Society for Conservation Biology

Protected Area Planning in the Tropics: Uganda's National System of Forest Nature Reserves
Author(s): P. C. Howard, T. R. B. Davenport, F. W. Kigenyi, P. Viskanic, M. C. Baltzer, C. J.
Dickinson, J. Lwanga, R. A. Matthews, E. Mupada
Source: Conservation Biology, Vol. 14, No. 3 (Jun., 2000), pp. 858-875
Published by: Blackwell Publishing for Society for Conservation Biology
Stable URL: http://www.jstor.org/stable/2641444
Accessed: 27/10/2010 12:05

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=black.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

Blackwell Publishing and Society for Conservation Biology are collaborating with JSTOR to digitize, preserve
and extend access to Conservation Biology.

http://www.jstor.org
Protected Area Planning in the Tropics: Uganda's
National System of Forest Nature Reserves

P. C. HOWARD,* T. R. B. DAVENPORT, F. W. KIGENYI, P. VISKANIC,

M. C. BALTZER,C. J. DICKINSON, J. LWANGA, R. A. MATTHEWS,AND E. MUPADA
Forest Department, Spring Road, P.O. Box 1752, Kampala,Uganda

Abstract: Uganda is one of the most biologically diverse countries in Africa, with mucl of its biodiversity
represented in a system of 10 national parks, 10 wildlife reserves, and 710 forest reserves, covering 33,000
km2 (14%) of the country's area. Wefocus on the role of the forest reserves in biodiversity conservation and
describe a procedure we developed to design a national system offorest nature reserves. In the late 1980s a
policy was instituted to dedicate half the area of forest reserves to sustainable timber production and the
other half to environmental protection (with 20% as nature reserves). To select suitable sites, a 5-year, US$1-
million program of biodiversity and resource assessment was undertaken, focusing on five biological indica-
tor species groups and covering all the major forest reserves. Based on data generated by the field studies, we
ranked each forest in terms of various criteria-(species richness, rarity, value for nonconsumptive uses, tim-
ber production, and importance to local communities) -and used an iterative site selection procedure to
choose the most suitable combination offorests for nature reserve establishment. Our procedure maximized
complementarity in representing species and habitats in reserves across the whole protected-area system. We
initially selected sites using purely biological criteria but later modified our procedure to ensure that oppor-
tunity costs and potential land-use conflicts were minimized. Our preferred network of sites included 14 for-
ests that, in combination with the existing national parks, would account for 96% of species represented in
the country's protected areas. These 14 forests were classified as 'prime" and "core"sites and were selectedfor
the establishment of large nature reserves (averaging 100 km2). The addition of 25 smaller "secondary"for-
est nature reserves (averaging 32 kmi2) would protect more than 99% of the indicator species.

Planeaci6n de Areas Protegidas en el Tr6pico: Sistema Nacional de Reservas Forestales Naturales de Uganda
Resumen: Uganda es uno de lospaises mas diversos biol6gicamente en Africa, mucha de su diversidad estd
representada en un sistema de 10parques nacionales, 10 reservas de vida silvestre y 710 reservasforestales,
cubriendo 33,000 km2 (14%) del drea delpais. Nos enfocamos alpapel de las reservasforestales en la conser-
vaci6n de la biodiversidad y describimos un procedimiento que desarrollamos para diseniar un sistema na-
cional de reservasforestales naturales. A finales de los 1980s una decisi6n politica fue tomada para dedicar
la mitad del drea de las reservas forestales para la producci6n sostenida de madera y la otra mitad para la
protecci6n ambiental (con 20% como reservas naturales). Para seleccionar sitios aptos se llev6 a cabo un
programa de evaluaci6n de la biodiversidad y los recursos por cinco anios con un costo de $1 mill6n de
d6lares. El programa estuvo enfocado a cinco grupos de especies como indicadores biol6gicos y cubri6 todas
las reservasforestales mayores. En base a datos generados en estudios de campo, clasificamnoscada bosque
en base a varios criterios (riqueza de especies, rareza, valor por usos diferentes al consumo, producci6n de
madera, e importancia para las comunidades locales). Utilizamos un procedimiento de selecci6n de sitio iter-
ativo para seleccionar la combinaci6n mds viable de bosques para el establecimiento de reservas naturales.
Nuestro procedimiento maximiz6 la complementariedad en la representaci6n de especies y habitats en reser-
vas a lo largo del sistema de area protegida. Inicialmente seleccionamos sitios usando criterios puramente bi-
ologicos, pero despue's miodificamos nuestro procedimiento para asegurar que los costos de oportunidad y los
conflictos potenciales por uso del suelo fueran minimizados. Nuestra red de sitios preferida incluy6 14

*Curr.ent AfricaProgram,WildlifeConservationSociety,2300 SouthernBoulevard,Bronx,NY 10460, U.S.A.,[email protected]

adtdress.
Papersubmnitted
April5, 1999; revisedmanuscriptacceptedSeptember22, 1999.
858

Conservation Biology, Pages 858-875

Volume 14, No. 3, June 2000
Howard
et al. Planning
ForestNature
Reserves
in Uganda 859

bosques que en combinaci6n con losparques nacionales existentespodrfan tomar en cuenta 96% de las espe-
cies representadas en las dreas protegidas delpa(s. Estos 14 bosquesfueron clasificados como sitios 'primar-
ios" y "nucleos"y seleccionados para el establecimiento de reservas naturales (promediando 100 kmi2). El
agregar 25 reservas forestales naturales pequenias "secundarias" (promediando 32 km 2) protegerfa mas del
99% de las especies indicadoras.

Introduction demics; Stuart et al. 1990; Howard 1991; WCMC 1992;

Pomeroy 1993), many of which are already represented
Protected areas provide the cornerstone of most na- in an extensive system of protected areas. Much of the
tional strategies to conserve biodiversity (World Conser- country's biodiversity is concentrated in its closed can-
vation Monitoring Centre [WCMC]1992). In recent de- opy tropical forests, however, which have suffered mas-
cades there has been rapid growth in the number of sive clearance during the course of this century. They
designated reserves in many regions of the world are now reduced to about 3%of the country's land area,
(McNeely et al. 1994), although some tropical countries whereas they originally covered about 20% (Hamilton
have begun to lose reserves because of difficulties in fi- 1984; Struhsaker 1987). Timber harvesting, hunting,
nancing effective management and the need to satisfy and other human activities within the remaining forest
growing human populations (Lowry & Donahue 1994; reserves are having major ecological effects, including
Wass 1995). The challenge facing protected area author- local extinctions of species (Howard 1991; Plumptre
ities is to ensure that any consolidation of reserve net- 1996; Johns 1997; Struhsaker1997).
works is adequately planned and managed so that re- Uganda's protected areas cover about 33,000 km2
serve networks of the twenty-first century are fully (14%of the country's area) and fall into three categories:
representative of each country's biodiversity. Systems of national parks, wildlife (formerly game) reserves, and
complementary reserves will have the greatest chance forest reserves (World Conservation Union 1992). Le-
of being maintained if they are selected to minimize the gally, no one is allowed to reside in, cultivate, or graze
opportunity costs involved in dedicating land to conser- livestock in any of these areas, although in practice vio-
vation rather than using it to address other development lations are widespread. The national parks and wildlife
goals. reserves comprise a few, generally large, areas (20 sites,
We describe an approach used to plan a nature re- averaging 1074 kM2), whereas forest reserves are much
serve system in Uganda, one of the most biologically di- more numerous, often quite small and widely dispersed
verse countries of Africa. We illustrate how the princi- (710 sites, averaging 21 kM2). Forest reserves make up a
ples of conservation biology can be applied to such a little less than half the total area protected and occur in
process and specifically how the biologically ideal con- all areas of the country, encompassing a wide range of
figuration of reserves can be modified to incorporate forest, savanna, and other vegetation types. They are of-
consideration of alternative land-use options in the de- ten surrounded by rural communities and are required
sign of reserve networks. We also show that although it to not only satisfy the subsistence needs of local people
might be theoretically desirable to create the largest pos- for forest products such as building poles and firewood,
sible reserves in each of Uganda's major vegetation but also support permitted commercial use such as tim-
types, a fully representative network of reserves requires ber harvesting, and provide environmental protection
the inclusion of a relatively large number of widely scat- (Hamilton 1984; Howard 1991).
tered, smaller sites. As a result of increasing awareness of the biodiversity
and other environmental values of Uganda's forests dur-
ing the late 1980s, a policy was instituted to designate
Uganda:a CaseStudy 20%of the forest reserve land (existing in 1988) as strict
nature reserves, 30% as low-impact "buffer zones," and
Uganda covers about 240,000 km2 of equatorial Africa the remaining 50% as areas managed primarily for the
and exemplifies many of the opportunities and con- sustained supply of forest products, particularlytropical
straints faced by conservationists in the tropics. Follow- hardwoods. These broad national allocations could be
ing many years of civil unrest, the country has recorded achieved either through designation of management
a decade of exceptional economic growth and com- zones within each forest or by dedicating entire reserves
pleted a transition to democratic government. There is to different uses. Uganda's "forest estate" comprises
growing awareness of environmentalissues and strong po- 15,950 kmi2,of which 3,190 km2 (20%) is (theoretically)
litical support for conservation. It is a country excep- destined to become forest nature reserve and 4,780 km2
tionally rich in species (although with few narrow en- (30%) to become low-impact buffer zones (excluding

Conservation Biology
Volume 14, No. 3, June 2000
860 PlanningForestNatureReservesin Uganda Howardet al.

