Advances in Anthropology
2012. Vol.2, No.3, 107-111
Published Online August 2012 in SciRes (http://www.SciRP.org/journal/aa) http://dx.doi.org/10.4236/aa.2012.23012
Copyright © 2012 SciRe s . 107
Indigenous Ecological Knowledge of a Human-Elephant
Interaction in Transmara District, Kenya: Implications for
Research and Management
Noah W. Sitati1*, Hellen Ipara2
1World Wide Fund f o r Nature, Nairobi, Kenya
2Moi University, Eldoret, Kenya
Received Mary 14th, 2012 ; revised June 12th, 2012; accepted June 24th, 2012
Indigenous ecological knowledge (IEK) of the Maasai community in the context of their interaction with
elephants around Masai Mara National Reserve (MMNR), Kenya is explored. Although Maasai commu-
nity land sustains a huge elephant population, it is experiencing increased human-elephant conflict (HEC).
Focus group discussions combined with scientifically collected data were used in assessing the relevance
of IEK to elephant related ecological research. The Maasai narrated their experiences with elephants
which were then formulated into hypotheses and tested scientifically by designing experiments that were
monitored to prove the authenticity of IEK. Respondents had in-depth knowledge of some key ecological
processes. Drunken people were more likely to be attacked by elephants, and elephant movement into ad-
jacent group ranches increased with increasing wildebeest density. Elephants mainly raided ripe or mature
crops while pupils within the elephant range performed poorly in national examinations. Based on this,
there is strong evidence that IEK could be used to design sustainable conservation strategies. It is recom-
mended that understanding of IEK in mitigating HEC and its subsequent integration into HEC decision
support system is necessary in order to resolve conflicts.
Keywords: Indigenous Knowledge; Masai Mara; Human-Elephant Conflict
Introduction
The semi-arid savannah ecosystem in Eastern Africa sup-
ports a large population of the African elephant (Loxodonta
africana africana) and pastoral communities. Pastoralism is
viewed as a compatible form of land-use with elephant conser-
vation (Western, 1994). However, concerns have been raised
over recent introduction of agriculture in the Mara ecosystem
(Sitati et al., 2003) by immigrant farmers, which some Maasais
are adopting. Consequently, human-wildlife conflicts have
ensued and threaten livelihoods and wildlife conservation (Si-
tati et al., 2003; Kamonjo et al., 2007). The conservation of
savannah ecosystems is therefore becoming a priority for con-
servation organisations (WWF, 2007). Understanding the new
challenge of conflict and putting in place the necessary mitiga-
tion measures are crucial for future survival of wildlife and
their habitats.
Among the approaches to elephant conservation, community
involvement has in recent years received widespread accep-
tance, despite recent critiques and project failures (Sitati &
Walpole, 2006; Parker & Osborn, 2006). This approach views
local participation as a prerequisite to sustainable conservation
of natural resources, and its promotion has been accompanied
by a raised awareness and appreciation of IEK (Ashenafi &
Leader-Williams, 2005).
Globally, IEK is not only recognised for its intrinsic value,
but also due to its potential instrumental value to science and
conservation. Local communities have in-depth knowledge
about their immediate environment and IEK has been used
where baseline data are needed to improve ecological research
and to design conservation programmes. Due to scarcity of
scientific data, IEK may offer an alternative source of informa-
tion which can be rooted in long-term ecological studies (Ashe-
nafi & Leader-Williams, 2005). Despite this, indigenous knowl-
edge on human-elephant interaction has neither been applied
nor well-documented i n e le p hant studies.
In this paper, we report results from an assessment elephant
related Maasai IEK with respect to ecological processes and
elephant-environment relationships.
Methods
Transmara lies in the south-west of Kenya on the border with
Tanzania, and encompasses the western portion of MMNR. The
district supports 540 - 820 residents and migratory elephants on
community land (Sitati, 2007), while Narok supports over 2000
elephants (KWS, 2006) some of which move up the escarpment
out of MMNR seasonally and cause conflicts like crop raiding,
property damage, and stopping children from going to school
among others (Sitati et al., 2003). Traditionally, over 25,000
affected Maasais attempt to mitigate the conflict.
