Beehive fences as a multidimensional conflict-mitigation tool for farmers coexisting with elephants

Abstract

Increasing habitat fragmentation and human population growth in Africa has resulted in an escalation in human-elephant conflict between small-scale farmers and free-ranging African elephants (Loxodonta Africana). In 2012 Kenya Wildlife Service (KWS) implemented the national 10-year Conservation and Management Strategy for the Elephant in Kenya, which includes an action aimed at testing whether beehive fences can be used to mitigate human-elephant conflict. From 2012 to 2015, we field-tested the efficacy of beehive fences to protect 10 0.4-ha farms next to Tsavo East National Park from elephants. We hung a series of beehives every 10 m around the boundary of each farm plot. The hives were linked with strong wire. After an initial pilot test with 2 farms, the remaining 8 of 10 beehive fences also contained 2-dimensional dummy hives between real beehives to help reduce the cost of the fence. Each trial plot had a neighboring control plot of the same size within the same farm. Of the 131 beehives deployed 88% were occupied at least once during the 3.5-year trial. Two hundred and fifty-three elephants, predominantly 20-45 years old entered the community farming area, typically during the crop- ripening season. Eighty percent of the elephants that approached the trial farms were kept out of the areas protected by the beehive fences, and elephants that broke a fence were in smaller than average groups. Beehive fences not only kept large groups of elephants from invading the farmland plots but the farmers also benefited socially and financially from the sale of 228 kg of elephant-friendly honey. As news of the success of the trial spread, a further 12 farmers requested to join the project, bringing the number of beehive fence protected farms to 22 and beehives to 297. This demonstrates positive adoption of beehive fences as a community mitigation tool. Understanding the response of elephants to the beehive fences, the seasonality of crop raiding and fence breaking, and the willingness of the community to engage with the mitigation method will help contribute to future management strategies for this high human-elephant conflict hotspot and other similar areas in Kenya.

Full text

Conservation Practice and Policy
Beehive fences as a multidimensional
conflict-mitigation tool for farmers coexisting
with elephants
Lucy E. King,1,2 ∗Fredrick Lala,3Hesron Nzumu,4Emmanuel Mwambingu,4
and Iain Douglas-Hamilton1,2
1Save the Elephants, P.O. Box 54667, Nairobi, Kenya
2Department of Zoology, University of Oxford, Oxford, OX1 3PS, U.K.
3Kenya Wildlife Service, Tsavo East National Park, P.O. Box 14–80300, Voi, Kenya
4Mwakoma Village, Sagalla, Taita-Taveta County, Kenya
Abstract: Increasing habitat fragmentation and human population growth in Africa has resulted in an
escalation in human–elephant conflict between small-scale farmers and free-ranging African elephants (Lox-
odonta Africana). In 2012 Kenya Wildlife Service (KWS) implemented the national 10-year Conservation and
Management Strategy for the Elephant in Kenya, which includes an action aimed at testing whether beehive
fences can be used to mitigate human–elephant conflict. From 2012 to 2015, we field-tested the efficacy of
beehive fences to protect 10 0.4-ha farms next to Tsavo East National Park from elephants. We hung a series of
beehives every 10 m around the boundary of each farm plot. The hives were linked with strong wire. After an
initial pilot test with 2 farms, the remaining 8 of 10 beehive fences also contained 2-dimensional dummy hives
between real beehives to help reduce the cost of the fence. Each trial plot had a neighboring control plot of
the same size within the same farm. Of the 131 beehives deployed 88% were occupied at least once during the
3.5-year trial. Two hundred and fifty-three elephants, predominantly 20–45 years old entered the community
farming area, typically during the crop- ripening season. Eighty percent of the elephants that approached the
trial farms were kept out of the areas protected by the beehive fences, and elephants that broke a fence were
in smaller than average groups. Beehive fences not only kept large groups of elephants from invading the
farmland plots but the farmers also benefited socially and financially from the sale of 228 kg of elephant-
friendly honey. As news of the success of the trial spread, a further 12 farmers requested to join the project,
bringing the number of beehive fence protected farms to 22 and beehives to 297. This demonstrates positive
adoption of beehive fences as a community mitigation tool. Understanding the response of elephants to the
beehive fences, the seasonality of crop raiding and fence breaking, and the willingness of the community
to engage with the mitigation method will help contribute to future management strategies for this high
human–elephant conflict hotspot and other similar areas in Kenya.