timber harvesting). Approximately 4,150 km2 (26% of

the total) is already included within the national parks, Stage One
and a further 390 km2 (2.4%) carries dual status as forest Identification Of Candidate Forests
and wildlife reserves. If these areas are considered part
of the 50%destined for conservation management, new
forest nature reserves and buffer zones should occupy
Stage Two
an additional 3,445 km2 (21.6%).
The main goals for the new nature reserves are to en- Biological Inventories And Resource
sure (1) that viable examples of undisturbed ecological Assessment
communities, populations of all species, and genetic re-
sources are maintained in perpetuity and (2) that the re-
generative capacity of adjacent timber-producing areas Stage Three
of the main production forests is enhanced through the Data Analysis And Site Selection
establishment of a nearby reservoir of seed material, dis- v
persal agents, and ecological services. To achieve these Step 1.
goals it is intended that the nature reserves become le- Evaluateeachforest's
gally protected against all forms of consumptive use and biodiversityimportance
remain under Forest Department management as an in- v
tegral part of the national forest estate. If they are too Step2.
small, their value in addressing the first goal will inevita- Evaluateeach forest's
bly be limited. As part of a wider protected-area system, naturereservesuitability
v
nature reserves should be selected to complement the
SteD3.
role of the national parks and wildlife reserves that are
Examinecomplementarity
generally much larger areas better suited to (1) conserv-
betweensites
ing viable populations of species that live at low density, v
(2) maintaining large-scale ecological processes and phe- Step4.
nomena such as migrations, and (3) protecting blocks of Initialsite selection
habitat large enough to be responsive to short- and long- v
term disturbance events (Noss 1994). Step 5.
Evaluatepotentialaltematives
andmakefinalselection
Methods
Our nature reserve planning procedure involved four Stage Four
principal stages (Fig. 1), starting with the identification Allocation Of Management Zones Within
of a limited number of forest reserves as candidates for Each Selected Forest
nature reserves. During the second stage these forests
became the focus of detailed biodiversity and resource Figure 1. Flow chart illustrating thefour main stages of
assessment. The third stage involved analysis of field our nature reserve planning procedure, and the steps
data and selection of forests for nature reserve establish- involved in the data analysis and site selection stage.
ment. Finally,we decided on the allocation of land to na-
ture reserves and other management zones within each
elsewhere in the protected area system or considered es-
of the selected forests.
pecially important from an international perspective.
The identification of candidate forests therefore in-
volved examination of the representation of 86 vegeta-
Identificationof CandidateForests
tion subtypes in each of Uganda's protected areas (i.e.,
In the first stage we identified a limited number of candi- gap analysis;Scott et al. 1993). We plotted protected area
date forests from which to select nature reserve sites. boundaries on vegetation maps at a scale of 1:500,000
This involved eliminating from consideration a large and used a transparent grid overlay to assess the extent
number of small forests representing widespread vegeta- of each vegetation subtype within each protected area
tion types. We decided to include as candidate forests by counting grid squares. Initially we did this for all na-
(1) all forest reserves exceeding 50 kM2, because these tional parks, wildlife reserves, and the largest forest re-
are most likely to sustain viable populations of the wid- serves (53 reserves >50 km2 in size). Then we examined
est range of species in the long term, and (2) additional these results before consulting the maps again to deter-
smaller reserves representing vegetation types not found mine where vegetation subtypes missing from these pro-

C:onservationBiology
Volume 14, No. 3, June 2000
Howardet al. PlanningForestNatureReservesin Uganda 861

um GE
\
\1 (I um\ r- - C\ N CN z t t Ni 11 0 O\1 \O 01 z E
N cl
r-C \, O 0 Z u \ GEN N Z r- N GI\ Z Z r- GI\ r \O C\

0 t O
Cn Cn t VE: C\ 0 N 0C O O C4 \0to \04 00 1 \O 00 N
* z r- x z z z V0 O c Z O 0 Xf \ t1 \ N O -

> -tt N " UV N N V-t \0 \ N <t \ -

w (N 0I \1 i
z

$~~~~~
0
r-
r- \11 cn C
00 VI (I CX 00 r- cn \\- r-
Ot\1 (In 0 r- VI\
UGGI\
mt
00~ \1
\ N r- \1I \,
N
0 \, r- C' 0 Z 0 N C\ 0

z O
- u VI\ ( ( N( N - O ON N
C\ -00 -\ 00 -0 t- z z -Vm O 00 F F 00 0 N
;n 0 \t C\ O 00 0 O N VE cn N 00 Vm Vm \1 \t r- G~
C\0 N C\ N yum r N

N rq r N -I N N N N r

X ( N \(N - -- (- (N ( (N- - (N N - - -, \ 0 (N
in - N /\ \

r-
V
n C\ \O I\ It
m n r- nNC\ 00 \0 O \ GE- N O
(N
N - N r <OOunFG4OG
(N-
\ F N
0 r- C N C\\N \N
\0 t

4 ~~~N CnC \ Z VOt r-r Cn 0 Va t C'n t>\ ZO Z X

Ft
C' C0n X0 Cn C\r
<

m
Q

- (N (N ~~~~~
r- (N ., \t---N 0 \t 00 \C (N
00
\ \ Nt \
r- 0 \C' X OnX 0 0 \z
\G
<0 Cn
0 O0
E
\t X
t C .\ 00 \\- V4 V0 4NF0 u00 00 \C Vm -
Fu
\\- C VE 00\t Z in

Q ~ ~ ( -

o-(N--CN(N(N(~1-(N(N- \t - - -
o 0~~~~~~~~~~~~~~0
I I (N -e (N I

u 0 0 ? i

a t X~~~~~~~~~~P
- \

O >Z

Ct\ C\ N 00 VI 00 C4 o 0 C\ ol 000 N o x z N VI I\ o o o GI CI C\ C\ C\ C
x0 I I \t \ N N N v
t \t

Conservation Biology
Volulme 14, No. 3, June 2000
862 PlanningForestNatureReservesin Uganda Howardet al.

um o
Qt \o r- F 00 Vo \o N 0o C' T- in in um V N r- VF uV N 0o Vm in 0o C' N in V- N 0
iN - 0 --t '-\r- 0--- - -i - - Gi O
o OV 0O \0
G -0 \-N X r Z

N 1N \O
-(t 0' -

0~~~~~~~~~~~~~~~~~~~~~~~~~~0
o
E~~~~~~~ ol N s N cn \z o \o N N o- 00 r- \zoXt\zXt \o
>
X

4z
o
>~~~~~~~~~ \1 tCnV N \t CNN oo Ct r- 00 \, i 0 0\1
\ \ \\ 0-

z ~~~. - -
GE
NN ( to
\, \1 C4 N N C' T- V- Z C'n C4 0 N 0 u m0 O 0 GE N \1 40 GE O O +Z e

z~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~lz
to

(N~~~~000(N V V~~~~~~V~~~(N01(N~~~~~~ -0

Vi \.@ N ~(N 0 N r-( N( N Q N

i C\ n V-- - - FtC\ - S
r 00 inV- i -1 N
~~
Vq 0In In
0 00 In
GE< NGE O N om000 Vt 1S

to

Q~ ~ ~ ~ ~ ~ ~ ~~~~~a toS
o - 0
Volume 14,3, June
No. 2000

I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I IZ)

i~~~- S
z ~ ~ .N
- ~ ~ ~ ~ ~ ~ ~ ~ ~ Q
.t> w
> N N N N N N N f%k~~~~~~~~~~~~~~~~~~~~~
to
?
9'~~~~~~~~~~~~~~~~~~~~~~~~~~~
t.RU v
S X o m N A S N w w . w -4 m N <z
~~o N <r<
-4-4^-4 o N , 4.t;~~~~~~~~~~~~~~~~~~~~~~~
X t .>
Cz ~~~~~~~~~~~~~~~~~~~~~~~~~~~O . 0
z>~~~~~~~~~~~~~~~~~~~~~~
e Z F 4 O0 t O F F GE GE t 4 t F N F O0 N F
N;'
,.sQ
>
F
GE

It zSi
, |-\?> t.r, cLer t t -z ut, N N _s -z G> t n~~~~~~~~~~~~~4-z
<) <)
44zz>A
0t<
O0
um
N $

r | t O0 X O0 X 1- 1 14 90 90 s0 4 ovn
m
oo
veoo<<OtN
,;4% OO-- >
ot90t Z
 et al.
Howar7d PlanningForestNatureReservesin Uganda 863

Table2. Rankingof top 20 Ugandanforests for species richness,rarityvalue, and biodiversityimportance.

