Large mammals of Mkomazi
S. Keith Eltringham, Ian A. Cooksey, William J.B. Dixon, Nigel E. Raine, Chris
J. Sheldrick, Nicholas C. McWilliam & Michael J. Packer
Mkomazi Game Reserve has never been noted for large populations of large mam-
mals but the diversity of species is as great as in any other east African protected
area. The size criterion defining a large mammal must, ultimately, be arbitrary. In
ecological terms, individual size correlates with individual contribution to the
rate at which ecological processes occur. The threshold size criterion for a large
mammal in Mkomazi was set at a weight of 3–5 kg or a height of about 50 cm, and
so would include, for instance, baboon and dikdik. The threshold is ecologically
and logistically justified for it includes individual ‘large mammals’ that are likely
to play significant roles in specific ecological processes and means that the chances
of observing individuals of the smaller species is reasonable across a range of
habitats, ensuring that widely distributed species have a fair chance of being ob-
served throughout their range.
Information on large mammal distribution and abundance patterns, and their
seasonal variations, is important to understanding the significance of large mam-
mals as an ecological driving force in Mkomazi. This information is also vital in
evaluating various management questions, such as the conservation importance of
Mkomazi for the persistence of large mammal species. For example, Mkomazi
has the only recorded gerenuk population in a protected area in Tanzania, as well
as nationally important populations of oryx and lesser kudu (TWCM 1991). Other
management considerations include the tourist potential of Mkomazi for game
viewing, the nature and extent of potential wildlife-human conflicts, possibilities
for sustainable harvesting of bush meat by local communities and the re-introduc-
tion of formerly-present species, such as the black rhino.
Large mammal species in Mkomazi
A full list of species recently and currently present in the reserve is given in the
checklist (Chapter 32). Megaherbivores, those weighing more than 1,000 kg, com-
prise elephant and giraffe. The equids are represented by plains zebra, and the pig
family by warthog and bush-pig. There are no hippopotamus, probably because
486 Mkomazi: ecology, biodiversity and conservation
there is no natural source of permanent water in the reserve except for the Umba
River in the far south-east. The eland and buffalo are the largest artiodactyls but
there are many medium-sized ungulates such as waterbuck, lesser kudu, fringe-
eared oryx, kongoni (Coke’s hartebeest), impala, Grant’s gazelle, gerenuk, bushbuck
and Bohor reedbuck. Smaller large mammal species include bush duiker, klip-
springer, steinbuck and dikdik.
The large carnivores are well represented in Mkomazi and include lion, leop-
ard, cheetah and two hyaenas, the spotted and striped. Smaller species comprise
wild dog (possibly extirpated from the reserve, although a group of 15 was sighted
during a seven-day period in March 1997 near Kisima), black-backed jackal, bat-
eared fox, aardwolf, ratel, serval, small-spotted genet, civet, five species of
mongoose, zorilla, serval, caracal and wild cat. Some of these carnivores are very
rare and populations of some may be threatened with extinction.
Other species have been recorded in the past. The wildebeest was common in
the 1930s and the greater kudu occurred until the mid-1950s. Occasional sightings
of the sable were reported up to the early 1950s. A group of 16 wildebeest was
introduced to the reserve in 1966 but it does not seem to have become established.
The black rhinoceros, which is the subject of a re-introduction programme, was
present until relatively recently.
Most species that might be expected in the reserve from its geographical posi-
tion are present with the exception of the hippopotamus. Information on the status
of the large mammals has been derived from a number of sources. The only previ-
ous study in any detail was that made by Harris (1972) from 1964 to 1967. He
established a series of line transects for estimating numbers supplemented by game
counts around water holes and some aerial censuses. Harris extrapolated his data
to provide estimates of the population size of some species and recorded a grada-
tion in large mammal numbers from high densities in the north-west to low densities
in the south-east. This trend reflects the rainfall pattern (see Chapter 2). Not all
species were zoned in this way. Elephants, for example, were found throughout
the reserve and oryx and zebra tended to be confined to the west-central and cen-
tral areas. The distribution varies between water dependent and water-independent
species and is influenced by vegetation, which is spatially heterogeneous. For sev-
eral of the larger mammal species, seasonal movements between Tsavo West
National Park in Kenya and Mkomazi mean that estimates of population size in
the reserve show marked seasonal variation.