Keywords: beehive fences, community, elephant deterrents, elephant management strategy, Elephants and
Bees Project, human–elephant conflict, participatory trials, Tsavo National Park
Cercos de Panales como una Herramienta Multidimensional para la Mitigaci´
on de Conflictos entre Agricultores y
Elefantes
Resumen: El incremento de la fragmentaci´
on del h´
abitat y el crecimiento de la poblaci´
on humana en ´
Africa
han resultado en un aumento del conflicto entre los peque ˜
nos agricultores y los elefantes africanos (Loxodonta
africana) libres. En el 2012, el Servicio de Vida Silvestre de Kenia (KWS, en ingl´
es) implement´
oanivelnacional
la Estrategia de Manejo y Conservaci´
on para el Elefante en Kenia con duraci´
on de 10 a˜
nos, la cual incluye
una acci´
on enfocada en probar si los cercos de panales pueden utilizarse para mitigar el conflicto humano
– elefante. De 2012 a 2015, probamos en el campo la eficiencia de los cercos de panales para proteger de los
∗email lucy@savetheelephants.org
Paper submitted April 28, 2016; revised manuscript accepted November 12, 2016.
1
Conservation Biology, Volume 00, No. 0, 1–10
C
⃝2017 Society for Conservation Biology
DOI: 10.1111/cobi.12898

2Beehive-Fence Elephant Deterrent
elefantes a diez granjas de 0.4 ha colindantes con el Parque Nacional Tsavo del Este. Colgamos una sere de
panales cada 10 m alrededor de los l´
ımites de cada lote agr´
ıcola. Los panales se conectaron con un alambre
fuerte. Despu´
es de una prueba piloto inicial en dos granjas, los ocho permanecientes de los diez cercos con
panales tambi´
en incluyeron panales-se˜
nuelo bidimensionales entre los panales verdaderos para ayudar a
reducir el costo del cerco. Cada lote de prueba tuvo un lote de control vecino del mismo tama˜
no dentro de
la misma granja. De los 131 panales implementados, el 88% fue ocupado por lo menos una vez durante la
prueba de 3.5 a˜
nos. Doscientos cincuenta y tres elefantes, predominantemente entre los 20 – 45 a˜
nos de edad,
entraron a la comunidad agr´
ıcola, com´
unmente durante la temporada de maduraci´
on de las cosechas. El
80 % de los elefantes que se acercaron a las granjas de prueba se mantuvieron fuera de las ´
areas protegidas
por los cercos de panales, y los elefantes que rompieron los cercos estuvieron dentro de grupos m´
as peque˜
nos
al promedio. Los cercos de panales no s´
olo hicieron que grupos grandes de elefantes no invadieran los lotes
agr´
ıcolas, sino los agricultores tambi´
en se beneficiaron socialmente con la venta de 228 kg de miel amigable
con los elefantes. Conforme se inform´
o sobre el ´
exito de la prueba piloto, doce agricultores m´
as pidieron
unirse al proyecto, lo que llev´
oaln
´
umero de granjas protegidas por cercos de panales a 22 y al de los panales
a 297. Esto demuestra la adopci´
on positiva de los cercos de panales como una herramienta comunitaria de
mitigaci´
on. Entender la respuesta de los elefantes a los cercos de panales, la temporalidad de las incursiones
hacia las cosechas y de la ruptura de los cercos, y la disponibilidad de la comunidad por participar en el
m´
etodo de mitigaci´
on ayudar´
a a contribuir con las siguientes estrategias de manejo para este gran punto
caliente del conflicto humano – elefante y en otras ´
areas similares en Kenia.
Palabras Clave: cercos de panales, conflicto humano – elefante, estrategia de manejo de elefantes, impedimentos
para elefantes, Parque Nacional Tsavo, Proyecto Elefantes y Abejas, pruebas de participaci´
on comunitaria
Introduction
Small-scale subsistence farmers in Africa are confronted
by unprecedented climatic changes (Rockströmetal.
2009) and face many complex challenges for survival,
including in some places an increase in resource compe-
tition with wildlife (Hill 2004). Because human–wildlife
conflict can negatively affect species survival (Cardillo
et al. 2005), damage traditional cultural tolerance for co-
existence with wildlife (Kissui 2008; Dickman 2010), and
often leads to strained relationships with wildlife man-
agers (Woodroffe et al. 2005; Baruch-Mordo et al. 2009),
human–wildlife conflict remains one of the greatest un-
resolved challenges for conservation. People in commu-
nities affected by wildlife can have extremely negative
attitudes toward those species and may retaliate by killing
wildlife, aiding poachers, or blocking tourist activities
(Sifuna 2005; Mackenzie & Ahabyona 2012; Benjaminsen
et al. 2013).
Conflict between farmers and free-ranging elephants
(Loxodona africana)iscommoninKenya(Sitati&
Walpole 2006; Graham & Ochieng 2008), and finding
practical tools to help mitigate this conflict and to create
avenues for productive coexistence remains a substantial
challenge (Okello & Amour 2008; Hoare 2012). In Kenya
both conservationists and local communities increasingly
believe there is no one effective elephant deterrent ap-
propriate for use in all circumstances. The consensus is
that multiple mitigation methods are needed for use by
wildlife managers and farmers for repeated and some-
times rotational use (Osborn & Parker 2002; Dublin &
Hoare 2004; Hoare 2012).
