 2019) and Evolution and Human
Behaviour (Apicella, 2019). A summary of his life and work can be
accessed, at the time of publication, at http://frankmarlowearchive.com/.
DATA AVAILABILITY STATEMENT
Certain participants are identifiable by their hunting reputations, espe-
cially when accompanied by their ages. For this reason, to protect par-
ticipant's anonymity, we have not made these data freely available
online.
ORCID
Duncan N. E. Stibbard-Hawkes https://orcid.org/0000-0002-6719-
9507
ENDNOTES
1
http://www.hemocue.com/en/knowledge-center/learning-center/
online-training.
2
It is for this reason that we report body fat % estimates and not the
impedance measure itself. This approach has also been successfully
implemented by Sherry and Marlowe (2007) in a previous study of the
Hadza.
3
For example, compare older peak reputation reported by Koster, 2010,
Blurton Jones, 2016 or Stibbard-Hawkes et al., 2018; Stibbard-Hawkes,
2019b with younger peak foraging efficacy in a large multi-population
sample reported by Koster et al., 2019 using returns measures.
REFERENCES
Apicella, C. L. (2014). Upper-body strength predicts hunting reputation
and reproductive success in Hadza hunter-gatherers. Evolution and
Human Behavior,35, 508–518.
Apicella, C. L. (2019). Frank Wesley Marlowe (1954-2019). Evolution and
Human Behavior,41,1–104.
Apicella, C. L., Crittenden, A. N., & Tobolsky, V. A. (2017). Evolution and
human behavior hunter-gatherer males are more risk-seeking than
females, even in late childhood. Evolution and Human Behavior,38,
592–603.
STIBBARD-HAWKES ET AL.17

Bailey, K. V., & Ferro-Luzzi, A. (1995). Use of body mass index of adults in
assessing individual and community nutritional status. Bulletin of the
World Health Organization,73, 673–678.
Bender, D. A. (2014). Nutrition metabolism (5th ed.). Boca Raton: CRC
Press.
Bennett, F. J., Barnicot, N. A., Kagan, I., & Woodburn, J. (1970). Helminth
and protozoal parasites of the Hadza of Tanzania. Transactions of the
Royal Society of Tropical Medicine and Hygiene,64, 857–880.
Bentley, G. R. (1985). Hunter-gatherer energetics and fertility: A
reassessment of the !Kung San. Human Ecology,13,79–109.
Berbesque, J., & Marlowe, F. W. (2009). Sex differences in food prefer-
ences of Hadza hunter-gatherers. Evolutionary Psychology,7, 601–616.
Berbesque, J. C., Marlowe, F. W., & Crittenden, A. N. (2011). Sex differ-
ences in Hadza eating frequency by food type. American Journal of
Human Biology,23, 339–345.
Berbesque, J. C., Marlowe, F. W., Pawn, I., Thompson, P., Johnson, G., &
Mabulla, A. (2012). Sex differences in Hadza dental wear patterns: A
preliminary report. Human Nature,23, 270–282.
Berbesque, J. C., Wood, B. M., Crittenden, A. N., Mabulla, A., &
Marlowe, F. W. (2016). Eat first, share later: Hadza hunter-gatherer
men consume more while foraging than in central places. Evolution and
Human Behavior,37,1–6.
Bird, R. (1999). Cooperation and conflict: The behavioral ecology of the
sexual division of labor. Evolutionary Anthropology,8,65–75.
Bliege Bird, R., & Bird, D. W. (2008). Why women hunt: Risk and contem-
porary foraging in a Western Desert Aboriginal community. Current
Anthropology,49, 655–693.
Blurton Jones, N. G. (2016). Demography and evolutionary ecology of Hadza
hunter-gatherers. Cambridge: Cambridge University Press.
Blurton Jones, N. G., & Marlowe, F. W. (2002). Selection for delayed matu-
rity. Human Nature,13, 199–238.
Bosy-Westphal, A., Later, W., Hitze, B., Sato, T., Kossel, E., Gluer, C. C., …
Muller, M. J. (2008). Accuracy of bioelectrical impedance consumer
devices for measurement of body composition in comparison to whole
body magnetic resonance imaging and dual X-ray absorptiometry.