Following Harris’ (1972) work, large mammal counts in Mkomazi were made
from the air by the Kenya Range Management Unit (now the Department of Re-
source Surveys and Remote Sensing). These were sample, not total counts, and
they were primarily carried out to count the elephants in the Tsavo ecosystem but
some other species were included and the results of the most recent count, made
in April 1994 (Inamdar 1994 & 1996), enables a comparison to be made with
conditions some thirty years earlier. Ground surveys of large mammals conducted
during the Mkomazi Ecological Research Programme (MERP) sought to deter-
mine current distributions and, where feasible, to estimate population sizes within
the reserve. In addition, systematic observations made in the west of the reserve
aimed to clarify seasonal variation in distribution and abundance.
The spatial and temporal patterns in distribution and abundance of large mam-
mals in Mkomazi are very variable. Elucidating these patterns necessitates the use
of a variety of survey methods, which would ideally sample all habitats several
times a year to capture seasonal variation.
Ground-based and, to a far lesser extent, aerial surveys of large mammal
populations were used in this study. For ground surveys, constraints of accessi-
bility to different parts of the reserve and of resource availability mean that in
practice the surveys are biased in various ways. A spatial bias exists in that the
north-western part of the reserve is the most intensely surveyed. In addition, the
‘sample’ for the whole reserve is small and is associated almost entirely with the
road network. A temporal bias exists as observations were not necessarily made at
times of day when most species are likely to be active, and the observations were
not seasonally representative, particularly outside the periodic survey sector. Fi-
nally, there are observation biases, caused by the varying visibility in different
vegetation types and at different times of year.
Ground-based surveys provide the majority of our presence records. Almost all
surveys were conducted from vehicles using existing roads in Mkomazi. When an
animal was sighted, the vehicle was stopped and the species, group size, location
and date were recorded. In most cases, the location was read from a global posi-
tioning system (GPS), using the same co-ordinate system and projection parameters
as the 1:50,000 maps of Tanzania. When animals were over 200 m from the vehi-
cle, their distance was estimated by eye and a bearing was taken using a standard
compass. These data were used within the Mkomazi geographical information
system (GIS—see Chapter 4) to calculate individual or group locations. Survey
routes were essentially confined to existing roads in Mkomazi, which introduced
various biases, as explained above. Bearing in mind these sampling biases, three
types of ground-based survey were employed.
Records of individuals or groups of large mammals (2,125 in total) were made
along all routes used by MERP researchers during 1994 to 1997 (see Figure 31.1a).
Observations were not systematic, being uncontrolled in terms of sampling effort
Large mammals of Mkomazi 487
488 Mkomazi: ecology, biodiversity and conservation
and with respect to time of day, time of year and location (most were made in the
western third of the reserve where most research occurred). Many of the observa-
tions were made during work to GIS map the roads, waterholes and other physical
features of Mkomazi. All species of large mammals were recorded. Further obser-
vations were made during the course of other research activities, when there was a
tendency to record only the less common species. The observations provide reli-
able, geo-referenced species presence data. Of records made without GPS location
Figure 31.1 Routes of
ground surveys of large
mammals in Mkomazi,
with 10 km grid squares.
(a) All tracks in the
reserve, along which most
of the opportunistic
sightings were made.
(b) The 1996 dry season
survey route, surveyed
(c) The 1996 periodic
survey routes in north-west
data, only those which could be precisely located, using described positions in
relation to Mkomazi GIS map features, were used for distribution mapping.
Dry season surveys
Systematic, dry season surveys were conducted along roads throughout the re-
serve (see Figure 31.1b). Four such surveys, each lasting three to four days in July
to September 1996 (13–16 July, 27–30 July, 19–22 August and 31 August–2 Sep-
tember), provided a spatially extensive (the survey route was 393 km) although
seasonally restricted, and very modest sample of the reserve. A total of 669 sightings
of individuals or groups mammals were made. In addition, the duration of each
survey meant that given locations were ‘observed’ at different times throughout
the day. Diurnal variation in animal behaviour means that the probability of ob-
serving a given species changes through the day, and that surveying in late morning
and early afternoon is the least ‘efficient’ for detecting individuals.