Existing options for human–elephant conflict (HEC)
mitigation include macro methods, ranging from the cre-
ation of wildlife corridors and national parks, and micro
farm-level methods, often used by families tackling one-
on-one conflict situations (Dublin & Hoare 2004). Tried
and tested methods include construction of passive (im-
movable) barriers, such as flashing solar lights, fences
strung with chili-oil soaked rags, ditches, watch tow-
ers, walls, buffer crops (e.g., chilies), and active (move-
able) deterrents such as fire or firecrackers, chili pow-
der bombs, guard dogs, and human patrols (Thouless &
Sakwa 1995; Osborn & Parker 2002; Graham & Ochieng
2008; Hoare 2012).
Common criticisms of established micro-level HEC mit-
igation methods are high establishment cost, elephant ha-
bituation, farmer fatigue, lack of funds for maintenance,
and poor uptake by the wider community even when
pilot studies prove successful (Sitati & Walpole 2006;
Graham & Ochieng 2008). Despite these challenges, mit-
igation can be successful when multiple methods are
used together or in random rotation to reduce habituation
(Sitati & Walpole 2006).
An emerging micromitigation method that combines
passive and active deterrent characteristics is the deploy-
ment of beehive fences around the outer boundary of
small-scale farms (King 2010; King et al. 2011). Elephants
are wary of foraging near African honey bees (Apis mellif-
era scutellata) (Vollrath & Douglas-Hamilton 2002) and
will run away from either the sound of (King et al. 2007)
or a threat of being stung by a swarm of honey bees
(King 2010). Elephants retreating from the threat of bees
emit a unique alarm call to nearby family members (King
Conservation Biology
Volume 00, No. 0, 2017

King et al. 3
et al. 2010). A 2-year field trial with 34 farms in Kenya
revealed that beehive fences reduced crop-raiding inci-
dents and farmers unanimously thought the fences were
successful due to the additional benefits derived from
the bee colonies pollinating crops and producing honey
they could sell (King 2010; King et al. 2011). Although
the positive response from the farmers was encouraging,
the effect of the open-ended fence design was not fully
explored (Hoare 2012) and the effect of the severe 2009
drought limited both hive occupation and crop-raiding
incidents (King et al. 2011).
In 2012, Kenya Wildlife Service (KWS) launched
a collaborative 10-year Conservation and Management
Strategy for the Elephant in Kenya. One of its 7 key
aims was to “[e]nhance human-elephant conflict mit-
igation by involving stakeholders at all levels in the
use of appropriate site-specific methods” (Litoroh et al.
2012). In cooperation with Save the Elephants, KWS,
and a community of subsistence farmers in Sagalla co-
existing with elephants next to Tsavo East National
Park, we tested the efficacy of a beehive fence de-
sign to keep elephants away from invading small-scale
farms.
Methods
Study Area
Sagalla is a rural farming community in Voi subcounty,
southern Kenya. It has 4 villages nestled at the base
of the steep slopes of Mount Sagalla and is approxi-
mately 3 km from the boundary of Tsavo East National
Park, which is home to approximately 6214 elephants
(Ngene et al. 2013). The land between Mt. Sagalla and
Mt. Kasigau to the south is a corridor where elephants
migrate out of Tsavo East into Taita-Taveta County and is
composed of a complex mosaic of community ranches
and wildlife-friendly sanctuaries connected on the west-
ern boundary by Tsavo West National Park. The 4 vil-
lages along the eastern slope of Mt. Sagalla (Kirumbi,
Mwakoma, Mwambiti, and Kajire) are all subject to
human–wildlife conflict. Mwakoma, the smallest of these
communities, has 150 households and is situated clos-
est to the park. Elephants primarily disturb front-line
farms (i.e., those on the eastern side of the road that
are first approached by elephants leaving Tsavo East
National Park) at night as they disperse from the park
in search of water and sustenance. The elephants often
break into the farms to eat crops or occasionally stored
grains, which they access by pushing down house walls
or pulling off roofs. Rangers are frequently called out
to these communities to drive elephants off and several
human fatalities have occurred, which has strained re-
lations between the community and KWS (Sakellariadis
2015).
Farm Selection
Mwakoma village is divided in half by a north-south access
road. Two-thirds of the farms are to the west of the road
under the steep slopes of the mountain. In May 2012, we
used ArcGis (version 10.1) to map the boundaries of all
52 front-line Mwakoma farms on the eastern side of the
road (see map in Supporting Information). We did not
deploy beehive fences on 17 of these farms because they
were not planted regularly and did not have permanent
homes. Of the remaining 35 farms, 2 already had beehive
fences from a successful 2009 pilot study (farms K and
M) (see King [2010] for pilot-study details), 1 was already
engaged in an HEC-reduction project (planting aloe vera,
farm 4), and 11 farms were along the main road and
were thus relatively protected due to human activity and
their greater distance from the park. This left 21 perma-
nent front-line farms as good candidates for beehive-fence
deployment.