Obesity Facts,1, 319–324.
Buchholz, A. C., Bartok, C., & Schoeller, D. A. (2004). The validity of bio-
electrical impedance models in clinical populations. Nutrition in Clinical
Practice,19, 433–446.
Clark, S. F. (2008). Iron deficiency anemia. Nutrition in Clinical Practice,23,
128–141.
Codding, B. F., Bird, R. B., & Bird, D. W. (2011). Provisioning offspring and
others: Risk-energy trade-offs and gender differences in hunter-
gatherer foraging strategies. Proceedings of the Royal Society B: Biologi-
cal Sciences,278, 2502–2509.
Crittenden, A. N., Conklin-Brittain, N. L., Zes, D. A., Schoeninger, M. J., &
Marlowe, F. W. (2013). Juvenile foraging among the Hadza: Implica-
tions for human life history. Evolution and Human Behavior,34,
299–304.
Deurenberg, P., Deurenberg-Yap, M., & Guricci, S. (2002). Asians are dif-
ferent from Caucasians and from each other in their body mass inde-
x/body fat per cent relationship. Obesity Reviews,3, 141–146.
Fitzpatrick, K. (2018) Foraging and Menstruation in the Hazda of Tanzania.
PhD thesis, University of Cambridge.
Flegal, K. M., Shepherd, J. A., Looker, A. C., Graubard, B. I., Borrud, L. G.,
Ogden, C. L., …Schenker, N. (2009). Comparisons of percentage body
fat, body mass index, waist circumference, and waist-stature ratio in
adults. The American Journal of Clinical Nutrition,89, 500–508.
Froehle, A. W., Wells, G. K., Pollom, T. R., Mabulla, A. Z., Lew-Levy, S., &
Crittenden, A. N. (2019). Physical activity and time budgets of Hadza for-
ager children: Implications for self-provisioning and the ontogeny of the
sexual division of labor. American Journal of Human Biology,31,1–13.
Gray, P. B., Crittenden, A. N., Apicella, C. L., Berbesque, C., Stibbard-
Hawkes, D. N. E., & Wood, B. (2019). In Memoriam: Frank
W. Marlowe (April 17, 1954–September 25, 2019). Human
Nature. https://doi.org/10.1007/s12110-019-09357-1.
Gurven, M. (2004). To give and to give not: The behavioral ecology of
human food transfers. Behavioral and Brain Sciences,27, 543–583.
Gurven, M., & Hill, K. (2009). Why do men hunt? Current Anthropology,50,
51–74.
Hawkes, K. (1991). Showing off: Tests of an hypothesis about men's forag-
ing goals. Ethology and Sociobiology,12,29–54.
Hawkes, K., & Bird, R. B. (2002). Showing off, handicap signaling, and the
evolution of Men's work. Evolutionary Anthropology,11,58–67.
Hawkes, K., O'Connell, J., & Blurton Jones, N. (2018). Hunter-gatherer
studies and human evolution: A very selective review. American Jour-
nal of Physical Anthropology,165, 777–800.
Hawkes, K., O'Connell, J. F., & Blurton Jones, N. G. (1991). Hunting income
patterns among the Hadza: Big game, common goods, foraging goals
and the evolution of the human diet. Philosophical Transactions of the
Royal Society, B: Biological Sciences,334, 243–250 discussion
250–251.
Hawkes, K., O'Connell, J. F., & Blurton Jones, N. G. (2001a). Hadza meat
sharing. Evolution and Human Behavior,22, 113–142.
Hawkes, K., O'Connell, J. F., & Blurton Jones, N. G. (2001b). Hunting and
nuclear families: Some lessons from the Hadza about Men's work. Cur-
rent Anthropology,42, 681–709.
Hawkes, K., O'Connell, J. F., & Blurton Jones, N. G. (2014). More lessons
from the Hadza about Men's work. Human Nature,25, 596–619.
Hawkes, K., O'Connell, J. F., & Coxworth, J. E. (2010). Family provisioning
is not the only reason men hunt. Current Anthropology,51, 259–264.
Hiernaux, J., & Hartono, D. B. (1980). Physical measurements of the adult
Hadza of Tanzania. Annals of Human Biology,7, 339–346.