Apart from generating presence data for species, the dry season surveys were
used to estimate numbers of individuals of certain species in Mkomazi. As al-
ready indicated, large mammal populations are generally bigger in the wet season.
The dispersed nature of wet season populations and the difficulties of observing
them in thicker vegetation resulted in inefficient sampling. The more open nature
of vegetation in the dry season means that transects ‘sample’, on average, a greater
area than they would in the wet season.
Transect widths were calculated from the average distances at which animals
were sighted, and multiplied by the route length to give an approximate sample
area. Estimates of total species abundance in the reserve were made by scaling up
the sample area density to the area of the reserve, for species which were sighted
ten or more times and when the distance of the animal from the road was greater
than two metres.
This method of abundance estimation ignores the influence of spatial hetero-
geneity in habitat type on variation in species abundance as it assumes that species
recorded during the survey have an equal chance of being observed in any part of
the reserve. Taken together, the high degree of habitat heterogeneity in Mkomazi,
the various biases associated with ground-based surveys (see above) and the fact
that the method is sensitive to low numbers of sightings mean that these abun-
dance estimates can only be treated as approximate indications.
At the same time as the dry season surveys were made, densities of large mam-
mals were calculated from counts made around three of the waterholes that were
surveyed by Harris (1972). This study was an attempt to detect changes or trends
that may have occurred over the past three decades. The values for biomass units
used in these calculations were the same as those used by Harris (1972). The cal-
culated densities cannot be exactly compared because Harris classified individuals
by age and sex whereas we did not distinguish between adults, instead using his
biomass figure for unclassified adults. In addition, any comparison between Harris’
Large mammals of Mkomazi 489
490 Mkomazi: ecology, biodiversity and conservation
and our observations must be made with great care because we could not take
account of the possible differences in water availability (itself a strong influence
on mammal densities) between observation periods.
The aim of these surveys was to represent temporal changes in species distribution
across the western third of the reserve during a one year study, in 1996. The sur-
veys were systematic and spatially intensive. They were carried out along four
road routes in north-west Mkomazi (see Figure 31.1c), chosen to represent major
habitats in north-west Mkomazi. Surveys were carried out twice each month dur-
ing January to November 1996, although for practical reasons the number of surveys
of each route in each month varied (see Table 31.1).
Surveys involved two or more observers inside a vehicle, and were standard-
ised to occur as soon after dawn or before dusk as was possible, when animal
activity was at its greatest. Presence data only were used in this study: a total of
875 sightings were made of individuals or groups.
Combining all of the ground-based surveys, a total of 3,542 geo-referenced
sightings, comprising 24,033 individuals, was made (see Table 31.2).
In addition to the ground-based surveys, regular flights, for various purposes,
have been made across the entire reserve by Tony Fitzjohn, resident at Kisima
since 1989. The largely opportunistic sightings have been used to supplement
data on species distributions and also to provide information on large mammal
movements. Several, more systematic, aerial surveys of the reserve were made by
SKE in July or August of 1994, 1995 and 1996.
In analysing the results of surveys of the distribution and abundance of large mam-
mals in Mkomazi, it is important to keep in mind that the reserve is part of the
greater Tsavo ecosystem and is at its southern limit. Animals move widely over
Table 31.1 Number of surveys on regular routes in north-west Mkomazi in 1996.
route name & length Jan Feb Mar Apr May June July Aug Sept Oct Nov
Kavateta (27.9 km) 1 3 2 1 1 0 1 2 1 2 1
Ngurunga (11.2 km) 0 4 2 1 1 1 1 2 1 3 1
Vitewini (37.6 km) 4 0 4 1 3 2 2 2 1 2 2
Zange (13.5 km) 0 4 4 2 0 1 2 2 1 2 0
the whole area and their recorded presence or absence is sensitive to relatively
small spatial changes in location, which may place them in Mkomazi or Tsavo.