Species richness Rarity value Overall biodiversity importance

Rank forest scorea (nmax. 10) forest scoreb (max. 10) forest scorec (nax. 20)
1 Kasyoha-Kitomid 8.3 Echuya 9.7 Bwindi 16.6
2 Bwinidd 8.2 Rwenzori 8.8 Kasyoha-Kitomi 15.6
3 Budongod 7.7 Mt. Moroto 8.5 Budongo 15.4
4 Kalinzu-Mara'd 7.6 Bwindid 8.4 Kalinzu-Mara' 14.8
S Kibaled 7.6 Mafuga 8.1 Mt. Moroto 14.7
6 Mt. Kadam 7.6 Mt. Elgon 8.0 Rwenzori 14.5
7 Mpanga 7.6 Budongod 7.7 Kibale 14.5
8 West Btigwe 7.5 Semlikid 7.7 Semliki 14.4
9 Mpigi 7.5 Kasyoha-Kitomid 7.3 Echuya 14.3
10 Kabtuika 7.4 Kalinzu-Mara'd 7.2 Otzi 14.1
11 Bugomad 7.3 Morongole 7.2 Mt. Kadam 14.1
12 Sango Bayd 7.2 Otzid 7.0 Bugoma 14.1
13 Otzid 7.1 Mt. Kei 7.0 Sango Bay 14.0
14 Era 6.9 Kibale'" 6.9 Mt. Elgon 13.9
15 Semlikid 6.7 Bugoma' 6.8 Mpanga 13.7
16 Nyangea-Napored 6.7 Sango Bayd 6.8 West Bugwe 13.6
17 Mt. Napakd 6.6 Sesse Islands 6.8 Kabuika 13.6
18 Igwe-Lvivunya 6.6 Nyangea-Napored 6.7 Mukono 13.5
19 Agoro-Agu 6.5 Mt. Napak d 6.7 Mpigi 13.5
20 Mujuzi 6.5 Mabira 6.7 Nyangea-Napore 13.4
aMean score for the five indicator taxa: score for each taxon calculated as mean valute of pairwuise comparisons of number of species recorded
at highest level of samplingpossiblefor each pair offorests (utsing the method of Prendergast et al. 1993b); top-scoringforest assigned 10, others
scaled dozvn proportionately.
bMean score for the five indicator taxa: score for each taxonzrepresents mean value for all species where each species is scored on a scale of 1-10
proportional to its known Africa-wide distribution and 1-10 based on number of Ugaandanforest site records with total score divided by 2 to
give maximum score of 10. Africa-wide distribution data were derived from the literature anzd each species scored as 10, Albertine Rift or
Uganda country endemic; 6, Afromonttane or Uganda-Kenya-Tanzania distribution; 4, widespread throughout either Central or West African
forest blocks or throughout eastern Africa (beyond Uganda-Kenya-Tanzaniia); 3, widespread aclross the Sud1an-Guineasavannca, across the cen-
tral African forest belt into Kenya and/or Tanzania, or throughout the Central and West African forest blocks; 2, known throughout two or
three of Africa's nzajor ecoregions; 1, known throughout sub-Saharan Africa. Scores for distribution within Uganda ar-e based on comnbined
data from this and previous studies: 10, known from one forest or protected-area site; 6, 2-3 sites; 4, 4-5 sites; 3, 6- 7; 2, 8-9 sites; 1, 10 or more
sites.
cScore derived as total of species richness plus rarity value scores.
dForests ranked amoong the top 20forests on criteria of both richness and rarity.

tected areas were represented in any of the (657) 1992 and March 1995 (Table 1). The objective for each
smaller forest reserves. These smaller reserves were team was to sample as many indicator taxa as possible,
then added to the list of candidate forests. exploring the full range of habitats within each forest re-
serve. Field time was allocated in proportion to the size
of each reserve, and strict field recording procedures
BiologicalInventoriesand ResourceAssessment
were maintained to track species accumulation rates.
The most intensive phase of the work involved detailed This facilitated later comparison between sites by means
biological inventory and other resource assessment of all of observed species totals for given levels of sampling
candidate forests as a more refined basis for selection of (Prendergast et al. 1993b). Return visits were made to
nature reserve sites. We used a species-based approach most forests by a different team during different seasons,
involving selected indicator groups that were chosen on although in Uganda's equatorial climate seasonal effects
the basis of practical and biological criteria, including on inventory results were minimal, except in the ex-
ease of sampling and the availability of resources (such treme north (Table 1). Specific field sampling tech-
as identification guides and reference collections), nec- niques for each of the five indicator groups were based
essary for their study. In order to include plants and ani- on current recognized methods, described elsewhere
mals that were as taxonomically different as possible, in- (Howard & Davenport 1996).
cluding representatives from flying and nonflying groups,
vertebrates and invertebrates, we selected woody plants,
five families of small mammals (Cricetidae, Gerbillidae, DataAnalysisand Site Selection
Muridae,Myoxidae, Soricidae), birds, butterflies, and two
families of large moths (Saturniidae,Sphingidae). Species lists generated through the fieldwork were used
Four inventory teams, each with five to six members, initially to compare and rank sites in terms of their biodi-
carried out the field work in 64 forests between May versity values. This provided a basis for nature reserve

Coinservationi Biology
Voltume 14, No. 3, June 2000
864 PlanningForestNatureReservesin Uganda Howardet a.

Table 3. Ranking of top 20 Ugandan forests for nature reserve suitability.

Biodiversity Timber Community

importance' Nonconsumptive potentialc used
Rank Forest (max. 20) useb (max. 10) (max. 20) (max. 20) Nature reserve suitability e
1 Mt. Moroto 14.7 7 0.6 0.5 20.6
2 Mt. Kadam 14.1 5 0.6 0.5 18
3 Mt. Napak 13.2 4 1.6 0.2 15.4
4 Morongole 13.2 4 0.6 1.2 15.4
5 Timu 12.1 2 0 0.5 13.6
6 Nyangea-Napore 13.4 4 4 0.2 13.2
7 Ogili 13.1 2 0 2.2 12.9
8 Agoro-Agu 12.8 4 3.6 0.8 12.4
9 Rom 12.2 3 3.6 0.2 11.4
10 Lwala 12.1 2 1.6 1.5 11
11 Otzi 14.1 3 4.6 2.5 10
12 Era 13.4 2 4 1.6 9.8
13 Kabuika 13.6 0 4 0.1 9.5
14 Labwor 12.8 2 4 1.6 9.2
15 Maruzi 10.7 1 3 0.7 8
16 Kasagala 12 2 4 2.7 7.3
17 Aswa 11 0 4 0.3 6.7
18 Kapimpini 11.2 0 4 0.4 6.8
19 S. Busoga 12.5 3 6.2 3 6.3
20 Kamusenene 10.4 0 4 0.2 6.2
aSee Table 2for derivation.
bDerived as sum of scores given for recreational (max. 4), watershed (max 4), and education-research value (max. 2).
cDerived as sum of scores for standing timber volume (max. 10), forest accessibility (max 3), sawmill investment (max 3), and plantation de-
velopment potential (max. 4).
dDerived from assessments (based on government census statistics) of number of potential resource users living in forest-edge subcounties per
unit area offorest, with highest-ranking forest assigned score of 20 and others scaled down proportionately.
eDe?.ivedas sum of biodiversity importance and nonconsumptive use, minus timber potential and community use.

site selection and development of a procedure to allo- potential for complementary uses, offset against its po-
cate forest land to different uses. tential for alternative uses. For example, a biologically
important area encompassing the highest parts of a
STEP 1: EVALUATION
OF BIODIVERSITY
IMPORTANCE mountain range with important watershed functions
and high potential for recreational use is likely to be par-
We first derived a biodiversity importance score for ticularly suitable as a conservation area, especially when
each forest based on species richness and the rarity it is poorly stocked with timber, inaccessible, and sur-
value of the species represented. We first considered rounded by sparsely populated communal lands.
each of the five indicator groups alone before calculat- We used nature reserve suitability scores as a means to
ing combined "average"scores for species richness and (1) ensure that "least cost" (suitable) nature reserve op-
rarityvalue for each site. For species richness we used a tions were taken in preference to more expensive alter-
measure of the relative number of species per unit area natives (even though these may be more efficient in
(Prendergast et al. 1993b; Table 2). For rarity we calcu- terms of the number of sites required to achieve biodi-
lated a score for each species based on the number of versity conservation objectives) and (2) adjust the area
Ugandan records for that species and its Africa-wide dis- dedicated to conservation uses within individual sites
tribution (Table 2). (Stage Four, Fig. 1).
This step in the data analysis and site selection proce-
STEP 2: EVALUATION
OF NATURERESERVESUITABILITY
dure involved scoring nature reserve suitability for each
site as conservation value minus alternative-use poten-
Nature reserves and other conservation areas are most tial. Conservation value was taken as the sum of the
likely to be maintained in the long term if they provide biodiversity importance score (described in step 1) and
additional benefits that complement their role in biodi- the potential for compatible nonconsumptive uses (i.e.,
versity conservation and if they are located in areas with tourism, watershed protection, education, and research);
little or no potential for alternative consumptive uses (Tables 2 & 3). Alternative-use potential was taken as
such as timber production or community use. The suit- the sum of a forest's commercial forestry prospects and
ability of any particular forest as a conservation area can the potential for community use of its resources (Table
thus be evaluated in terms of its biodiversity value and 3). Full details of the derivation of these scores are pro-