Assessment of Indigenous Ecological Knowledge
Twenty four Maasai community members with different
background experiences participated in focus group discussions
and gave historical accounts of the status, distribution and be-
haviour of elephants in the district, and past human-elephant
*Corresponding author.

N. W. SITATI, H. IPARA
relationships. During the discussions, leading questions were
either read to participants or asked indirectly and this elucidated
more questions. Discussions were complimented with other
probing techniques that depended on the responses in order to
gather information that participants could not disclose openly
and also to collect diverse views. Interesting responses were
treated with an echo probe to allow a participant to continue
talking while baiting probe was used to reaffirm what had al-
ready been learned, and to elicit further what participants were
reluctant to discuss.
Group discussions elicited information for testing scientifi-
cally among them an outline of some long term observations of
human-elephant interactions based on experience, and which
derived the following checklist questions: Do elephants have a
high preference for mature maize? Under what circumstances
are people attacked by elephants? Does wildebeest migration
increase the number of elephants on community land? Do pu-
pils located within the elephant range have lower mean scores
in national examinations? These questions formed the hypothe-
ses and were subjected to scientific investigation for confirma-
tion.
Elephant’s Preference for Mature Maize
To test the hypothesis that elephants damaged more mature
and dry maize than young and middle level maize, ten commu-
nity scouts were trained to enumerate crop-raiding incidents to
circumvent the problem of exaggeration of reported conflict by
farmers (Siex & Struhsaker, 1999). Farmers reported elephant
crop raiding incidents to their local enumerator who visited the
farm and recorded details of each incident on a standard form
(Hoare, 1999b) which included: level of maize maturity using
four categories; “young” (maize fields without tassels), “mid-
dle” (tasselled maize with immature cobs), “mature” (ready
green maize), and “dry” (maize getting ready for harvesting).
Data were, first, subjected to Kolmogorov-Smirnov test for
normality, and then the percentage of the different maturity
levels of maize was derived. The chi-square goodness of fit test
was used to test whether the observed proportions for a maize
maturity level differed from hypothesized proportions.
Assessment of Elephant Attacks on People
In order to test the hypothesis that elephant attacks on people
usually occur in the morning or at night, and that drunk people
are more likely to be attacked, focus group participants enu-
merated the names and areas of past cases of elephant attacks
on humans. The information was verified using Kenya Wildlife
Service (KWS) Occurrence Books (OBs) records and District
Compensation Committee documents. Families of victims were
visited and interviewed and information on victims’ sex, tribe
and state (whether drunk or sober), and time collected. A chi-
square goodness of fit test was used to test whether the ob-
served proportions for a categorical variable differ from hy-
pothesized proportions while a logistic regression was used to
analyse the factors that may determine the likelihood of ele-
phant attack on people (Drapper & Smith, 1981).
Monitoring of Elephant-Wildebeest Interaction
Elephant counts and wildebeest densities were used to test
the hypothesis that elephant movement between the reserve and
adjacent community lands increases with increasing wildebeest
density. Focus group participants identified three frequently
used elephant corridors between MMNR and communal lands
for monitoring elephant movements. Community scouts and
KWS rangers were stationed in the corridors and recorded ele-
phant numbers and time of using the corridors daily for 12
months. A monthly count of wildebeests was undertaken in
MMNR using seven permanent road transects to determine the
wildebeest density for 12 months. Wildebeests within a dis-
tance of 200 m on both sides of the road were counted for a
distance of 10 km, and the density calculated as the number of
wildebeests per km2. The Analysis of variance (ANOVA) and
correlations were used for analysis.