Movements in general are governed by rainfall (see Chapter 2) with Mkomazi
acting as a wet season retreat for many of the animals because of the higher rain-
fall than that in the neighbouring Tsavo West National Park. The seasonal
movements of some of the more important species are considered in Species dis-
Ground-based versus aerial survey methods
As indicated previously, aerial and ground-based surveys are subject to various
significant biases. An idea of the variation in ‘efficiency’ of each of the survey
methods used in this study can be gained from a comparison of observation data
‘simultaneously’ gathered by each method for the same area. A road transect was
driven around Magunda at the same time as an aerial count of species was made.
The results are given in Table 31.3.
Table 31.2 Summary of presence data for large mammal
species in Mkomazi, gathered during ground-based
species sightings individuals
dikdik 735 1,178
giraffe 477 2,761
kongoni 442 2,584
zebra 376 6,754
Grant’s gazelle 264 1,278
impala 248 1,886
lesser kudu 156 263
eland 138 1,187
steinbuck 118 138
gerenuk 110 207
buffalo 102 3,376
warthog 76 204
elephant 71 1,291
reedbuck 43 73
oryx 40 387
lion 37 243
waterbuck 32 124
jackal 29 43
duiker 28 31
bushbuck 20 25
total 3,542 24,033
Large mammals of Mkomazi 491
492 Mkomazi: ecology, biodiversity and conservation
The results highlight ways in which each survey technique under-samples spe-
cies presence and abundance. Aerial surveys are better at recording animals away
from roads but are less efficient at detecting individuals or small groups of ani-
mals, especially of smaller large mammals. The aerial counters almost certainly
flew over but failed to see a pride of eight lions with 11 cubs which was found by
the ground team. Lions are notoriously difficult to detect from the air and are best
surveyed from the ground. In addition, air-borne observers have less time in which
to make repeated checks on numbers, and the survey usually takes far less time,
reducing the relative opportunity to observe animals. In this case, the aerial survey
was completed within 50 minutes while the ground counts took several hours so
the surveys were not simultaneous. The difference in timing was probably respon-
sible for the discrepancy between the aerial and ground totals for buffalo and eland.
Most of the buffaloes seen from the air were in the hills south of Dindira and were
moving towards thick country. It is unlikely that they could have been detected
from the ground. The group of eland recorded on the ground count was certainly
not present when the aircraft flew over the region where they had been seen. Where
the animals were more widely distributed, the totals from the two methods agreed
Table 31.4 compares the estimates of the numbers of large mammals in Mkomazi
Game Reserve made in the 1960s by Harris (1972) with those made in the 1990s
by Inamdar and by this study in 1996. The 1996 estimates are conjectural and are
based on the 1996 dry season systematic ground surveys, supplemented by Fitz-
john’s aerial observations. The 1996 estimates are mainly of comparative value
and probably do not represent the true totals. The 1994 aerial totals are also taken
into account although aerial surveys are known to underestimate numbers, par-
Table 31.3 The numbers of animals observed during simultaneous aerial and
ground counts of large mammals in part of Mkomazi on July 28, 1996.
species aerial total ground total difference
buffalo 87 0 - 100%
eland 6 76 + 92%
elephant 14 21 + 33%
giraffe 46 26 - 43%
Grant’s gazelle 5 11 + 55%
kongoni 25 10 - 60%
waterbuck 3 3 0%
zebra 216 181 - 16%
lion 0 19 + 100%
ticularly of the smaller species. Any one-off count represents only a snapshot of
the situation and needs to be treated with caution.
In view of the various techniques used and the seasonal differences when the
counts were made, close comparisons are not justified but the results suggest that
there has been little change except for a tendency towards an increase in numbers
of the larger species (which may, however, be due to improved sampling tech-
niques). The huge apparent increase in eland in 1994 is probably due to sampling
error in the aerial survey total. Buffalo, giraffe and zebra have shown substantial
increases and of the large mammals, only elephant has shown a marked decrease.
Although not included in the analysis, the black rhino has also decreased, from
several hundred to none. The declines in elephant and rhino are not surprising,
given the known extensive poaching for ivory and horn in the intervening years.