Conservation Biology
Volume 14, No. 3, June 2000
Howard et al. Planning Forest Nature Reserves in Uganda 865

vided in the Forestry Nature Conservation Master Plan erence to sites that were suitable for nature reserve es-
(Uganda Forest Department 1999). tablishment. Therefore the complementarity table was
continued with the addition of all sites classified as be-
STEP 3: COMPLEMENTARITY
ANALYSIS
ing of high nature reserve suitability, followed by sites of
medium suitability, and finally sites of low nature reserve
Although it is useful to evaluate sites for biodiversity im- suitability.In this way we deliberately disadvantagedfor-
portance and nature reserve suitability, site selection ests that were heavily stocked with timber and located
based solely on such criteria is likely to be inefficient be- in densely populated rural areas in the search for nature
cause several of the more biologically important or suit- reserve sites that could be more readily protected and
able sites may be similar. Protecting several examples of maintained at lower costs in terms of alternative uses
a particular community, habitat, or suite of species may foregone.
be difficult to justify if others remain unprotected. For We considered a wide range of alternatives, details of
this reason, we constructed complementaritytables based which are provided elsewhere (Uganda Forest Depart-
on species records, starting with the richest site and fol- ment 1999). Exploring them enabled us to reach deci-
lowed by the one that complemented it best by adding sions about the criteria to use in deriving the final pre-
the most species not already represented in the first site. ferred complementarity table, which was built up in
We built up a list of sites in this way, each one adding to layers, each involving a defined group of sites. At the top
the total list of species represented, until the full com- of the final table were national parks (a group of 9 sites;
plement of species known from all of Uganda's pro- Mgahinga excluded for lack of data). These were fol-
tected areas was included. This list represents the "mini- lowed by a second group of 3 sites that contributed at
mum critical set" of sites required to represent all least 2%of species overall in the second complementar-
species at least once. The procedure has been used ity analysis in which sites were introduced to minimize
widely elsewhere (Margules et al. 1988, 1994; Vane- selection of those where land-use conflict would be
Wright et al. 1991; Pressey et al. 1993, 1994; Johnson likely. In the following order we then added (1) sites
1995; Lombardet al. 1995). with concentrations of species not found elsewhere in
We combined data for the five indicator groups in the Uganda's protected areas (6 additional sites each ac-
complementarity analysis as an average of the percent- counting for at least 1%of the species total within any
age of plants and the percentage of animals added by taxon); (2) sites contributing at least 1% of remaining
each forest. The animal complement was derived by av- species within at least one taxon (5 sites); (3) sites with
eraging the complements added for each of the four ani- a significant number of species and/or at least one "en-
mal groups, expressed as a percentage. Thus, in the demic" species not found elsewhere in Uganda's pro-
overall assessment, 50%of the recorded complement of tected areas (18 sites); and (4) sites supporting a vegeta-
species contributed by each forest was attributableto its tion subtype not already represented (7 sites).
trees and shrubs, whereas 12.5% was contributed by
each of the animal groups (butterflies, birds, mammals,
STEP 4: INITIALSELECTIONOF NATURERESERVESITES
and moths).
In a significant departure from normal complementar- All forest sites included in the final complementarity ta-
ity analysis we then carried out a series of further analy- ble were thus selected for nature reserve establishment.
ses to explore different configurations of the protected- To differentiate between sites according to their impor-
area system. The normal procedure-adding sites to tance in the national protected-area system, we distin-
maximize the number of species represented at each guished three categories: prime, core, and secondary.
stage-produces the most "efficient" system, account- The more biologically important prime and core sites
ing for the greatest number of species in the fewest pos- were allocated substantially larger areas for designation
sible sites. We recognized, however, that it may be pref- as nature reserves so as to maximize their chances of re-
erable from an economic, social, or political perspective maining viable in the long term. We defined prime sites
to adopt a less efficient system, involving selection of a as those contributing more than 2% of the total pro-
greater number of sites, if these sites can be selected to tected-area species complement in the final complemen-
minimize potential land-use conflicts. Therefore, we be- tarity analysis (i.e., the second group in the complemen-
gan our next complementarity analysis by considering tarity table), core sites as those that represent the only
areas designated as national parks at the top of the table, known locality for at least 1%of the protected-area sys-
followed by steep mountain catchment reserves. Biodi- tem's species total (third group), and any other sites
versity within these areas can be considered preferen- contributing 1-2% of species within any taxon in the fi-
tially protected because the necessary political decisions nal complementarity analysis (fourth group). Secondary
and institutional mechanisms are already in place. The sites were defined as those where (1) 0.5-1% of species
next stage was to determine which forest reserves added known from the protected-area system as a whole occur
the most species to these preselected areas, giving pref- only at that site, (2) at least one species of (possible) in-

Conservation Biology
Volume 14, No. 3, June 2000
866 PlanningForestNatureReservesin Uganda Howardet at.

ternational conservation concern occurs, (3) a vegeta- the size of a forest's nature reserve allocation to allow
tion type occurs that is not otherwise represented, or low-impact uses in areas immediately adjacent to each
(4) 0.5-1% of species within any taxon are contributed nature reserve. For each forest the larger of these three
to the final complementarity analysis (fifth and sixth options was chosen.
groups).

POTENTIALALTERNATIVE
STEP 5: EVALUATE NATURERESERVESITES Results
Recognizing that the application of rigid criteria may not
CandidateForests,BiologicalInventories,and
be practical (for example, somebody may seek to estab-
lish a gold mine in one of the selected sites), we exam- ResourceAssessment
ined the potential for alternatives to the preferred sites. Vegetation analysis revealed that all 22 major vegetation
We carried out cluster analysis, based on the TWINSPAN types are represented somewhere in Uganda's pro-
procedure for each taxon, to identify groups of forests tected-area system, although 32 of the 86 recognized
supporting similar suites of species that might be substi- subtypes are missing (or represented by <50 kM2) from
tuted for one another if necessary. the country's national parks and wildlife reserves (ex-
cluding the degraded Karamojawildlife reserves). These
32 subtypes all occur in forest reserves, and representa-
Allocationof ManagementZoneswithinEachForest
tive areas were identified and included in the list of can-
Having selected our list of forests, we proceeded to allo- didates for nature reserves. Altogether 64 of Uganda's
cate a proportion of each site's total area for designation 710 forest reserves were selected for further investiga-
as nature reserves, recognizing the relative importance tion, including all 53 reserves of >50 km2 and 11 smaller
of the three categories of prime, core, and secondary. In ones representing these 32 vegetation subtypes (Table
making these allocations, we decided to apply specified 1; Fig. 2).
absolute minimum areas, aimed at maximizing the long- The biological inventory program involved approxi-
term maintenance of biodiversity at smaller sites. For mately 6000 person-days of field sampling in 64 forests
prime forest sites we allocated 30-35% to nature re- and resulted in about 6400 butterfly species site records,
serves, subject to an absolute minimum size of 100 km2. 8200 bird records, 750 mammal records, 2250 moth
For core sites we allocated 20-30% to nature reserves, records, and 11,500 tree records (Table 1). The results
subject to an absolute minimum size of 50 km2 (not to are described in detail in the 33-volume Forest Biodiver-
exceed 70% of a site's total area because local people sity Report series (Howard & Davenport 1996).
cannot be totally excluded). For secondary sites we allo-
cated 10-20% to nature reserves, subject to an absolute
minimum size of 20 km2 (not to exceed 70% of a site's DataAnalysisand Site Selection
total area).
BIODIVERSITYIMPORTANCE
AND NATURERESERVESUITABILITY
Within each category we decided on nature reserve al-
locations for each forest on the basis of their nature re- For illustrative purposes we have provided a list of the
serve suitability scores, with low, medium, and high suit- top 20 forests ranked in terms of species richness, rarity
ability ratings corresponding to (arbitrary) 5% intervals value, and overall biodiversity importance (Table 2). Our
in allocations. To satisfy the requirement that 20%of the results confirm earlier work (e.g., Hamilton 1974; Struh-
(1988) forest estate be designated as nature reserve, we saker 1981) that has consistently rated the larger closed-
included five new forest national parks (formerly forest canopy rainforests of western Uganda as being of great-
reserves, redesignated while the planning process was est biodiversity importance. Unfortunately, only four of
underway) in the prime sites category and made theoret- the most important forests for biodiversity (Mt Moroto,
ical nature reserve allocations for them as if they were Otzi, Kadam, and Nyangea-Napore) are also among the
still forest reserves (in practice there is no intention to top 20 for nature reserve suitability (Table 3). This is be-
designate nature reserves within national parks). cause the most valuable forests biologically are often im-
To satisfy the policy goal of designating an additional portant timber producing areas, located in the more
30%of the forest estate primarilyfor environmental pro- populated parts of the country where local community
tection, we made buffer zone allocations as either (1) demands for forest products are high.
the remainder of areas (not already allocated as theoreti-
cal nature reserves) that lie within national parks and
ANALYSISAND INITIALSELECTIONOF NATURE
COMPLEMENTARITY
wildlife reserves; (2) all areas of >150 slope minus na-
RESERVESITES
ture reserve allocations already made, on the assumption
that all steep slopes will be dedicated to conservation as Most of the forests included in the assessment program
either nature reserves or buffer zones; or (3) an area half (52 out of 64) support at least one species not recorded

Conservation BiologyZ
Volume 14, No. 3, June 2000
Howard et al. Planzing Forest Nature Reserves in Uganda 867

l l UDAN
~~~~~~~S
I \ ~
~~~~~~~~~~~~~~~~~~~~~~~KIDEPO
VALLEy

UGANDA Mgr-orongole

Equator
~ ~ ~~~~~~M tLokun
Nyangea-NaporeLwa
Tims
Era
46AswaRiver
Rom
Zoka at

( ~ ~ ~~~~~
0 Kil:ak Ogili

AJAI'sISLANDWR Wiceri

N / Opitit
Labswor Hlis
1
+ MUR~~~~CFIISON FALL
~~
\_\N~NAIOAL PARK , Mt Moroto

KARl)MA WR Napak
BUGUNtGA WRll

DEMOCRATIC Kibeka

REPUBLIC ALBERT.. ongo Maruzi Hills,

//-_Kada)
OF CONGO

Wabisi-Wajala NIT ELGON

J Bugoma Kabu-MUJ
Kamusenene KapimpaiMuj
NATIONAL PARK

RO MaR as Bu
SEMLIKI
NATIONAL PARK KTaala bP Kasagala
Luungal
Itwara 4 Bwezigolo-Gunga
40 Mbr
QtlEEN ELtZABETg 0 NsoweMpanza?SamaaWest
> ; Bug-ve