Performance of Schools in National Examination
Finally, to test the hypothesis that pupils from elephant range
have lower mean scores than those from outside the range, a
comparison between the mean scores of 96 schools in the pri-
mary school national examination, both in (n = 31) and out (n =
65) of the elephant ranges was made. The schools’ mean scores
for performance in national examinations were collected from
the District Education Office and used to test the hypothesis
that schools within the elephant range have lower mean scores
in national examinations. Nonetheless, since not all the pupils
within the elephant range experienced problems with elephants,
the mean scores of 277 pupils who sat for the national examina-
tion in 1999, both within and outside the elephant range, were
obtained from 18 randomly selected primary schools. Other
details collected included: the number of days absent, distance
covered to school, and whether or not elephants interfered with
the pupils. The mean scores for schools and pupils with respect
to elephant range were compared using ANOVA.
Results
Elephants Preference for Mature and Dry Maize
According to IEK based on group discussions, elephants
have high preference for mature and dry maize, and disappear
from farming areas once maize has been harvested until the
next planting season. Results also revealed that a total of 329
farms were raided by elephants. Data collected on levels of
maize maturity were normally distributed (Kolmogorov-Smir-
nov z = 4.57, p = .000). Since maize was planted in February
and March, elephants did not raid farms with very young maize.
In April all maize was either young or middle, while from May
to September maize was mainly mature or dry. The second
planting season was September and October leading to mature
and dry maize from November to January. Surveys of raided
farms showed that elephants attacked more mature (53.7%, n =
261) and dry maize (37.2%, n = 181) than young (3.3%, n = 16)
or middle (5.8%, n = 28) maize cro ps (2 = 177.565, df = 3 , p = .000,
Figure 1).
Elephant Attacks on People
Group discussions revealed that the majority of the people
attacked by elephants were those who stayed out late in bars
and in homesteads during traditional ceremonies and walked
back home late. People who engaged in early morning activities
also risked being attacked. A total of 56 elephant attack cases
on people were recorded between 1960 and 2000. Out of 53
such cases, 17 people were reported as “drunk”, and 36 were
Copyright © 2012 SciR es . 108

N. W. SITATI, H. IPARA
0
10
20
30
40
50
Young Middle MatureDry
Level of maize maturity
% Farms raided
Figure 1.
Maize maturity level and percentage composition of raided farms by
elephants.
“sober”. The proportion of human deaths to injuries was not
varied (2 = 2.283, df = 1, .131) over time. However, the
Maasai suffered more attacks (2 = 23.11, df = 1, p = .001) than
non-Maasai. More males were attacked than females (2 =
25.83, df = 1, p = .001), and the six females (13%) attacked
were all Maasai. Most elephant attacks on people occurred
early in the morning (0600 - 0900 hr) and late in the evening
(1900 - 2900 hr), while fewer (2 = 14.075, df = 2, p = .001)
cases occurred during mid-day and afternoon (Figure 2).
The logistic model for factors that might have determined the
likelihood of an elephant attack on people produced a goodness
of fit of 73.3% of observed to expected values. Results showed
that of the factors examined including a person’s state, sex,
tribe and time of attack, the state of the person, whether drunk
or sober, was the only factor that determined the likelihood of
attack (Table 1).
Wildebeest Migration and Elephant Movement
The IEK from group discussions revealed that seasonal mi-
gration of wildebeests from Serengeti into the Mara ecosystem
from July drove elephants out of the reserve into adjacent
community lands. A total of 13,059 elephants were recorded
moving between MMNR and community lands between July
1999 and June 2000. The composition of elephant groups using
the corridors differed between months (F11,981 = 14.365, p
= .000). Mixed herds dominated between August and Decem-
ber while more bulls used corridors from January to July. In-
creased elephant movement coincided with the end of the crop
raiding season meaning that crop raiding was mainly by the
resident elephant population. Elephant numbers correlated posi-
tively (p = .001) w i t h wildebeest density (Figure 3).
Schools and Pupil Performance in Relation to
Elephant Range
According to group discussions, school children were af-
fected by the presence of elephants and had to report to school
late and/or leave early to avoid elephants on the way. This af-
fected their performance in national examinations. Five year
averages of the mean scores of 96 schools were obtained for
1995 to 1999. The mean scores were normally distributed (Kol-
mogorov-Smirnov z = .765, p = .601). ANOVA showed there
were differences between scores and the location of schools
within and outside the elephant range (F1,94 = 19.54; p = .001).
