|UFDC Home||myUFDC Home | Help ||
|Table of Contents|
|Bigo Bya Mugenyi|
|The Bakopi in the Kingdom of Buganda...|
|Continuity through change in the...|
|The vegetation of Southwest...|
|Territorial behaviour and population...|
|Index to Volume 33 (1969)|
ALL VOLUMES CITATION THUMBNAILS DOWNLOADS PAGE IMAGE ZOOMABLE
STANDARD VIEW MARC VIEW
This item has the following downloads:
|Table of Contents|
Front Cover 1
Front Cover 2
Table of Contents
Bigo Bya Mugenyi
The Bakopi in the Kingdom of Buganda 1900-1912
Continuity through change in the social history of Kibuli
The vegetation of Southwest Kigezi
Territorial behaviour and population in the grey-backed fiscal shrike
Index to Volume 33 (1969)
Back Cover 1
Back Cover 2
THE UGANDA SOCIETY,
Price Shs:. ISI- -(Uganda)
THEV BAROPI IN THE ~ i KI oio-o BUGANDA-
1900-1912 7 J... ATA
CONTI ~RrTaoUGYI CHANGE IN THE
SOCIAL HI MRouV orIau R. M. Somean,
Tim VE-GPTATION. OF:86TJTHWS rAz Aiw
TERRrropjAL likHAVOt% AND, POPULAfON IN
THERLY-B"LCKED FACAL 'SHRIKE- W' R.-\\
Binor HANNINGTON: A, MISCELLANY T .B
1TiV UGANDA MUTINYWMDA-L OF. Rv. F. RwnoH B. Ta
A No-rF ON TIax Foon HABIT OF. THE
GIRAFE IN, EAsr AFRICA z R_ L' row
Averoom:Myth, Memoir and Moral Adonition J i
S. W. KnonvaA D. K. MARPAppIAa7-
B. I MULYATv D. K. MARPHAkfiv
Ahitoyg#anor-Kitara (By-A. R. Dunbar) G. N ztw
.- he tethi der(ebral Orti in der 1Kulturland
whaftthchen Entw klwung Bugandas (By G. Kadc) M. W~Ayw
Y Cnk Sxqslatl (y Go"chutidt) -7 :J VEATH0P
NOTES ON CONITRIBUTIORS.- ------
UGANAnn BrMaunca0ni 1969- Compiled by -B. W. LANotAike
ExkeOllncy Ahe Presiderit of Uganda, Dr. A, Mlmton Otote,
rau Profmsor B. W. Langlandfis
'h rsdent Professor HK El-Abd
TeVice-Preident Mrs. D. M. Etooni
The Hun, Secretary Mr. S, C. Griniley, xR.E..
The H~on, Treasurer Mr. W. S. Kajubi
_4e Hon, Editor' Mr` F. X. Katete
The Hon. Librarian Mr. D. Kavulu
:Dr. W, B, Bantage Mr. W. A. R. Mc Crae
Mr, H. J. Bevin Mr. J. P. Ocittu
Mr. J, E. Compo Mr. H. Sassoon
Lr L, Dxon_ Dr. W. A. van Eck
Hon. S&ereya Mr. M. Kaggwa
Hon. 'I'll",r;Mrs. J. Bevmn
Hfon. Editor, Professor B. W. Langlands
non. Librarian: Mrs. F. Wapenyi
Hon. Audtos Hon. Legal-Adviser:
MesrsCooper Brothers & Co. Mr. R. A. Couni'hark
Hon. Vzce-Presidena t Hon. Lfe Members:
Trito Winyi Gafabusa c.B.E. Mr. E. B. Haddon
Reeed Father J. P. Crazzolara Mr. H. B. Thomas, ouE.
C6piin C, R. W. Pitmnan, C.B.E.R.., w Mc.. Professor A- W. Williams, c.ux.e
,Sir A, R. Cook, O.M.a., O.B.E. 1950N-51 Rev. Fr. F. B. Gaffney, w. F.