MTRWVENZO.
r'Matiri
~Kasana-Kasambva
p,0q/ Baleba E Y
NATIONAL PARK NTOALE PAK D Igwve-Luvunva

,/ isangi Nsowe *?~oI ~

QUEENELIZABH is vtag
NATIONAL PARK Kazooba
. Z7 .i*?O* \ :
Ji KYAMBURAWVR
LAKE
EDWARD Kasyoha-Kitomi Jubiya |

Kalinzu- Sesse Islands

KIGEZI WR Maraniagambo K Mu :. ..

g a ~~~~~~~~~~~~Klianabolola
LAKE
*
L.MBURO Sango Bay VICTORIA
BWINDI IMWlPENETRABLE NATIONAL PARK
NATIONAL PARK
Rwh
9Zz ~~~~~Rwoho
% Mafuga >_

Echuya TANZANIA 0 0 2
Kilometres

Figure 2. Location of Uganda's nationalparks, wildlife reserves (WR, horizontal lines), and the 64forests (black)
investigated during biological inventory and resource assessment.

elsewhere (Uganda Forest Department 1999). Thus a When we compared the results of examining site com-
protected-area system designed to conserve all species plementarity on purely biological grounds (Table 5)
would need to include at least these 52 sites. For illustra- with those derived from our modified procedure (Table
tive purposes we provide a list of the most important 37 6), we found in each case that 13 sites were required to
sites, including all those that represent the only Ugandan account for 90% of species and 20 sites were needed to
site for at least one narrow endemic and/or those that reach the 95%level. If sites were selected on purely bio-
represent at least 0.5% of species within any indicator logical criteria, six forest reserves would be included in
group in that forest only (Table 4). the top 13 sites, but two of these become redundant

Conservation Biology
Volume 14, No. 3, Jtune 2000
868 PlanningForestNatureReservesin Uganda Howardet al.

Table4. Numberof knownfroma single forest and correspondingnumberof these forest-specificspecies thatare Ugandanor
regional endemicsa

Trees Butterflies Birds Mammals Moths

Forest" endemics total endemics total endemics total endemics total endemics total
Mt. Elgonc 8 18 0 1 4 5d ? id 2 3
Rwenzoric 3 lod 1 2 2 3 5 5d o 0
Budongoc 4 32d 0 2 1 4 0 0 ? 4d
Kalinzu-Mara'C 0 2 3 9d 0 0 ? id 0 O
Kibale 0 3 0 2 1 1 0 0 0 1
Mt. Morotoc 1 3 7 9d 7 13d 2 3 4 8d
Labwor 2 4 0 3 0 1 0. 0 0 0
Nyangea-Napore 3 4 1 4e 0 0 0 0 0 0
Bugoma 0 0 0 7e 0 0 0 0 0 2e
Mt. Kadam 1 1 0 1 0 0 0 0 0 0
Kasyoha-Kitomic 1 3 1 0ld O 1 0 0 0 0
Mt. Keic 0 3 1 7e 2 4 0 id ? 2e
Mabira 1 0 6 0 1 0 0 0 1
Agoro-Agu 4 7e 0 0 0 0 0 0 0 0
Semlikic 1 6e 1 34 1 37 0 id ? 3d
Mt. Napak 0 2 0 2 0 0 0 0 1 1
Otzi 2 7e 1 3 0 0 0 0 0 0
S.Busoga 1 2 0 2 0 0 0 0 0 1
Bwindic 1 6e 7 lod 4 ge 0 id I 1
Sango Bayc 0 3 1 8d 1 2 0 0 0 1
Morongole 2 8e 1 2 0 0 0 0 0 0
Timu 5 6e 2 5e 0 0 0 0 0 0
Rom 1 3 0 0 0 0 0 0 0 0
Kasagala 1 2 0 1 0 0 0 0 0 0
Rwoho 1 1 0 3 0 0 0 0 0 0
Itwara 1 1 0 1 0 0 0 0 0 0
Erac 1 3 0 0 0 0 0 0d O 1
Kijanabolola 0 1 1 1 0 0 0 0 0 0
Lwala 0 0 1 1 0 0 0 0 0 0
Echuya 3 5e 3 4e 0 1 0 0 0 0
Mafuga 2 3 1 1 0 0 0 0 0 0
West Bugwe 1 2 0 1 0 0 0 0 0 0
Igwe-Luvunya 0 1 1 1 0 0 0 0 0 0
Mpigi 1 2 0 0 0 0 0 0 0 0
Sesse Islandsc 0 8e 1 3 0 o 1 0d O 1
Jubiya 0 0 2 3 0 0 0 0 0 0
Lokung 1 1 0 0 0 0 0 0 0 0
aEnde,nic refers to species apparently restricted Uganda, the Afromontane, or Somalia-Maasai regional centers of endemism (White 1983).
bForests are listed if they support at leastfive trees,four butterflies, five birds, or two moths (0.5% of the protected-area total) or at least one en-
demic recorded in no other Ugandan forest.
CCoreforests for nature-reserve establishment.
dNumbers shown represent >1% of species recorded from all protected areas combined.
eNumbers shown represent 0.5-1% of species recordedfrom all protected areas combined.

when the analysis is modified to preselect national parks 1999) and can therefore be represented in the pro-
and suitable forests. In the final complementarity analy- tected-area system as forest nature reserves. Examina-
sis, nine of Uganda's national parks (group 1) account tion of these data resulted in the addition of seven sites
for 77% of species, and the addition of a further 14 for- in the final complementarity analysis (Table 6).
ests (groups 2-4) accounts for more than 96%of species
in 23 prime and core sites (Table 6). A more complete
ALTERNATIVE
NATURERESERVESITES
protected-area system, (accounting for over 99% of spe-
cies), would include 49 sites (Table 6). Cluster analysis of each of the five indicator taxa showed
Altogether, 32 (out of 86) vegetation subtypes are not that sites were consistently grouped into six or seven
represented (or are underrepresented) in the country's main groups, loosely corresponding to (1) dry northern
national parks and wildlife reserves, but most of them mountains and hill savannas, (2) moist savannas, (3)
occur in at least one forest (Uganda Forest Department moist savanna-forest mosaics, (4) mid-elevationlakeshore

Conservation Biology
Volume 14, No. 3, June 2000
Howardet al. PlanningForestNatureReservesin Uganda 869

Table 5. Complementarity table showing the minimum critical set of sites based on biological criteria without consideration of legal
status of sites or opportunity costs of their selection.*
Butterflies Birds Mammals Moths Trees Fauna Combined
(n = 736) (n = 959) (n = 90) (n = 207) (n = 985) Trees average
Site add % add % add % add % add % (total %) (total %) add % total %
Budongo 34.51 37.43 26.67 62.80 47.21 47.2 40.3 43.78 43.75
Queen Elizabeth National Park no data 29.09 8.89 no data 9.24 56.4 49.8 14.11 53.12
Bwindi 10.33 9.38 18.89 11.11 8.32 64.8 62.3 10.38 63.54
Kidepo Valley National Park no data 11.57 10.00 no data 3.96 68.7 67.7 7.37 68.22
Semliki 15.63 5.21 11.11 2.90 2.34 71.1 76.4 5.52 73.73
Moroto 6.93 1.98 4.44 5.80 3.65 74.7 81.2 4.22 77.96
Mt. Elgon 2.04 1.25 2.22 2.90 4.37 79.1 83.3 3.23 81.19
Otzi 2.99 0.21 2.22 2.90 4.37 83.5 85.3 3.22 84.38
Rwenzori 1.49 0.31 7.78 0.48 1.83 85.3 87.9 2.17 86.59
Kasyoha-Kitomi 4.48 0.10 0.00 0.00 1.52 86.8 89.0 1.33 87.90
Mt. Kei 1.90 0.73 1.11 1.93 0.71 87.5 90.4 1.06 88.96
Lake Mburo National Park no data 0.73 1.11 no data 1.12 88.6 90.9 1.02 89.76
Sese Islands 1.77 0.00 1.11 0.48 1.12 89.7 91.7 0.98 90.72
Labwor Hills 0.95 0.10 2.22 0.48 1.02 90.8 92.7 0.98 91.73
Sango Bay 2.31 0.31 0.00 1.93 0.61 91.4 93.8 0.87 92.59
Kibale 1.49 0.10 0.00 1.93 0.51 91.9 94.7 0.70 93.29
Morongole 0.68 0.00 0.00 0.00 1.02 92.9 94.9 0.59 93.90
Echuya 1.36 0.10 0.00 0.00 0.71 93.6 95.2 0.54 94.40
Timu 0.95 0.00 0.00 0.00 0.81 94.4 95.5 0.52 94.96
Kalinzu-Mara' 1.49 0.10 1.11 0.00 0.30 94.7 96.1 0.49 95.41
Era 0.00 0.00 1.11 0.48 0.51 95.2 96.5 0.45 95.86
Nyangea-Napore 0.95 0.00 0.00 0.00 0.51 95.7 96.8 0.37 96.27
Mabira 1.36 0.21 0.00 0.97 0.10 95.8 97.4 0.37 96.62
Murchison Falls National Park no data 1.04 0.00 0.00 0.20 96.0 97.7 0.36 96.87
Bugoma 1.49 0.00 0.00 0.97 0.00 96.0 98.3 0.31 97.17
Agoro-Agu 0.14 0.00 0.00 0.00 0.51 96.5 98.3 0.28 97.42
Napak 0.27 0.00 0.00 0.48 0.30 96.9 98.5 0.25 97.68
Mpanga 0.00 0.00 0.00 0.48 0.30 97.2 98.6 0.21 97.88
Mafuga 0.14 0.00 0.00 0.00 0.30 97.5 98.7 0.17 98.08
West Bugwe 0.14 0.00 0.00 0.00 0.30 97.8 98.7 0.17 98.23
Aswa River 0.00 0.00 0.00 0.00 0.30 98.1 98.7 0.15 98.39
Zoka 0.00 0.00 0.00 0.00 0.30 98.4 98.7 0.15 98.54
South Busoga 0.27 0.00 0.00 0.48 0.10 98.5 98.9 0.15 98.69
Opit 0.27 0.00 0.00 0.00 0.10 98.6 99.0 0.12 98.79
Kasagala 0.14 0.00 0.00 0.00 0.20 98.8 99.0 0.12 98.89
Kagombe 0.41 0.00 0.00 0.00 0.10 98.9 99.1 0.10 98.99
Rwoho 0.41 0.00 0.00 0.00 0.10 99.0 99.2 0.10 99.09
Rom 0.00 0.00 0.00 0.00 0.20 99.2 99.2 0.10 99.19
Mujuzi 0.00 0.00 0.00 0.00 0.20 99.4 99.2 0.10 99.30
Kijanabolola 0.14 0.00 0.00 0.00 0.10 99.5 99.2 0.07 99.35
Jubiya 0.54 0.00 0.00 0.00 0.00 99.5 99.4 0.07 99.45
Itwara 0.14 0.00 0.00 0.00 0.10 99.6 99.4 0.07 99.50
Igwe-Luvunya 0.14 0.00 0.00 0.00 0.10 99.7 99.4 0.07 99.55
Mpigi 0.14 0.00 0.00 0.00 0.10 99.8 99.5 0.07 99.65
Zika 0.00 0.00 0.00 0.48 0.00 99.8 99.6 0.06 99.70
Ogili 0.00 0.00 0.00 0.00 0.10 99.9 99.6 0.05 99.75
Kibeka 0.00 0.00 0.00 0.00 0.10 100.0 99.6 0.05 99.80
Kabuika-Muj' 0.27 0.00 0.00 0.00 0.00 100.0 99.7 0.03 99.85
Matiri 0.27 0.00 0.00 0.00 0.00 100.0 99.7 0.03 99.85
Kitechura 0.27 0.00 0.00 0.00 0.00 100.0 99.8 0.03 99.90
Nsowe 0.27 0.00 0.00 0.00 0.00 100.0 99.9 0.03 99.95
Bukaleba 0.14 0.00 0.00 0.00 0.00 100.0 99.9 0.02 99.95
Kadam 0.14 0.00 0.00 0.00 0.00 100.0 99.9 0.02 99.95
Lwala 0.14 0.00 0.00 0.00 0.00 100.0 99.9 0.02 99.95
Bwezigola-Gunga 0.14 0.00 0.00 0.00 0.00 100.0 100.0 0.02 100.00
* Table shows percentages of the total number of species known from the protected-area system. Columns headed "add %" show the percentage
added to the total by each site through the addition of species not already represented in sites higher on the table. Columns headed "total %"
show the cumulative percentage of species represented by sites at and above the level concerned.