The seasonal changes in numbers were investigated by Harris (1972) from counts
made in three study areas surrounding semi-artificial water holes. Changes be-
Table 31.4 Minimum estimates of the numbers of large mammals based on aerial and
ground counts made in Mkomazi Game Reserve between 1964 and 1967 by Harris
(1972), in 1994 by Inamdar (1994) and in 1996 by the present authors. Not all species
were counted on each occasion. Numbers in parentheses are standard errors.
date and season
species 1960s wet 1994 wet 1996 dry
buffalo 750 1,858 (1,569) – a
eland 500 2,421 (1,279) 473 (1,313)
elephant 3,000 477 (304) 314 (149) b
gerenuk 250 17 (16) 933 (141)
giraffe 250 545 (76) 979 (84)
Grant’s gazelle – – 306 (89)
impala 600 801 (348) 3,564 (2,470)
dikdik – – 55,978 (8,153)
kongoni 1,000 511 (200) 840 (229)
lesser kudu 250 426 (71) 5,739 (2,417)
oryx 400 102 (97) – c
steinbuck – – 554 (339)
warthog – – 1,460 (704)
waterbuck 150 17 (16) –
zebra 400 460 (178) 1,438 (741) d
a Buffalo were observed but numbers were too small for estimating population size
b 500 elephant were estimated by Harris (1972) to be present in 1960s dry season
c 100 oryx were estimated by Harris (1972) to be present in 1960s dry season
d 100 zebra were estimated by Harris (1972) to be present in 1960s dry season
Large mammals of Mkomazi 493
494 Mkomazi: ecology, biodiversity and conservation
tween wet and dry seasons are shown in Table 31.5. The biggest difference was
recorded at Dindira, which is in the north-western corner of the reserve and which
is the only one of the study areas to hold permanent water. Hence its attraction for
wildlife in the dry season.
These counts were repeated in 1996 and the results are included in Table 31.5.
Too much cannot be deduced from these comparisons because of the somewhat
different techniques employed but they provide evidence of an increase in the
numbers and biomass of large mammals in the northern sector of the reserve.
Mbula is curious in that the biomass more than doubled although there was little
increase in numbers. A similar trend is apparent in the 1960s when the biomass in
the wet season was double that in the dry season although numbers remained the
same. These observations suggest that the species composition of the large mam-
mals around this waterhole is liable to fluctuate. Harris did not record eland or
zebra at Mbula and the presence of these species in 1996 may explain the discrep-
ancy. Alternatively it could be the sporadic appearance of elephants that is
responsible. The Kavateta figures were influenced by the absence of giraffe and
zebra in the 1960s and lower densities of impala and kongoni. These differences
may be due to the presence of cattle in the 1960s and their absence in 1996.
Although the trends noted are not in themselves very convincing, they all point
towards a possible increase in the numbers of most large mammals. Differences in
densities between Harris’ and our studies might, however, result from differences
in environmental circumstances (such as water availability) at the times of obser-
vation, rather than from population trends. The increase at waterholes is largely
attributable to two species, giraffe and zebra. These have relatively low standard
errors in the road count estimates and it is very likely, therefore, that the perceived
increases in these species are genuine. The decline in elephants is also likely to be
real for the same reason.
Best estimates of herbivore abundance in Mkomazi are given in Table 31.6.
These are based mainly on the 1996 dry season systematic ground surveys, whose
totals are listed in Table 31.4, but include a subjective ‘expert knowledge’ element
(SKE and Tony Fitzjohn).
Table 31.5 Numbers and biomass (kg) of large mammals per km2 around Mkomazi
waterholes in wet and dry seasons in the 1960s (from Harris 1972) compared with
those recorded during the dry season in 1996. (Kavateta was named Mzara by Harris.)
dry wet dry
waterhole number biomass number biomass number biomass
Dindira 23.7 12,705 7.7 2,082 31.6 17,329
Mbula 8.7 1,452 8.7 3,638 8.8 3,058
Kavateta 3.4 261 6.5 752 46.9 11,953
Figures 31.2–31.13 show the distributions of species for which there are numer-
ous geo-referenced sightings made in 1994–97 or which are important from a
management point of view. In considering these maps, it is very important to keep
in mind the biases associated with each survey technique. The maps do not neces-
sarily represent the limits of species distributions in Mkomazi, although they may
do so. Aerial observations, particularly of species movements, are incorporated in
the following notes on elephant and buffalo. The letters in parentheses in the text
refer to the seasonal movements shown on these maps.