Higham, J. P. (2014). How does honest costly signaling work? Behavioral
Ecology,25,8–11.
Hill, K., & Hurtado, A. M. (1996). Ache life history: The ecology and demogra-
phy of a foraging people. Hawthorne: Aldine de Gruyter.
Hill, K., & Kintigh, K. (2009). Can anthropologists distinguish good and
poor hunters? Implications for hunting hypotheses, sharing conven-
tions, and cultural transmission. Current Anthropology,50, 369–378.
Howell, N. (1979). Demography of the Dobe !Kung (2nd ed.). New York:
Hawthorne.
Jaggernath, M., Naicker, R., Madurai, S., Brockman, M. A., Ndung'u, T., &
Gelderblom, H. C. (2016). Diagnostic accuracy of the HemoCue Hb
301, STAT-site MHgband URIT-12 point-of-care hemoglobin meters
in a central laboratory and a community based clinic in Durban,
South Africa. PLoS One,11,1–11.
Karastergiou, K., Smith, S. R., Greenberg, A. S., & Fried, S. K. (2012). Sex
differences in human adipose tissues—The biology of pear shape. Biol-
ogy of Sex Differences,3,1–12.
Kelly, R. L. (2013). The Lifeways of hunter-gatherers: The foraging Spectrum
(3rd ed.). Cambridge: Cambridge University Press.
Koster, J. (2010). Informant rankings via consensus analysis. Current
Anthropology,51, 257–258.
Koster, J., McElreath, R., Hill, K., Yu, D., Shepard, G., van Vliet, N., …
Ross, C. (2019). The life history of human foraging: Cross-cultural and
individual variation. bioRxiv, 574483 https://www.biorxiv.org/
content/10.1101/574483v1.
Kyle, U. G., Bosaeus, I., De Lorenzo, A. D., Deurenberg, P., Elia, M.,
Gomez, J. M., …Pichard, C. (2004). Bioelectrical impedance analysis—
Part I: Review of principles and methods. Clinical Nutrition,23,
1226–1243.
Lardi, A. M., Hirst, C., Mortimer, A. J., & McCollum, C. N. (1998). Evaluation
of the HemoCue for measuring intra-operative haemoglobin concen-
trations: A comparison with the coulter max-M. Anaesthesia,53,
349–352.
Marlowe, F. W. (2000a). Paternal investment and the human mating sys-
tem. Behavioural Processes,51,45–61.
18 STIBBARD-HAWKES ET AL.

Marlowe, F. W. (2000b). The patriarch hypothesis. Human Nature,11,
27–42.
Marlowe, F. W. (2003). A critical period for provisioning by Hadza men
implications for pair bonding. Evolution and Human Behavior,24,
217–229.
Marlowe, F. W. (2004). Marital residence among foragers. Current Anthro-
pology,45, 277–283.
Marlowe, F. W. (2007). Hunting and gathering: The human sexual division
of foraging labor. Cross-Cultural Research,41, 170–195.
Marlowe, F. W. (2010). The Hadza: Hunter-gatherers of Tanzania. Los
Angeles: University of California Press.
Mascie-Taylor, C. G. N., & Goto, R. (2007). Human variation and body
mass index: A review of the universality of BMI cut-offs, gender and
urban-rural differences, and secular changes. Journal of Physiological
Anthropology,26, 109–112.
Morris, L. D., Osei-Bimpong, A., McKeown, D., Roper, D., & Lewis, S. M.
(2007). Evaluation of the utility of the HemoCue
301 haemoglobinometer for blood donor screening. Vox Sanguinis,93,
64–69.
Pantopoulos, K., Porwal, S. K., Tartakoff, A., & Devireddy, L. (2012). Mech-
anisms of mammalian iron homeostasis. Biochemistry,51, 5705–5724.
Peterson, J. T., Repovich, W. E., Eash, M., Notrica, D., & Hill, C. R. (2007).
Accuracy of consumer grade bioelectrical impedance analysis devices
compared to air displacement Plethysmography. Medicine & Science in
Sports & Exercise,39, S373.