Conservation Biology
Volume 14, No. 3, Julne2000
Table6. table for sites selected for Uganda'snationalnetworkof forest naturereserves.a
Complementarity

Butterflies Birds Mammals Moths Trees Fauna Combined

Qualifying (n = 736) (n =959) (n = 90) (n = 207) (n = 985) Trees average add total
Site groupb add % add % add % add % add % (total %) (total %) % %
Semliki
National Park 1 41.98 45.36 34.44 38.94 34.11 34.1 40.2 37.15 37.16
Kidepo Valley
National Park 1 no data 27.42 13.33 no data 9.04 43.1 50.4 9.61 46.77
Bwindi National
Park 1 10.33 7.82 20.00 17.31 11.68 54.8 64.2 12.77 59.51
Queen Elizabeth
National Park 1 no data 10.22 2.22 no data 6.60 61.4 67.3 4.85 64.36
Mt. Elgon
National Park 1 3.40 1.56 5.56 6.73 6.19 67.6 71.7 5.25 69.66
Rwenzori
National Park 1 1.63 0.31 8.89 2.40 2.13 69.7 75.0 2.72 72.37
Kibale National
Park 1 4.76 0.42 no data 8.65 1.22 71.0 78.4 2.34 74.68
Murchison Falls
National Park 1 no data 1.98 no data no data 1.83 72.8 78.9 1.16 75.85
Lake Mburo
National Park 1 no data 0.63 1.11 no data 1.93 74.7 79.4 1.18 77.06
Budongo 2 6.25 1.15 2.22 10.10 7.21 81.9 84.3 6.07 83.11
Otzi 2 4.76 0.10 2.22 2.88 4.06 86.0 86.8 3.28 86.39
Mt. Moroto 2 4.21 1.77 3.33 4.81 1.73 87.7 90.3 2.63 89.01
Mt. Kei 3 2.17 0.52 1.11 1.92 0.81 88.5 91.7 1.12 90.11
Sesse Islands 3 1.63 0.00 1.11 0.48 1.42 89.9 92.5 1.11 91.22
Sango Bay 3 2.04 0.21 0.00 1.45 0.51 90.5 93.5 0.72 91.98
Kasyoha-Kitomi 3 2.72 0.10 0.00 0.00 0.71 91.2 94.2 0.71 92.68
Kalinzu-Mara' 3 1.63 0.00 10.9 0.00 0.30 91.5 94.9 0.49 93.19
Era 3 0.27 0.00 1.09 0.48 0.51 92.0 95.3 0.48 93.64
Labwor Hills 4 0.68 0.10 2.17 0.00 1.02 93.0 96.1 0.88 94.55
Nyangea-
Napore 4 1.36 0.00 0.00 0.00 0.81 93.8 96.4 0.58 95.10
Echuya 4 1.22 0.10 0.00 0.00 0.61 94.4 96.7 0.47 95.56
Mabira 4 1.36 0.21 0.00 0.97 0.10 94.5 97.4 0.37 95.96
Bugoma 4 1.49 0.00 0.00 0.97 0.00 94.5 98.0 0.31 96.26
Timu 5 0.95 0.00 0.00 0.00 0.71 95.2 98.2 0.47 96.71
Morongole 5 0.27 0.00 0.00 0.00 0.81 96.0 98.3 0.44 97.17
Agoro-Agu 5 0.14 0.00 0.00 0.00 0.51 96.5 98.3 0.27 97.42
Mt. Napak 5 0.27 0.00 0.00 0.48 0.30 96.9 98.5 0.25 97.68
West Bugwe 5 0.14 0.00 0.00 0.00 0.30 97.2 98.5 0.17 97.83
Mafuga 5 0.14 0.00 0.00 0.00 0.30 97.5 98.6 0.17 98.03
Rwoho 5 0.41 0.00 0.00 0.00 0.20 97.7 98.7 0.15 98.18
South Busoga 5 0.27 0.00 0.00 0.48 0.10 97.8 98.9 0.15 98.33
Mpigi 5 0.14 0.00 0.00 0.00 0.20 98.0 98.9 0.12 98.43
Kasagala 5 0.14 0.00 0.00 0.00 0.20 98.2 98.9 0.12 98.54
Rom 5 0.00 0.00 0.00 0.00 0.20 98.4 98.9 0.10 98.64
Kijanabolola 5 0.14 0.00 0.00 0.00 0.10 98.5 99.0 0.07 98.74
Jubiya 5 0.54 0.00 0.00 0.00 0.00 98.5 99.1 0.07 98.79
Igwe-Luvunya 5 0.14 0.00 0.00 0.00 0.10 98.6 99.1 0.07 98.84
Itwara 5 0.14 0.00 0.00 0.00 0.10 98.7 99.2 0.07 98.94
Mt. Kadam 5 0.14 0.00 0.00 0.00 0.00 98.7 99.2 0.02 98.94
Lwala 5 0.14 0.00 0.00 0.00 0.00 98.7 99.2 0.02 98.94
Ogili 5 0.00 0.00 0.00 0.00 0.10 98.8 99.2 0.05 98.99
Kitechura 5 0.27 0.00 0.00 0.00 0.00 98.8 99.3 0.03 99.04
Zulia 6 no data no data no data no data 0.00 98.8 99.3 0.00 99.04
Lokung 6 0.00 0.00 0.00 0.00 0.00 98.8 99.3 0.00 99.04
Taala 6 0.00 0.00 0.00 0.00 0.00 98.8 99.3 0.00 99.04
Kazooba 6 0.00 0.00 0.00 0.00 0.00 98.8 99.3 0.00 99.04
Wabisi-Wajala 6 0.00 0.00 0.00 0.00 0.00 98.8 99.3 0.00 99.04
Kyambogo 6 0.00 0.00 0.00 0.00 0.00 98.8 99.3 0.00 99.04
Zika 6 0.00 0.00 0.00 0.97 0.00 98.8 99.5 0.12 99.14
Totals 98.23 100.00 99.90 100.03 98.78 99.18
"Table shows percentages of the total number of species known from the protected-area system. Columns headed "add %" show the percentage
added to the total by each site through the addition of species not already represented in sites higher on thvetable. Columns headed "total %"
showJthe cumulative percentage of species represented by sites at and above the level concerned.
bSites were added in groups, with members of each group added in order of decreasing species contribution. group 1, national parks; group 2,
site contributes >2% of species complement of national protected-area system; group 3, within any taxon, >1% of species unique to forest;
group 4, site contributes 1-2% of species complement of national protected-area system; group 5, site supports at least one unique species of con-
servation significance and/or, within any taxon, 0.5-1% of species unique to forest; group 6, site supports vegetation type not otheruwiserepre-
sented in protected-area system.
Howardet al. PlanningForestNatureReservesin Uganda 871

Table 7. List of Ugandan forests selected for nature reserve (NR) establishment, criteria used, and areas to be designated.