During a village meeting, 64 participants identified the
10 farms most affected by elephant raids and owned
by active farmers whom they believed would benefit
most from a beehive fence. We used a semistructured
questionnaire to interview owners of these 10 farms to
gather baseline data on their socioeconomic status and
historical information on elephant incidents (Sakellari-
adis 2015). We selected 8 farms for beehive fences based
on socioeconomic conditions and intensity of past crop-
raiding events. To this group, we added 2 farms from the
pilot study that still had operational beehive fences.
In a change from the pilot-study fences (King 2010) and
to reduce costs, every other beehive was replaced with
a 2-dimensional plywood dummy beehive; thus, fences
had 12 real beehives and 12 dummy beehives interlinked
around each 0.4-ha plot (Fig. 1) (King 2014). The real
beehives and dummy beehives were alternatively hung
1.5 m off the ground between live fence posts cut from
locally available species of Commiphora, a tree that re-
grows once planted. These live posts reduced the effort
and cost of replacing termite-weakened posts and even-
tually will grow into trees, providing natural shade and
food for the bees and an additional physical barrier to
elephants. Each hive and dummy were interlinked with
plain fencing wire (Fig. 1). Elephants attempting to pass
between the hives would set multiple beehives swinging.
The swinging caused the bees to fly out and chase away
the elephants (King et al. 2007; King 2010).
We built 6 new beehive fences during the dry season
(June –August) of 2012 and 2 more in February 2013.
Two of the new beehive fences incorporated 12 Kenyan
Top Bar Hives (KTBH) with 18 frames each that were con-
structed by community carpenters at a cost of $35/bee-
hive (all monetary units are in U.S. dollars unless other-
wise noted). Langstroth hives were sourced from Honey
Care Africa for $60/hive and were adapted with wooden
blocks nailed to the side of the hive to facilitate hanging
Conservation Biology
Volume 00, No. 0, 2017

4Beehive-Fence Elephant Deterrent
Figure 1. Construction of Langstroth beehives and Kenyan Top Bar Hives (KTBH) used in elephant-deterrent
beehive fences and setup and components of the fences. The waxed combs of the Langstroth hives help attract bees
and provide ample storage for honey in the super box above, separated from the brood chamber by a horizontal
queen excluder. The KTBH hives have a vertical queen excluder separating the front and back chambers for brood
and honey respectively. Real hives and dummy hives (2-dimensional plywood) hang alternately along plain
fencing wire attached to regrowing Commiphora sp. posts. Iron sheets tied onto the posts help deter honey badger
attacks.
(Fig. 1). The Langstroth hives consisted of a large lower
brood box of 10 large frames and a secondary upper super
box of 10 small frames for honey collection. From each of
the small frames a bar with approximately 1 kg of honey
could be harvested once the colony was settled. Both
KTBH and Langstroth hives had a mesh queen excluder to
separate brood from honey (Fig. 1). Materials for a 240-m
beehive fence constructed with 12 Langstroth beehives
and 12 dummy hives cost on average $850. A 12 KTBH
and 12 dummy beehive fence cost $550. The 10 trial
beehive fences around 10, 0.4-ha plots incorporated 68
Langstroth hives (6 fences) and 63 KTBH hives (4 fences).
Each of the 10 farm plots protected by a beehive fence
was paired with a control plot of the same size within
the same farm.
We trained each participating farmer in beekeeping
and to record both hive occupation events and elephant
movements around the beehive fences. We trained one
community member to help the farmers with honey har-
vesting and major fence maintenance and to track ele-
phant movements and crop raiding with a GPS. Dung and
urine markings were observed to help identify whether
elephants were male or female, and the length and width
of each clear hind footprint were measured to help es-
tablish age (Lee & Moss 1997). We could not accurately
measure night-guarding efforts by the farmers or compare
our data with any quantitative crop-raiding data prior to
the trail because of a lack of detailed elephant-incident
records.
Significant Events during the Trial
Eighteen months into the trial we constructed a commu-
nity office and honey processing room that enhanced our
Conservation Biology
Volume 00, No. 0, 2017

King et al. 5
Figure 2. The (a) number of elephants in the community over the project period (after December 2013
monitoring effort and railway construction activity increased); (b) bee occupation of Kenyan Top Bar Hives
(KTBH) (bars) and their occupation relative to rainfall (line); and (c) bee occupation of Langstroth hives.
data-collection capacity and enabled us to process honey
and beeswax on site. In December 2013, construction
began on the new high-speed standard gauge railway be-
tween Mombasa and Nairobi. Railway construction along
a7-m-highridgealongtheboundaryofTsavoEast,gen-
erated substantial mechanical disturbance along the park
boundary. Concurrently, the number of elephant crop-
raiding incidents increased 500% from 39 elephant visits
from June 2012 to December 2013 to 199 elephant visits
during the second half of our study period, (June 2014 to
December 2015) (Fig. 2).