Pontzer, H., Raichlen, D. A., Wood, B. M., Emery Thompson, M.,
Racette, S. B., Mabulla, A. Z., & Marlowe, F. W. (2015). Energy expen-
diture and activity among Hadza hunter-gatherers. American Journal of
Human Biology,27, 628–637.
Pontzer, H., Raichlen, D. A., Wood, B. M., Mabulla, A. Z., Racette, S. B., &
Marlowe, F. W. (2012). Hunter-gatherer energetics and human obe-
sity. PLoS One,7,1–8.
Prentice, A. M., & Jebb, S. A. (2001). Beyond body mass index. Obesity
Reviews,2, 141–147.
Sherry, D. S., & Marlowe, F. W. (2007). Anthropometric data indicate nutri-
tional homogeneity in Hadza foragers of Tanzania. American Journal of
Human Biology,118, 107–118.
Smith, D., Dyble, M., Major, K., Page, A. E., Chaudhary, N., Salali, G. D., …
Mace, R. (2019). A friend in need is a friend indeed: Need-based shar-
ing, rather than cooperative assortment, predicts experimental
resource transfers among Agta hunter-gatherers. Evolution and Human
Behavior,40,82–89.
Smith, K. M., Larroucau, T., Mabulla, I. A., & Apicella, C. L. (2018). Hunter-
gatherers maintain assortativity in cooperation despite high levels of
residential change and mixing. Current Biology,28, 3152–3157.e4.
Stevens, J. (2003). Ethnic-specific revisions of body mass index cutoffs to
define overweight and obesity in Asians are not warranted.
International Journal of Obesity and Related Metabolic Disorders,27,
1297–1299.
Stibbard-Hawkes, D. N. E., Attenborough, R. D., & Marlowe, F. W. (2018).
A Noisy signal: To what extent are Hadza hunting reputations predic-
tive of actual hunting skills? Evolution and Human Behavior,39, 639.
Stibbard-Hawkes, D. N. (2019a). Costly signaling and the handicap princi-
ple in hunter-gatherer research: A critical review. Evolutionary Anthro-
pology,28, 144–157.
Stibbard-Hawkes, D. N. (2019b). No association between 2D:4D ratio and
hunting success among Hadza hunters. Human Nature,1–21.
Strassmann, B. I. (1999). Menstrual cycling and breast cancer: an evolu-
tionary perspective. Journal of women's health,8, 193–202.
Sullivan, K. M., Mei, Z., Grummer-Strawn, L., & Parvanta, I. (2008).
Haemoglobin adjustments to define anaemia. Tropical Medicine and
International Health,13, 1267–1271.
Számadó, S. (2011). The cost of honesty and the fallacy of the handicap
principle. Animal Behaviour,81,3–10.
Wang, L., & Hui, S. S.-C. (2015). Validity of four commercial bioelectrical
impedance scales in measuring body fat among Chinese children and
adolescents. BioMed Research International,2015,1–8.
WHO (2005). Worldwide prevalence of anaemia 1993–2005 [Technical
report]. Geneva: WHO.
WHO (2006). WHO Global Database on Anaemia: United Republic of Tanza-
nia [Technical report]. Geneva: World Health Organization.
WHO (2011). Haemoglobin concentrations for the diagnosis of Anaemia and
assessment of severity. Geneva: World Health Organization.
Wood, B. M., & Marlowe, F. W. (2013). Household and kin provisioning by
hadza men. Human Nature,24, 280–317.
Wood, B. M., & Marlowe, F. W. (2014). Toward a reality-based under-
standing of Hadza Men's work. Human Nature,25, 620–630.
Woodburn, J. (1964) The social organisation of the Hadza of North Tan-
ganyika [Doctoral thesis]. University of Cambridge, Cambridge, UK.
SUPPORTING INFORMATION
Additional supporting information may be found online in the
Supporting Information section at the end of this article.
How to cite this article: Stibbard-Hawkes DNE,
Attenborough RD, Mabulla IA, Marlowe FW. To the hunter go
the spoils? No evidence of nutritional benefit to being or
marrying a well-reputed Hadza hunter. Am J Phys Anthropol.
2020;1–19. https://doi.org/10.1002/ajpa.24027
STIBBARD-HAWKES ET AL.19