Area Selection criteria" NR NR NRarea Slope>15? NP/WR Bufferarea

Site categorya Forest (km2) 1 2 3 4 5 6 suitabilityc (%) (ki2) (ki2) statusd (ki2)e
Prime Semliki 219 + + + 3 35 100 0 219 119
Bwindi 321 + + + 2 30 100 207 321 221
Mt. Elgon 1192 + + + 2 30 358 569 1192 834
Rwenzori 996 + + + 3 35 349 818 996 647
Kibale 558 + + 2 30 167 223 558 391
Budongo 825 + + + 2 30 248 27 248 335
Otzi 188 + + 3 35 100 107 0 50
Mt. Moroto 483 + + + + 3 35 169 258 0 85
Subtotal 4782 1590 2209 3534 2682
Core Mt. Kei 384 + + + + 2 25 96 1 0 48
Sesse Islands 43 + + + + 2 25 30 1 0 13
Kalinzu-Mara' 584 + + + 1 20 117 37 443 485
Sango Bay 151 + + + + 2 25 50 0 0 25
Era 74 + + + 3 30 50 14 0 25
Kasyoha-Kitomi 390 + + + + 2 25 98 162 0 65
Labwor Hills 437 + + + 3 30 131 233 0 102
Nyangea-Napore 417 + + 3 30 125 223 52 98
Echuya 35 + + 1 20 25 26 0 10
Bugoma 401 + 2 25 100 6 0 50
Mabira 300 + + + 1 20 60 6 0 30
Subtotal 3216 882 709 495 950
Secondary Agoro-Agtu 235 + + 3 20 47 136 0 89
Morongole 151 + + 3 20 30 129 63 99
Timu 118 + + 3 20 24 5 0 12
Mt. Kadam 399 + + 3 20 80 306 0 226
Mt. Napak 203 + + 3 20 41 103 0 62
South Busoga 163 + + 3 20 33 0 0 17
Rom 109 + 3 20 22 79 0 57
Kasagala 103 + + 3 20 21 0 0 10
Rwoho 90 + 2 15 20 51 0 31
Itwara 86 + 2 15 20 20 0 10
Kijanabolola 65 + 2 15 20 11 0 10
Lwala 59 + 3 20 20 46 0 26
Mafuga 38 + 1 10 4 35 0 0
West Bugwe 30 + 1 10 20 0 0 10
Igwe-LuvLnya 20 + 1 10 14 1 0 6
Mpigi group 261 + 1 10 26 4 0 13
Jubiya 36 + 2 15 20 0 0 10
Lokung 13 + 2 15 9 0 0 4
Zulia 1026 + 3 20 205 420 480
Ogili 53 + 3 20 20 43 0 23
Taala 92 + 1 10 20 6 0 10
Kazooba 74 + 2 15 20 4 0 10
Wabisi-Wajala 87 + 2 15 20 1 0 10
Kyambogo 89 + 2 15 20 0 10
Kitechura 53 + 2 15 20 10 0 10
Subtotal 3653 794 990 483 1246
Total 11651 3266 3908 4512 4878

aPrime sites allocated 30-35% nature reserve (based on nature reserve suitability), suibject to a minimum of 100 kM2;core sites allocated 20-
30% nature reserve, minimum 50 km2 but not to exceed 70% offorest; secondary sites allocated 10-20% nature reserve, minimum 20 km2 but
not to exceed 70% offorest.
bCriteria. 1, site contributes >2% of nationalprotected-area system species complement; 2, withginany taxon >1% of species unique toforest; 3,
site contributes 1-2% of national protected-area system complement; 4, within any taxon 0.5-1% of species uinique toforest; 5, site supports at
least one unique species of conservation significance (i.e., broadly endemic); 6, site supports vegetation type not otherwise represented in pro-
tected-area system.
cNature reserve suitability: 3, high (top 20 sites); 2, medium (middle 20 sites); 1, low (bottom 20 sites).
dArea (kin2) lying within a national park (NP) or wildlife reserve (WR).
eArea (kin2) to be designated as buffer zone so as to include all NPs, WRs,areas of >150?slope, and land bordering NRs.

Conservation Biology
Voluime 14, No. 3, Juine2000
872 PlanningForestNatureReservesin Uganda Howardet al.

Mt. Moroto Dry, northern

Labwor Hills and
mnountains
Nyangea-Napore Mt. Napak hill savannas
Mt. Kadam Morungule
Rom Timu

Agoro-Agu Lwala
Otzi Maruzi
Era Kapimpini
Kasagala Kamusenene
Kibeka Ogili
Wabisi-Wajala Lokung
Aswa River Zoka Moist
Kabuika savannas
Rwoho Kisangi
Kazooba Nsowe
Kijanablola
Kidepo Valley Mt Kei |
MurchisonFalls L. Mfburo

Kagombe Namwasa
Luunga Taala Moist
Bwezigolo-Gunga Itwara savanna/forest
Kasana-Kasambya Matiri mosaic

Mpanga Zika |
Mpigi _ Mid-elevation
South Busoga West Bugwe lakeshore
Igwe-Luvunya Jubiya forests
Sesse Islands Mujuzi
Bukaleba

Kalinzu-Mara Budongo
Kasyoha-Kitomi Bugoma Mid-elevation
Kibale Mabira
Sango Bay western
Semliki l forests

Queen Elizabeth|
Figure 3. Classification of 64 Ugan-
Mt. Rwenzori Echuya Highland dan protected areas included in our
Mafuga forests
study, based on TWJNSPANcluster
Mt. Elgon Bwindi analysis for birds.

forests, (5) mid-elevation western forests, and (6) high- only provide for the long-term protection of 99% of
land forests (Uganda Forest Department 1999). Within Uganda's species (Table 6) but should also ensure that
these major groupings, subgroups of up to about eight possible conflict over resource use within areas of dual
sites were generally distinguished, so for any given status is avoided (through formal recognition of pro-
taxon, sites were classified into 15-17 distinct sub- tected forest areas), and that steep slopes are given nec-
groups (example for birds provided in Fig. 3). essary protection throughout the forest estate.

Allocationof ManagementZoneswithinEachForest
Discussion
The government's broad policy objective can be satis-
fied by allocating 3,266 km2 (20.5% of the estate) to 44 Our work addresses a specific new government policy,
nature reserve sites with an additional 4,878 km2 of namely to protect 20% of Uganda's forest estate as na-
buffer zones (Table 7). These allocations should not ture reserves, to allocate 30%to low-impact uses, and to

Conservation Biology
Volume 14, No. 3, June 2000
Howard el al. Planning Forest Nature Reserves in Uganda 873

dedicate 50%to sustainable timber production. The 20- groups (unpublished data). We found, however, that
30-50% allocation was a political decision, to seek a bal- complementary sets of sites selected on the basis of
ance between a multitude of conflicting demands on the their bird or butterfly faunas included at least as many
country's forest resources. Although the case for more species overall as those selected on the basis of the
land to be committed to nature reserves was made and whole data set (Howard et al. 1998). Thus, in future site-
considered earlier in the process, our mandate at the selection exercises of this kind it should be possible to
time was to advise on implementation of the allocation, limit the number of indicator groups to just one or two,
so we do not consider the policy decision further here. thereby reducing site-assessment costs considerably.
The nature reserve system we describe is now being put An obvious shortcoming of our method is that it com-
in place. bines data from five biological indicator groups with a
There is no generally accepted way to design a pro- wide range of other criteria, and in so doing it is inevita-
tected-area network, and no two scientists or teams bly biased to some extent by value judgements. These
could be expected to produce identical recommenda- influence the selection of criteria, the weightings given
tions. This is because biodiversity hotspots for one to each, and the derivation of scores for certain site at-
taxon do not necessarily coincide with those of another tributes where data are unavailable (e.g., timber inven-
(Prendergast et al. 1993a; Dobson et al. 1997; Howard tory data) or difficult to assess quantitatively(e.g., tourism
et al. 1998) and because the biological characteristics of potential). Nevertheless, this bias is at least transparent
different groups dictate radically different area and habi- and to some extent consensual, and we believe our
tat requirements for maintenance of viable populations. methods demonstrate a pragmatic application of conser-
Our approach attempts to provide for the conservation vation science to a real situation and may provide a use-
of all of Uganda's biodiversity by including a few rela- ful model for adaptation elsewhere.
tively large reserves-(particularly important for the A network of sites selected through the standard pro-
larger animals and trees)-and a larger number of com- cedure of complementarity analysis (using number of
plementary smaller ones. The larger parks and reserves species added as the selection criterion) is not necessar-
are essential because only they can maintain viable pop- ily optimal from an economic, social, or political per-
ulations of certain keystone species such as elephants spective and does not necessarily lead to optimal land-
and the full complexity of ecological interaction and use decisions. It may be better to protect the country's
phenomena such as migration. The smaller reserves are biodiversity in a larger number of sites, if these are areas
essential because they provide the only opportunity to with low potential for other uses and where protection
protect examples of habitats and species that do not oc- would provide additional complementary benefits such
cur in the larger sites. as watershed protection. Our analysis illustrates how
Our focus has been the design of a protected-area sys- such considerations can be built into complementarity
tem that is fully representative of Uganda's biodiversity, analysis using relevant criteria to add sites to the list in-
even though much of that biodiversity is widespread stead of relying entirely on the number of species added
elsewhere and may be more efficiently protected in by each new site.
other countries. Uganda's location in eastcentral Africa A surprising result of the biodiversity inventory was
means that the country lies in a zone of overlap between the large number of sites (52 of the 64 investigated) sup-
ecological regions that extend well beyond its borders porting at least one species not recorded elsewhere. Al-
(White 1983). Ideally it would be desirable to undertake though we expect that further sampling would have re-
the type of inventory and analysis that we have done on duced this number, we preferredto adopt a "precautionary
a continent-wide basis. This would inevitably show that principle" (Myers 1993), ensuring that each species was
many of the reserves selected by our analysis for Uganda represented at least once in the country's protected-area
are made (at least theoretically) redundant by reserves system. This necessitated selection of a rather large num-
elsewhere. ber of (relatively) small sites. Although we might other-
The program involved more than 100 person-years of wise have selected fewer larger sites, recent evidence
work and cost something in excess of US$1 million. This suggests that our approach may be optimal in terms of
makes it unlikely that similar programs will be under- maximizing species number per unit area (Lomolino
taken elsewhere in the tropics without strong justifica- 1994).
tion (but see Balmford & Gaston 1999). It is therefore The planning we describe is an importantstage in an on-
important to consider whether the program was worth- going process. It provides a strong teclhical basis and co-
while and what lessons can be drawn from this work to herent national framework for adapting present manage-
develop more cost-effective site-selection procedures. ment practices and defining management zones within
Much work remains to be done, but we found that a each forest. The process is now being taken to the local
minimum complementary set of sites selected to include level, and detailed site-managementstrategiesare being de-
all vegetation types (Langdale-Brownet al. 1964) would veloped in consultationwith local communities. The aim is
account for only 81.6% of species from our indicator to establish appropriate management zones within each