As the trial developed, demand for beehive fences
increased from Mwakoma’s unfenced farmers. Conse-
quently, with 30 beehives donated from the Tsavo KWS
office, a further 3 Mwakoma beehive fences were added
in mid-2014 (Farms HN, KR and S) (Supporting Informa-
tion). Demand from neighboring Mwambiti community
resulted in us establishing 9 additional beehive fences in
2015. Hence, at the end of the 43-month trial in Decem-
ber 2015, the project had expanded to 22 farms and had
a total of 297 beehives. We present data from only the
first 10 farms with beehive fences, but we considered the
likely reasons for this progressive demand for beehives
by the wider community.
Results
During the 43-month trial all 10 of the beehive fences
became naturally occupied with wild African honey
bees. Out of the 131 beehives, 116 were occupied at
least once; 98% of the Langstroth hives and 78% of the
KTBH hives were occupied at least once during the
trial. Occupation rates increased as rainfall increased
Conservation Biology
Volume 00, No. 0, 2017

6Beehive-Fence Elephant Deterrent
with the onset of the biannual rainy seasons (November
-December and March-April each year), which triggered
both crop and natural vegetation growth that provided
ample water and foraging sources to attract wild bee
swarms to the beehive fences (Fig. 2).
A total of 253 elephant visits were recorded within
<30 m of the 10 farms protected by beehive fences.
Each elephant visit was recorded as a unique data point.
Of the 238 elephants that directly approached the farms
with beehive fences 80% (n=190) were deterred from
entering the area protected by the fence and significantly
fewer (20%, n=48) broke through the beehive fences
(mean deterrent events per farm =19 [SE 7.9], mean
breakthroughs per farm =4.8 [SE 1.3]; 1-tailed paired t
test, p=0.046). Of the remaining subsample of 15 ele-
phants, 11 elephants walked through a gap in the fence
without touching any part of it (usually an opening in
the fence near the house), and 4 elephants walked along
a portion of the beehive fence to enter a gap without
breaking any part of the fence (Fig. 3).
Group sizes of elephants entering the community var-
ied from single bulls to groups of 10; the median group
size was 3. Elephants were most commonly (73%) in small
group of 1–3. The remaining 27% came in large groups of
4–10. Compared with the typical size of elephant groups
visiting the community, elephants managing to break into
farms through a beehive fence did so in proportionally
smaller group sizes. Thirteen groups of 1–3 (88%) broke
the fence (median group size of 2), and 3 large groups of
4 elephants (12%) broke through a beehive fence across
the 10 farms (Fig. 4). Camera trap footage and 238 foot-
print measurements collected from 59 elephants showed
it was likely the majority of the elephants were bulls, but
females were also observed raiding crops. Sixty-eight per-
cent of the crop-raiding elephants entering the farming
community were estimated to be 20–45 years old.
Out of the 89 total events of elephant groups approach-
ing the farms, there were 65 events where elephants went
within 10 m of the fences. Of these 65 events, 39 elephant
groups (n=114 elephants) turned away at the beehive
fence line and 26 groups (n=50 elephants) broke the
beehive fences. Two elephants were deterred by the ac-
tion of breaking the fence, and 48 elephants broke the
fence and gained access to the farm. Fences with low
hive occupancy at the contact points (0-25% of hives
occupied) showed both high breakthrough events (n=
11 groups) and high deterrent events (n=15 groups),
which suggests that even without bees occupying the
hives some elephant groups were deterred by the fence
structure itself. Where occupation of hives was high (60–
100% of the hives occupied), only 7 elephant groups
broke through over the 43 months, and 14 groups were
kept out by the fence (Supporting Information).
Rainfall influenced when elephants entered the com-
munity: 70% of elephants entered the community and
broke the fences in the driest period (!30 mm rainfall
in the 28 days prior to their visit) (Fig. 5). During times
of heavy rainfall (>91 mm over the previous month),
6 elephants entered the community and none broke a
beehive fence, which indicates a reduced need for crop
species as a nutritional resource during and immediately
after high rainfall.