BUFFALO. The seasonal movements of buffalo mirror those of the elephant to a
large extent and the routes followed in the north-west are almost exactly the same.
Some enter Mkomazi from the north-west (A) while a second wave (B) cross over
from Tsavo West National Park and spreads out over the plain between Kavateta
and Vitewini. In general, most of the buffalo occur either in the north-west or in
the south-east of the reserve with few in between (Figure 31.2). They are quite
common along the border with Kenya and three herds totalling some 350 to 500
criss-cross the border south-east of Kavuma Hill (C). Another 100 or so are resi-
dent along the Kenya border near Mabata. It is not possible to give an accurate
Table 31.6 Best estimates of the numbers of herbivores present in
Mkomazi based on ground counts made in 1996, supplemented by
incidental observations and subjective assessments.
species number comments
buffalo 2,000 wet season total
dikdik 100,000 subjective assessment
elephant 300 /1,000 dry/wet season totals
gerenuk 1,000 imprecise total
Grant’s gazelle 200
impala 5,000 conservative estimate
lesser kudu 6,000
steinbuck 600 conservative estimate
warthog 1,500 probably an overestimate
Large mammals of Mkomazi 495
496 Mkomazi: ecology, biodiversity and conservation
See text for
figure for total numbers in the reserve. A population of around 2,000 would seem
to be reasonable for the wet season.
DIKDIK is the species most frequently seen from the roads in Mkomazi (Figure
31.3). It was almost never seen on road counts passing through the seasonal swamps
or open vegetation.
ELAND are most likely to be seen in the central regions of the reserve, particularly
around the Maore waterhole although they are found in most regions (Figure 31.4).
Some 100 eland are known to enter the reserve from Kenya in the wet season
between Maore and Kamakota. As far as total numbers are concerned, extrapola-
tion from the ground surveys gives a figure of 473. General impressions, which
admittedly are notoriously unreliable, tend to support this figure.
ELEPHANTS may be found anywhere within the reserve although their distribution
is markedly clumped (Figure 31.5). In the north-west, elephants enter the reserve
in the wet season and some (A) spread south-west to the Mbula and Gulela Hills
and beyond. A second group (B) moves to the region between the Gulela and
Mzara Hills. At the same time, 100 or more elephants (C) move out of the forested
hills on either side of Dindira Dam and pass on to the plains between Zange and
Ngurunga, where they mix with the Kenyan elephants. Smaller movements (D)
across the border from Kenya to Kavateta occur if there is water in the dam. Simi-
lar small scale movements (E) across the border occur near Maore waterhole. A
group of resident elephants occurs on the western side of the Mzara Hills but they
may move out into Tsavo in the wet season (F). A more substantial wet season
immigration takes place in the Mzara/Maore region (G). Some of these elephants
pass to the west around Hafino Hill (H) and a few get as far as Kisiwani Village,
where they may raid crops, and even fewer to Njiro Gate. Most pass between
Kisima and Tussa Hills (I) to meet up with those that moved to the east of Hafino
Hill. Mating is commonly observed in this region. Some migrate further to the
south-east into the thick vegetation around Kamakota (J) where they mingle with
elephants that have entered the reserve near Kavuma (K). Altogether some 400 or
so elephants may be present in this area. Some 25 to 30 resident elephants occur
north of Mabata in the far south-east of the reserve (L) but they may move a little
way into Kenya from time to time or south to the Umba River. The total number of
elephants present in the reserve during the wet season is around 1,000 but this falls
to less than 100 in most dry seasons.
GERENUK are present throughout the reserve (Figure 31.6) but they are not easy to
see and estimates of their numbers are probably too low. The 1996 dry season
population size estimate of 933 suggests a healthy population commensurate with
the size of the reserve.
Large mammals of Mkomazi 497
498 Mkomazi: ecology, biodiversity and conservation
See text for
GIRAFFE were known to occur more or less throughout the reserve, although not
uniformly so, and there are places where they are never seen, possibly because of
heavy poaching in such regions. They are most numerous in the north-west but are
present throughout the reserve (Figure 31.7), being comparatively rare in the cen-
GRANT’S GAZELLE are restricted to the western half of the reserve (Figure 31.8), and
while not present in large numbers are most numerous in the far west. The species
was not recorded on the 1994 aerial survey, although it is not clear whether the
species was excluded from the counts. The total from the 1996 ground counts was
only 306. Even so this is likely to be too high as about half of the area included
does not support gazelles and it is unlikely that there are more than a couple of
hundred in the reserve altogether.