There were only three boarding schools which had a high mean
grade (316.39 ± 32.74, n = 3) compared to day schools (298.65
± 59.37, n = 93). However, scores for boarding schools did
Table 1.
Factors determining elephant attacks on humans based on logistic re-
gression.
Variable B SE Wald p
Constant –5.014 2.159 5.389 .0203*
State (drunk or sober) 2.816 1.154 5.959 .0146**
Level of signif ica nce sh ow n w ith * = p < .05, ** = p < .01.
not differ from day schools (F1,94 = .263, p = .609). Finally, a
boarding school within the elephant range had a mean score of
348.66 compared with the two boarding schools outside the
elephant range whose mean scores were 317 and 283 respec-
tively.
Although pupils from the elephant range had low mean grade,
the scores for 277 pupils were normally distributed (Kolmo-
gorov-Smirnov z = .857, p = .454). ANOVA showed that there
was a difference between other factors that may influence per-
formance (mean scores) including distance from home to
school (F3,273 = 10.346; p = .001); tribe (F1,275 = 12.101; p
= .001); length of absenteeism (F3,261 = 7.76; p = .001); and,
school location relative to elephant range (F1,275 = 7.70; p
= .006).
Discussion
Due to past experiences with elephants, the Maasai practise
intensive guarding of crops as they mature. The Maasai believe
that elephants usually send a few animals to investigate the
status of crop maturity and then appear in big numbers when
maize is mature. Study results revealed that the levels of maize
maturity determined the seasonal patterns of crop raiding (Ta-
ble 1). Young and middle level maize farms were raided mainly
when: 1) farms were located along elephant routes; 2) an area
had young, middle and mature maize and/or 3) elephants went
out to inspect the level of maize maturity. In the later case,
maize maturity raiding appears explanatory and conforms to
indigenous knowledge. Elephants come in contact with agri-
cultural land more easily during their natural movement pat-
terns and not necessarily to raid crops (Sukumar, 1989; Hoare,
1997; Osborn, 1998). However, the ability to sense the maturity
level of maize has not been understood. Mature maize is pre-
ferred because of its high nutritive value and higher percentage
of moisture during the dry season than grass and browse
(Osborn, 1998).
According to Sam et al. (1997) crop raiding occurred when
crops were mature during the dry season, and natural forage is
in short supply. Hence mature crops provide an important dry
season food source (Osborn, 1998). In contrast, Hoare (1997)
reports that crop raiding did not necessarily occur when crops
were mature. These contradictory findings suggest that crop
raiding may either be opportunistic, implying a preference for,
rather than reliance upon, crops as a source of food.
Results revealed a strong correspondence between the
Maasai IEK and experience of human-elephant interaction. This
however, does not imply that Maasai knowledge of ecological
interactions is equal to that of scientific ecology. A trained
ecologist may benefit from in-depth dialogue with local people
(Donovan & Puri, 2004). Based on this, participants suggested
that 1) elephants are more likely to attack men than women; 2)
most elephant attacks on people occurred mostly in the evening
or in the morning (Figure 2); and 3) drunk people were more
Copyright © 2012 SciRe s . 109

N. W. SITATI, H. IPARA
Copyright © 2012 SciRe s . 110
Figure 2.
Time and frequency of elephant attacks on peo p l e i n t ransmara distr ict.
0
5
10
15
20
25
30
35
40
45
Jan
Feb
March
Apr
May June
July Aug
Sept
Oct
Nov
Dec
Mean density of elephants
0
200
400
600
800
1000
1200
Density of wildebeest (sq.km)
Elephant
Wildebeest
Figure 3.
Relationship between wildebeest density in Masai Mara national reserve and the mean
(±SE) number of elephants moving into communal areas between January and Decem-
ber 2000.
likely to be attacked by elephants (Sitati et al., 2003). The
scientific data generated confirmed all the three suggestions as
true.