During the last 2 years of the project (January 2014
to December 2015), the 6 farmers with Langstroth bee-
hive fences had a mean monthly occupancy rate in their
68 hives of 68% and produced a total of 255 bars of
honey (mean 42.6 [SD 21.7]), which yielded 206 kg of
raw honey, significantly more honey than the 4 farmers
using KTBH hives. These farmers experienced a lower
mean hive occupancy rate of 32% and all had poor honey
harvests, producing 23.5 bars of honey (mean 5.9 [SD
5.75]; ttest, p<0.01) among them that yielded 22 kg
of raw honey. The 6 Langstroth farmers made a total
income of $627, whereas the 4 KTBH farmers made $57
(Supporting Information). For subsistence farmer W, who
was living off his pension of $22 per month, his honey
income for 2015 represented a 21.5% increase in his cash-
based income for the year.
The 228 kg of honey produced by all 10 farmers was
bottled and produced 480 jars of honey in 450-g glass
jars, which cost $0.48 each. Lables cost $0.30 each. Each
jar was sold locally within Kenya for 600 Kenyan shillings
($7) each. Total income was $3388. Once the payments
to farmers ($684) and costs of honey jars ($230), labels
($170), honey extractor ($600), beesuits ($390 for 6),
buckets ($40 for 10), and miscellaneous tools ($140)
had been taken into account (total expenses $2254), the
profit from honey sales was approximately $1134.
Discussion
Our 43-month trial provided ample time to thoroughly
test the beehive fence system as a mitigation tool against
crop-raiding elephants and to observe the acceptance
and adoption of the method by the wider community.
We could not compare our crop-raiding data with crop-
raiding incidents prior to the establishment of the fences,
but the community thought the fences effectively re-
duced crop-raiding. Thus, additional community mem-
bers requested to join the project, and we supported
establishment of 22 beehive fences by the end of the trial.
Although 20% of the elephants that directly ap-
proached the beehive fences broke the fences and ac-
cessed the farms, the farmers accepted that the method
was helping them because the number of elephants that
broke through was typically small (median group was
2), they could scare single or smaller elephant groups
away on their own, and they were being compensated
for damage caused by infrequent fence breaks by honey
sales. There were 7 incidents of elephants succeeding
in breaking the fence where the proportion of occupied
Conservation Biology
Volume 00, No. 0, 2017

King et al. 7
Figure 3. Effectiveness of the 10
beehive fences against 253
elephants over 43 months of the
trial in Mukoma Village.
Figure 4. Results of elephant visits to
farms with beehive fences.
beehives on either side of the break point was high. It is
possible that either these few elephants had no experi-
ence with bees prior to the event or they rushed through
the fence and moved quickly away from the hive into the
farm, where the bees were less likely to fly after them in
the dark.
The income generated by the honey provided an in-
centive to maintain the beehive fences even when it was
not crop-raiding season. Often farmer-managed elephant
fences fall into disrepair after a breakthrough or after
harvest because materials are expensive and elephant
events are not that frequent (King 2010). However, for
the beehive fence farmers, maintenance or breakages
were typically swiftly dealt with so as to avoid honey
losses from ants and honey badgers and to maximize
this alternative income-generating project all year round
(Raja 2015). The honey provided farmers with a financial
incentive (Supporting Information) and a cultural and
social benefit because honey is prized in Sagalla, is often
given as a gift during family events, and is highly valued by
medicine men as a method for protection against illness
and bad spirits (Raja 2015). These cultural and medici-
nal markets for honey in Sagalla provide an alternative
destination for the beehive-fence products. By providing
farmers with a mitigation tool that they felt fully engaged
in all year round, the farmers were able to maximize their
own returns depending on how much effort they put into
maintenance and beekeeping activities.
Conservation Biology
Volume 00, No. 0, 2017

8Beehive-Fence Elephant Deterrent
Figure 5. Frequency of elephants
visiting farms and breaking
through beehive fences relative to
biannual rainfall periods.
The profit from the honey sales was enough to pay
70% of the annual salary of the beehive fence officer,
which suggested that with careful equipment mainte-
nance the beehive-fence project with 131 beehives was
potentially a year away from becoming self-sustaining.
Self-sustainability depends on the existence of a consis-
tent local buyer, which is likely because there are a num-
ber of supermarkets and tourist lodges near Voi Town and
Tsavo East National Park. However, had the Langstroth
farmers been required to fund 100% of the cost of the
materials for their fences themselves it would have taken
the 5 best honey producers an average of 6.8 years at the
2015 rate of honey production to pay back the expense of
their 68 Langstroth hives. This time frame is considerable
but could be reduced by additional sales of value-added
products, such as candles and lip balms, and by higher
profits from higher crop yields. It is likely that start-up ma-
terial costs for Langstroth beehive fences ($850/240 m)
will need to be funded initially by one-off donations or
microloans. Although a considerable upfront cost, this
is significantly lower than the expense of installing a 6-
strand electric fence to keep elephants out, which costs
at least $9000/1000 m (Kioko et al. 2008) and requires
continual maintenance and expensive upkeep.