IMPALA is one of the commoner species and is generally widely distributed in the
reserve although it is most abundant in the western half (Figure 31.9). Antelopes
of this size and coloration are not easily seen from the air and, as with gerenuk, the
1994 aerial count of 801 is certainly too low. The 1996 dry season estimate of
3,564 suggests a population of several thousand.
KONGONI (Coke’s hartebeest) are more or less distributed throughout the reserve
although not uniformly so as there are regions of higher density, particularly near
waterholes (Figure 31.10). Like other large ungulates, they show seasonal move-
ments between Tsavo and Mkomazi and groups of 40–50 accompany eland and
zebra on their passage through the Maore region.
LESSER KUDU are probably the most numerous of the larger antelopes in Mkomazi
but because of their cryptic markings and the dense cover they inhabit, they are
not easily seen. They are distributed widely throughout the reserve (Figure 31.11
but are rare in the Ibaya region due, no doubt, to the lack of suitably thick country
ORYX are widely distributed in Mkomazi (Figure 31.12), although the population
size appears relatively low. A herd was regularly seen throughout the Mkomazi
Ecological Research Programme at Kavateta.
ZEBRA are found most frequently in the western half of the reserve (Figure 31.13)
but are present in the eastern half. The species is particularly numerous around
Ibaya and in the vicinity of the Maore waterhole. Up to 400 zebra accompany
eland and kongoni in the wet season movements from Kenya.
Large mammals of Mkomazi 499
500 Mkomazi: ecology, biodiversity and conservation
Data on the distribution of carnivores are limited, largely because relatively low
population sizes mean that species are rarely encountered during surveys. The
smaller carnivores are regularly seen and population sizes appear to be healthy.
Compared with populations elsewhere in similar habitats, the spotted hyaena is
very rare. The incidence of melanism in the serval seems to be high.
Although cattle, sheep, goats and donkeys are not supposed to be present, large
numbers of cattle, in particular, have been recorded within Mkomazi. This is per-
haps not surprising given the problems of law enforcement, in part associated
with the elongated shape of the reserve. The numbers involved run into thou-
sands. Domestic stock were counted in the 1994 aerial survey (Inamdar 1994) and
totals of 23,557 cattle (s.e. + 12,530) and 4,739 sheep or goats (s.e. + 2,356) were
estimated although not all of these were within the reserve’s borders. Cattle enter
the reserve at Pangaro, where there is a dam just outside the reserve boundary,
north of Ndea across to Kavateta, possibly including herds from Lake Jipe in Kenya,
and on the southern boundary towards the Ngurunga region. Herds have been
recorded at Kamakota in the centre of the reserve and in the south-eastern region.
Mammal species re-introductions
Following the judgement of the Tanzanian Wildlife Division in the late 1980s that
Mkomazi was in a florally and faunally degraded state, a decision was made to
rehabilitate the reserve. The Division formed the Mkomazi Project under the su-
pervision of a project manager. One of the aims of the Project was to re-introduce
species to the reserve that had recently been extirpated. Three species were ini-
tially selected for re-introduction: black rhinoceros, cheetah and wild dog. As
cheetah re-established naturally in Mkomazi, attention was focused on the black
rhino and wild dog. In view of the costs involved with re-introducing species, the
UK-based George Adamson Wildlife Preservation Trust was invited to assist with
the rehabilitation programme. The Trust provides funds for equipment and sup-
ports the activities of Tony Fitzjohn.
There was a population of at least 150 rhinoceros in the reserve as recently as the
mid-1960s but the species became extinct, largely due to poaching. A re-intro-
duction programme obviously needs a source of animals but the general decline of
the species throughout Africa to dangerously low levels was a problem. There
Large mammals of Mkomazi 501
502 Mkomazi: ecology, biodiversity and conservation
would be no justification for moving animals from where they were native unless
they were at grave risk of being killed. It became known that authorities in South
Africa wanted to dispose of a population of 35 black rhino in the Addo National
Park because they were a subspecies that was not native to the area. The popula-
tion was descended from a group of seven animals that had been shipped to South
Africa in the early 1960s from the Tsavo ecosystem, and so belonged to the appro-
priate subspecies, Diceros bicornis michaeli, for re-introduction to Mkomazi.