Migration corridors exist in most elephant ranges, and this
influences elephant movement (Soule & Gilpin, 1991; Sam et
al., 1997). In Luangwa Valley, elephants moved from the valley
to the higher slopes of the Muchingas to get Musuku (wild fruit),
or to a pool of medicated water known as Chipatala (hospital of
sick and wounded elephants) (Melland, 1938). Elephants in
Transmara District move up the escarpment to feed on Acacia,
which is not in the northern part of the reserve and to access
forest products and salt licks (Sitati et al., 2003).
Elephant corridors along the escarpment in Transmara influ-
enced elephant movement between the reserve and communal
land especially during the migration of wildebeests. Elephants
from the reserve move up the escarpment every evening and
return to the reserve in the morning (Sitati et al., 2003). Ac-
cording to Laws et al. (1975) and Leader-Williams et al. (1990b),
elephants form large herds in situations where they feel threat-
ened and these aggregation behaviour and their movement
while it is dark possibly serves as a security measure on com-
munity land. However, bulls have been reported to venture into
risky movement (Sukumar, 1991; Thouless & Dyer, 1992) to
access high nutritive food for reproductive purposes. Despite
this, elephants are now losing their fear of humans (Tchamba,
1995).
Although elephants are often cited as interfering with learn-
ing activities (Ngure, 1995; Mwathe et al., 1998), this allega-
tion has never been explored scientifically. Study results how-
ever, showed that pupils and teachers lose many hours as ele-
phants sometimes block all routes between home and school.
Some schools also close earlier than usual, to enable children to
look for safer routes back home, and sometimes adults have to
escort children to school. Such factors could definitely have a
negative impact on pupil performance and this probably ex-
plains why schools outside the elephant range have higher
mean scores than schools within the elephant range. Despite
this, distance from school was the most important factor in
determining pupil performance as pupils who are exhausted
from a long walk cannot concentrate in class, are often absent,
and do not have enough time to finish their assignments. Pupils
with unfinished assignme nt s may also av oi d school.
Elephants were often used as an excuse for absenteeism and
late arrival in schools, which was difficult for teachers to verify.

N. W. SITATI, H. IPARA
According to Mwathe et al. (1998), elephants alone may not be
responsible for pupils’ absenteeism.
Hence the establishment of boarding schools is seen as the
best mitigation measure. Additionally, some Maasais have re-
sorted to having two homesteads. One homestead is located
near a school so that children can attend while the other, mainly
for livestock, is away in elephant areas.
This study has demonstrated that the Maasai community has
detailed knowledge of elephants and their ecology and interac-
tion with people. The relevance of Maasai IEK to elephant
research and management is threefold. First, the congruence
between Maasai and scientific knowledge of elephants suggests
that local experts may be consulted for rapid and reliable eco-
logical assessments since local experts often recognize unique
elephant characters better than external scientists. Secondly,
understanding the cause of HEC shows that there is great po-
tential for combining elephant conservation and mitigating of
conflict and improving local livelihoods. Lastly, local concepts
and understanding of ecological processes may provide valu-
able inputs to ecological research, and locally rooted knowl-
edge based on long-time observations may be crucial, espe-
cially in the absence of a scientific base.
REFERENCES
Ashenafi, Z. T., & Leader-Williams, N. (2005). Indigenous common
property management in the central highlands of Ethiopia. Human
Ecology, 33, 539-563. doi:10.1007/s10745-005-5159-9
Donovan, D. G., & Puri, R. K. (2004). Learning from traditional knowl-
edge of non-timber forest products: Penan Benalui and the autecol-
ogy of Aquilaria in Indonesian Borneo. Ecology a n d S ociety, 9, 3.
Hoare, R. E. (1997). The effects of interactions with humans on ele-
phant populations of the Sebungwe region, Zimbabwe. Ph.D. Thesis,
Harare: University of Zimbabwe.
Hoare, R. E. (1999). A standardised data collection andanalysis proto-
col for human-elephant conflict situations in Africa. Nairobi: IUCN
African Elephant Specialist Group.