The onset of the biannual rainy seasons triggered in-
creases in hive occupations, which were followed natu-
rally by the growth of crops. There was typically a period
when the seedlings were growing during the rainy season
when only a few elephants approached the farms. It was
only once the rains slowed or stopped completely and
the bee colonies and crops had had a chance to mature
that the elephants started to enter the community more
frequently looking for ripe crops and water. Elephant
visitation going into the dry seasons coincided with the
maximum hive occupation of bees that were benefiting
from the plethora of natural flora, creating a synchronic-
ity between occupied hives and the highest number of
elephants attempting to raid crops.
Beehive fences constructed with Langstroth beehives
were consistently more successful at attracting bees,
keeping bees after harvest events, and retaining bees
between dry seasons. This was likely because the thicker
wood helped keep the hives cool and the brood box
prevented the queen from being disturbed during honey
harvesting. The fences with KTBH hives had a mean oc-
cupancy rate of only 32% during the last 2 years of the
trial. The farmers struggled to keep their bees during the
dry seasons; honey harvests were low; and the 4 KTBH
farmers made little money from their beehive fences.
Lack of financial benefit to the beehive fences may re-
sult in a reduction of incentive for the KTBH farmers
over time; thus, farmers should invest more initially by
purchasing Langstroth hives. Higher rates of hive occu-
pancy appeared to have a positive impact on elephant
deterrent success relative to low occupancy (Supporting
Information), so skills or equipment that can sustain high
occupancy are worth investing in. Without bees to trigger
the conditioned retreat response (King et al. 2007; King
2010), the likelihood of elephants becoming habituated
to the physical beehive fence is likely to increase the
longer hives in a fence remain empty.
The acceptance and consequent adoption of beehive
fences in the wider community (12 additional farms
joined after the trial) may be the best sign the beehive
fences were perceived as effective by the farmers. Al-
though there were some maintenance problems (new
grass shades, replacement posts needed, or occasional
honey badger attacks), the beehive fences typically gen-
erated more money for the farmer than the minimal main-
tenance costs because materials such as grass thatching
Conservation Biology
Volume 00, No. 0, 2017

King et al. 9
and Commiphora posts could largely be gathered sus-
tainably and for free from the surrounding bush. As with
most elephant-deterrent methods, at $850/0.4 ha initial
start-up costs are prohibitive for subsistence farmers, but
as news of the success of the trial spread, hives were
donated to the farmers from individuals within and out-
side Kenya. The KWS contributed 30 donated beehives
halfway through the trial. Distributing a portion of park
entrance fees to communities bordering national parks
to build beehive fences could be a tangible way for
wildlife departments to share revenues and thus reduce
conflict and enhance community relations with park
management.
Our data suggest beehive fences are a viable tool for
small-scale farmers and wildlife managers to use when
communities are struggling with high levels of HEC. The
beehives are relatively affordable to build and maintain,
and the multidimensional, practical, social, and cultural
benefits of honey production and pollination services
from the bees make it a method willingly being adopted
by rural African farmers.
Acknowledgments
We thank the community and farmers of Mwakoma and
Sagalla in Voi County for their welcome and collabora-
tive participation in trials. The KWS and the Republic
of Kenya granted research permission (permit number
NCST/RCD/12B/013/18). We thank W. Lelukumani and
M. Koskei for mapping help, F. Vollrath and J. Soltis for
advice, and our interns for data compilation and beehive
fence construction. We are grateful for funding from Dis-
ney Conservation Fund, The St. Andrews Prize for the
Environment, The Rufford Foundation, EKCT, Save the
Elephants, and numerous private beehive donors.
Supporting Information
The experimental setup (Appendix S1), the relationship
between number of elephant events and proportion of
beehives (Appendix S2), and the amount of honey har-
vested (Appendix S3) are available online. The authors
are solely responsible for the content and functionality
of these materials. Queries (other than absence of the
material) should be directed to the corresponding author.
Literature Cited
Baruch-Mordo S, Breck SW, Wilson KR, Brocerick J. 2009. A tool box
half full: how social science can help solve human–wildlife conflict.
Human Dimensions of Wildlife 14:219–223.
Benjaminsen TA, Goldman MJ, Minwary MY, Maganga FP. 2013. Wildlife
management in Tanzania: state control, rent seeking and community
resistance. Development and Change 44:1087–1109.
Cardillo M, Mace GM, Jones KE, Bielby J, Bininda-Emonds ORP, Sechrest
W, Orme CDL, Purvis A. 2005. Multiple causes of high extinction
risk in large mammal species. Science 309:1239–41.
Dickman AJ. 2010. Complexities of conflict: the importance of consid-
ering social factors for effectively resolving human–wildlife conflict.
Animal Conservation 13:458–466.
Dublin HT, Hoare RE. 2004. Searching for solutions: the evolution of
an integrated approach to understanding and mitigating human–
elephant conflict in Africa. Human Dimensions of Wildlife 9:271–
278.