Guidelines for the re-introduction of species have been prepared by the Species
Survival Commission of IUCN (IUCN, 1987). These guidelines provided the frame-
work for the Mkomazi rhino re-introduction project, which was assessed and
approved by experts from the South African National Parks Board (Knight &
Morkel, 1994). It was decided not to release the animals straight into the wild but
to hold them for a number of years in a sanctuary enclosed by an electrified fence
within the reserve. The fence was completed in late 1996 and covers an area of
about 43 km2. The first four rhinoceros arrived on 4 November 1997 and were
kept in bomas for a few weeks to recover from the translocation and to acclimatise
to their new circumstances, before being released into the sanctuary.
The wild dog is an endangered species and is still persecuted throughout its range
in Africa. An attempt to re-establish the species in Mkomazi was considered to be
an important contribution to the survival of the species. There are ethical prob-
lems, however, in taking specimens from the wild for re-introduction in former
parts of the species’ range, unless the population concerned is at immediate risk of
being killed. This appeared to be the case with three groups that were located at
Engassumet on the Maasai Steppe, about 100 km from Mkomazi. The dens were
dug out in June 1995 and 25 pups (15 males and 10 females) were collected and
brought to holding pens at Kisima, close to the rhino sanctuary, before threats to
poison the dogs could be carried out. The parents of the captured pups survived
long enough to produce further litters so the exercise did not seriously affect the
status of the species in that area.
The captive dogs are breeding successfully, with the first young born in March
1997. The plan is to maintain a breeding stock in captivity and to release groups in
a series of re-introductions as well as to provide source individuals for re-intro-
ductions elsewhere. Four males were sent to Kenya to provide ‘fresh blood’ for a
re-introduction project. They were introduced to four wild-caught females with
hunting experience and after some time, two males and all the females were re-
leased into Tsavo National Park. The pack eventually found its way back to
Mkomazi where its progress is being monitored through radio-tracking.
Large mammals of Mkomazi 503
504 Mkomazi: ecology, biodiversity and conservation
The large mammal fauna of Mkomazi is diverse, resulting from both a high degree
of habitat heterogeneity and the inclusion of the reserve in the greater Tsavo eco-
system. Although population sizes in general are not great, the reserve contains
important populations of several herbivore and carnivore species, such as gerenuk,
oryx, lesser kudu, leopard and cheetah. The seasonal movements of large mam-
mals into and out of Mkomazi indicate that the reserve probably plays a key role in
the population viability of several species, by providing important wet season re-
Information on large mammal distribution and abundance is necessary for man-
agement planning purposes. The data generated by this study provide an essential
baseline of information on species presence and status in Mkomazi. This informa-
tion can be used to evaluate the potential for developing tourism within the reserve,
to predict impacts of different kinds of utilisation of the reserve, as well as to
provide the basis for monitoring impacts. The information would also be useful
in planning the re-introductions of large mammal species to the reserve.
Our thanks to the members of MERP for many of the opportunistic large mammal
observations and to the Mkomazi rangers who also participated in the systematic
Harris, L.D. (1972) An ecological description of a semi-arid East African ecosys-
tem. Range Science Department Science Series No. 11, Colorado State Uni-
Inamdar, A. (1994) Wildlife Census Mkomazi April 1994: an interim report on
results of an aerial census of the Mkomazi Game Reserve. Unpublished re-
port, Worldwide Fund for Nature, Nairobi.
Inamdar, A. (1996) The ecological consequences of elephant depletion. Doctor
of Philosophy thesis, University of Cambridge.
IUCN (1987) The IUCN position statement on translocation of living organisms.
Knight, M. & Morkel, P. (1994) Assessment of the proposed Mkomazi rhino sanc-
tuary, Mkomazi Game Reserve, Tanzania. Unpublished Report.
TWCM (1991) Wildlife Census: Mkomazi 1991. Tanzania Wildlife Conservation
Monitoring, Arusha, Tanzania.