Kamonjo, N. W., Sitati, N. W., Adano, W. R., Leader-Williams, N., &
Stephenson, P. J. (2007). Assessing the economic costs and benefits
of maize farming in elephant ranges in Transmara District, Kenya. In
M. Walpole, & M. Linkie (Eds.), Mit igating human-elephant conflict:
Case studies from Africa and Asia (pp. 71-82). Cambridge: FFI Pub-
lication.
KWS (2006). Total dry season aerial counts of elephants in the Mara
ecosystem. Nairobi: KWS Unpublished Report.
Laws, R. M., Parker, I. S. C., & Johnstone, R. C. B. (1975). Elephants
and habitats in North Bunyoro, Uganda. East African Wildlife Jour-
nal, 8, 163-180.
Leader-Williams, N., Albon, S. D., & Berry, P. S. M. (1990b). Illegal
exploitation of black rhinos and elephant populations: Patterns of de-
cline, law enforcement and patrol effort in Luangwa Valley, Zambia.
Journal of Applied Ec ol o gy , 27, 1055-1087.
doi:10.2307/2404395
Melland, F. H. (1938). Elephants in Africa. London: Country L ife Ltd.
Mwathe, K. M., Ruhui, J. M., Macharia, O. K., & Warutere, J. M.
(1998). Naari community fence post-project: Environmental impact
assessment. The social, economic and environmental impacts of an
elephant barrier on a farming community in Eastern Kenya. Unpub-
lished Report. Nairobi: KWS Field Report.
Ngure, N. (1995). People-elephant conflict management in Tsavo,
Kenya. Pachyderm, 19, 20-26.
Osborn, L. (1998). Ecology of crop raiding elephants. Pachyderm, 25,
39-40.
Parker, G. E., & Osborn, F. V. (2006). Investigating the potential for
chilli Capsicum spp. to reduce human-wildlife conflict in Zimbabwe.
Oryx, 40, 343-346. doi:10.1017/S0030605306000822
Sam, M. K., Haizel, C., & Barnes, R. F. W. (1997). Crop raiding by
elephants during the 1996 harvest season in the Red Volta Valley,
Upper East Region, Ghana. WWF Project Report. Gland: WWF.
Siex, K. S., & Struhsaker, T. T. (1999). Colobus monkeys and coconuts:
A study of perceived human-wildlife conflicts. Journal of Applied
Ecology, 36, 1009-1020. doi:10.1046/j.1365-2664.1999.00455.x
Sitati, N. W. (2007). Dry and wet season elephant survey in Transmara.
WWF Unpublished Re p o r t. Gland: WWF.
Sitati, N. W., & Walpole, M. J. (2006). Assessing farm-based measures
for mitigating human-elephant conflict in Transmara District, Kenya.
Oryx, 40, 279-286. doi:10.1017/S0030605306000834
Sitati, N. W., Walpole, M. J., Smith, R. J., & Leader-Williams, N.
(2003). Predicting spatial aspects of human-elephant conflict. Jour-
nal of Applied Ecology, 40, 667-677.
doi:10.1046/j.1365-2664.2003.00828.x
Soule, M. E., & Gilpin, M. E. (1991). The theory of wildlife corridor
capability. In A. Denis, & J. H. Richard (Eds.), Nature conservation
2: The role of corridors. Surrey Beatty & Sons.
Sukumar, R. (1991). The management of large mammals in relation to
male strategies and conflict with people. Biological Conservation, 55,
93-102. doi:10.1016/0006-3207(91)90007-V
Sukumar, R. (1989). The Asian elephant: Ecology and management.
Cambridge: Cambridge University Press.
Tchamba, M. N. (1995). The problem elephants of Kaele: A challenge
for elephant conservation in northern Cameroon. Pachyderm, 1, 26-
33.
Thouless, C. R., & Dyer, A. (1992). Radio-tracking of elephants in
Laikipia District, Kenya. Pachyderm, 15, 34-39.
Western, D. (1994). Elephant and people: Statement by the director of
KWS. Swara, 29, 4-5.
WWF (2007). Species action plan for conservation and management of
African elephants. Gl and : WWF International.
Copyright © 2012 SciRe s . 111