Graham MD, Ochieng T. 2008. Uptake and performance of farm-
based measures for reducing crop raiding by elephants Loxodonta
africana among smallholder farms in Laikipia District, Kenya. Oryx
42:76–82.
Hill CM. 2004. Farmers’ Perspectives of conflict at the wildlife–
agriculture boundary: some lessons learned from African
subsistence farmers. Human Dimensions of Wildlife 9:279–
286.
Hoare R. 2012. Lessons from 15 years of human–elephant conflict mit-
igation: management considerations involving biological, physical
and governance issues in Africa. Pachyderm 51:60–74.
King LE. 2010. The interaction between the African elephant and the
African honeybee and its potential application as an elephant deter-
rent. PhD dissertation. University of Oxford, Oxford, United King-
dom.
King LE. 2014. Beehive fence construction manual. Save the Elephants,
Nairobi, Kenya. Available from http://elephantsandbees.com/wp-
content/uploads/2014/05/Beehive-Fence-Construction-Manual-
2014-ss.pdf (accessed Oct 2014).
King LE, Douglas-Hamilton I, Vollrath F. 2007. African elephants run
from the sound of disturbed bees. Current Biology 17:R832–
R833.
King LE, Douglas-Hamilton I, Vollrath F. 2011. Beehive fences as ef-
fective deterrents for crop-raiding elephants: field trials in northern
Kenya. African Journal of Ecology 49:431–439.
King LE, Soltis J, Douglas-Hamilton I, Savage A, Vollrath F. 2010. Bee
threat elicits alarm call in African elephants. PLOS ONE 5(e10346)
DOI: 10.1371/journal.pone.0010346.
Kioko J, Muruthi P, Omondi P, Chyiyo PI. 2008. The performance
of electric fences as elephant barriers in Amboseli, Kenya. South
African Journal of Wildlife Research 38(1):52–58.
Kissui BM. 2008. Livestock predation by lions, leopards, spotted hyenas,
and their vulnerability to retaliatory killing in the Maasai steppe,
Tanzania. Animal Conservation 11:422–432.
Lee PC, Moss C. 1997. Statural growth in known-age African elephants
(Loxodonta Africana). Journal of Zoology 236:29–41.
Litoroh M, Omondi P, Kock R, Amin R. 2012. Conservation and Man-
agement Strategy for the Elephant in Kenya. Kenya Wildlife Service,
Nairobi.
Mackenzie CA, Ahabyona P. 2012. Elephants in the garden: finan-
cial and social costs of crop raiding. Ecological Economics 75:72–
82.
Ngene S, Njumbi S, Nzisa M, Kimitei K, Mukeka J, Muya S, Ihwagi F,
Omondi P. 2013. Status and trends of the elephant population in
the Tsavo–Mkomazi ecosystem. Pachyderm 53:38–50.
Okello MM, D’Amour DE. 2008. Agricultural expansion within Kimana
electric fences and implications for natural resource conservation
around Amboseli National Park. Kenya Journal of Arid Environments
72:2179–2192.
Osborn FV, Parker GE. 2002. Community-based methods to reduce crop
loss to elephants: experiments in the communal lands of Zimbabwe.
Pachyderm 33:22–31.
Raja N. 2015. The ‘beehive fence’ project: a socio-economic, cultural
and environmentally appropriate human-elephant conflict mitiga-
tion strategy. BS thesis. University of Durham, United Kingdom.
Rockstr¨
om J, et al. 2009. A safe operating space for humanity. Nature
461:472–5.
Conservation Biology
Volume 00, No. 0, 2017

10 Beehive-Fence Elephant Deterrent
Sakellariadis A. 2015. Shifting the dynamic equilibrium among percep-
tions, attitudes, and behaviors. MS thesis. Yale School of Forestry
and Environmental Studies, New Haven, Connecticut.
Sifuna N. 2005. Providing compensation for damage caused by wildlife:
a case study of Kenya with particular reference to elephants. Journal
of Social Development in Africa 20:7–39.
Sitati NW, Walpole MJ. 2006. Assessing farm-based measures for miti-
gating human-elephant conflict in Transmara District, Kenya. Oryx
40:279–286.
Thouless CR, Sakwa J. 1995. Shocking elephants: fences and crop
raiders in Laikipia district, Kenya. Biological Conservation 72:99–
107.
Vollrath F, Douglas-Hamilton I. 2002. African bees to control African
elephants. Naturwiss 89:508–511.
Woodroffe R, Lindsey P, Roma˜
nach S, Stein A, ole Ranah SMK.
2005. Livestock predation by endangered African wild dogs (Ly-
caon pictus) in northern Kenya. Biological Conservation 124:225–
234.
Conservation Biology
Volume 00, No. 0, 2017