Conservation Biology
Volume 14, No. 3, June 2000
874 PlanningForestNatureReservesin Uganda Howardet al.

forest so that biodiversity is protected and the legitimate Johns, A. G. 1997. Timber production and biodiversity conservation in
demands of local people are satisfied while also provid- tropical rain forests. Cambridge University Press, Cambridge,
United Kingdom.
ing for the sustained production of timber and other Johnson, N. C. 1995. Biodiversity in the balance: approaches to setting
products. geographic conservation priorities. Biodiversity Support Program,
The new nature reserves will be crucial to conserving Washington, D.C.
Uganda's forest biodiversity, but their protection must Langdale-Brown,I., H. A. Osmaston, and J. G. Wilson. 1964. The vege-
be undertaken alongside other complementary conser- tation of Uganda and its bearing on land-use. Government Printer,
Entebbe, Uganda.
vation measures. Adjacent areas of the same forests that Lombard,A. T., A. 0. Nicholls, and P. V. August. 1995. Where should
are used for timber production or other uses will also nature reserves be located in South Africa?A snake's perspective.
play an important role in biodiversity conservation Conservation Biology 9:363-372.
(Johns 1997). Management techniques that enhance Lomolino, M. V. 1994. An evaluation of alternative strategies for build-
this role, such as judicious use of girth limits in selective ing networks of nature reserves. BiologicalConservation69:243-249.
Lowry, A., and T. P. Donahue. 1994. Parks, politics and pluralism: the
logging systems and retention of seed trees and strips of demise of national parks in Togo. Society and Natural Resources 7:
riverine forest at the time of felling, must also be applied. 321-329.
Others have discussed these techniques, and they are Margules, C. R., A. 0. Nicholls, and R. L. Pressey. 1988. Selecting net-
beyond the scope of this paper. Nevertheless it is impor- works of reserves to maximize biological diversity. Biological Con-
tant to recognize the synergy between these different as- servation 43:63-76.
Margules, C. R., R. L. Pressey, and A. 0. Nicholls. 1994. A scientific ba-
pects of forest management. sis for establishing networks of protected areas. Pages 327-350 in
P. L. Forey, C. J. Humphries, and R. I. Vane-Wright,editors. System-
atics and conservation evaluation. Clarendon Press, Oxford, United
Kingdom.
Acknowledgments McNeely, J. A., J. Harrison, and P. Dingwall, editors. 1994. Protecting
nature: regional reviews of protected areas. World Conservation
We thank members of the Uganda Forest Department Union, Gland, Switzerland.
Myers, N. 1993. Biodiversity and the precautionary principle. Ambio
field inventory teams and R. Badaza, D. Duli, D. Hafashi- 22:74-79.
mana, I. Kapalaga, A. Katende, R. Kityo, and R. Na- Noss, R. 1994. Cows and conservation biology. Conservation Biology
banyumya, who managed the fieldwork and assisted 8:613-616.
with specimen identification and data collation. E. Olet, Plumptre, A. J. 1996. Changes following sixty years of selective timber
A. Rodgers, R. Murtland, and A. Finch provided crucial harvesting in the Budongo forest reserve, Uganda. Forest Ecology
and Management 89:101-113.
technical and administrative support. We thank E. Diner- Pomeroy, D. 1993. Centres of high biodiversity in Africa. Conservation
stein, J. Gronow, D. Moyer, W. Hawthorne, A. Rodgers, Biology 7:901-907.
and four anonymous referees for helpful comments on Prendergast, J. R., R. M. Quinn, J. H. Lawton, B. C. Eversham, and D.
earlier versions of the manuscript. Most of this work was W. Gibbons. 1993a. Rare species, the coincidence of diversity
funded by the European Union (Project 6100.31. hotspots and conservation strategies. Nature 365:235-237.
Prendergast, J. R., S. N. Wood, J. H. Lawton, and B. C. Eversham.
42.015), the Government of Uganda, and the Global En- 1993b. Correcting for variation in recording effort in analyses of di-
vironmental Facility (Project UNO/RAF/006/GEF). P. V. versity hotspots. Biodiversity Letters 1:39-53.
was supported by the Italian government and J. S. L. by Pressey, R. L., C. J. Humphries, C. R. Margules, R. I. Vane-Wright,and
the United Kingdom's Darwin Initiative. P. H. Williams. 1993. Beyond opportunism: key principles for sys-
tematic reserve selection. TREE8:124-128.
Pressey, R. L., I. R. Johnson, and P. D. Wilson. 1994. Shades of irre-
placeability: towards a measure of the contribution of sites to a res-
LiteratureCited ervation goal. Biodiversity and Conservation 3:242-262.
Scott, J. M., et al. 1993. Gap analysis: a geographic approach to protec-
Balmford, A., and K. J. Gaston. 1999. Why biodiversity surveys are tion of biological diversity. Wildlife Monographs 123.
good value. Nature 398:204-205. Struhsaker,T. T. 1981. Forest and primate conservation in East Africa.
Dobson, A. P., J. P. Rodriguez, W. M. Roberts, and D. S. Wilcove. 1997. AfricanJournal of Ecology 19:99-114.
Geographic distribution of endangered species in the United Struhsaker,T. T. 1987. Forestry issues and conservation in Uganda. Bi-
States. Science 275:550-553. ological Conservation 39:209-234.
Hamilton, A. C. 1974. Distribution patterns of forest trees in Uganda Struhsaker, T. T. 1997. Ecology of an African rainforest: logging in
and their historical significance. Vegetatio 29:21-35. Kibale and the conflict between conservation and exploitation.
Hamilton, A. C. 1984. Deforestation in Uganda. Oxford University University Press of Florida, Gainesville.
Press, Nairobi, Kenya. Stuart, S. N., R. J. Adams, and M. D. Jenkins. 1990. Biodiversity in sub-
Howard, P. C. 1991. Nature conservation in Uganda's tropical forest re- SaharanAfrica and its islands: conservation, management and sus-
serves. World Conservation Union, Gland, Switzerland. tainable use. Occasional paper no. 6. Species SurvivalCommission.
Howard, P. C., and T. R. B. Davenport, editors. 1996. Forest biodiver- World Conservation Union, Gland, Switzerland.
sity reports. Volumes 1-33. Uganda Forest Department, Kampala. Uganda Forest Department. 1999. Forestry nature conservation master
Howard, P. C., P. Viskanic, T. R. B. Davenport, F. W. Kigenyi, M. Balt- plan. Kampala,Uganda.
zer, C. J. Dickinson, J. S. Lwanga, R. A. Matthews, and A. Balmford. Vane-Wright,R. I., C. J. Humphries, and P. H. Williams. 1991. What to
1998. Complementarity and the use of indicator groups for reserve protect? Systematics and the agony of choice. Biological Conserva-
selection in Uganda. Nature 394:472-475. tion 55:234-254.

Conservation Biology
Volume 14, No. 3, June 2000
Howard et at. Planning Forest Nature Reserves in Uganda 875

Wass, P., editor. 1995. Kenya's indigenous forests: status, management 20. United Nations Educational, Scientific, and Cultural Organiza-
and conservation. World Conservation Union, Gland, Switzerland. tion, Paris.
World Conservation MonitoringCentre (WCMC). 1992. Global biodiver- World Conservation Union. 1992. Protected areas of the world: a re-
sity: status of the Earth'sliving resources. Chapman & Hall, London. view of national systems. Volume 3. Afrotropical. Gland, Switzer-
White, F. 1983. The vegetation of Africa. Natural resources research land.

7'4

Conservation Biology
Volume 14, No. 3, June 2000