E. J. Wayland, CB.Ex. 1951-52 Professor A. W. Williams, c-.B.E
H4. fl. -Hunter, C.B.E., L.L.D. 1952-58 Sir J. N. Hutchinson, c.U.P., eFx~s.
SH. Jowitt, 0.4.r.. 1953-54 Mr. J, DM Jameson. Ox.BE
Sir ff. R. Hone, KG.0,MG., K"D.C., 1,954-55 Dr. Audrey 1. Richards, c.B.E.
M~e., go,1955--56 Rev. Dr. IL C. Trowell, o>w.s.,
SMr. SykC6 o.B.E. 19,%-57 Mr. D. K. Marphatia, i.B.E.
WMr N. V. nranett 1957,%8 Mr. AC Barringon Ward
SCaptxin C- R. S. Pitman, c.BE. 195859 Dr. H. F Mori
ID.s.o., M.C. 1959L-4 Professor A W. fiouthall
SMr. SL Kulubiya, C.B.E. 196 1 Nfi. I. a, ). 4WiranCCe O'sr..
SMr. &. A. Temple Perkins 1961-62 Mr.B E., R. Kirwan, mB.B.E
(rMr. R. A. SnwmHl 19624-4 Mr. W. &. K~ajbi
$'Or, K- A. Davies, cx.uo., OXB.E. 1964-65 Dr. M. Posnansy
14Mr. G, H. H. Hlopkirus, OB.E. 1965- Dr. W. B. Bxg
7 Amrs K. M. Trowell w.B.E. 196"47 Profesow & J. Baker, mO.BE
B Dr4 W. I., Eggeling 1967-8 Dr. M. &MA Kiwanuka
CayunaMr. E, Kironde Rev. ILH. Sharpe
swcetary:, Mrs. J, Bevnn
THE JOURNAL OF THE UGANDA SOCIETY
THE UGANDA SOCIETY
Including a Uganda Bibliography published in
association with the Milton Obote Foundation
Copyright: Uganda Society 1969
BIGO BYA MUGENYI- - - - - M. POSNANSKY 125
THE BAKOPI IN THE KINGDOM OF BUGANDA
1900-1912 - - - - - --J.A.ATANDA 151
CONTINUITY THROUGH CHANGE IN THE
SOCIAL HISTORY OF KIBULI - - R. M. SOLZBACHER 163
THE VEGETATION OF SOUTHWEST KIGEZI - A. HAMILTON 175
TERRITORIAL BEHAVIOUR AND POPULATION IN
THE GREY-BACKED FISCAL SHRIKE - - W. B. BANAGE 201
BISHOP HANNINGTON: A MISCELLANY - H. B. THOMAS 209
THE UGANDA MUTINY MEDAL OF REV. F. ROILING H. B. THOMAS 213
A NOTE ON THE FOOD HABITS OF THE
GIRAFFE IN EAST AFRICA - R. L. BRAHMACHARY 214
ADDENDUM: Myth, Memoir and Moral Admonition J. A. ROWE 217
S. W. KULUBYA - - - D. K. MARPHATIA 220
B. K. MULYANTI - - - D. K. MARPHATIA 220
A history of Bunyoro-Kitara (By A. R. Dunbar) G. N. UZOIGWE 221
Die Stellung der Zentral Orte in der Kulturland
schaftlichen Entwicklung Bugandas (By G. Kade) M. WINGENS 224
Kambuya's cattle (By Goldschmidt) - - J. WEATHERBY 227
NOTES ON CONTRIBUTORS - - - - - 220
UGANDA BIBLIOGRAPHY 1969-Compiled by B. W. LANGLANDS 229
INDEX TO VOLUME 33
Uganda Journal, 33, 2 (1969) pp. 125-150.
BIGO BYA MUGENYI
By M. POSNANSKY
The earthworks at Bigo bya Mugenyi (0 10'N, 31 15'E) on the south
bank of the Katonga river were described and figured by E. J. Wayland in
the Uganda Journal for 1934, following their first description in 1909 by D. L.
Baines, the Masaka District Collector.1 Though the first plan of the site was
produced by Major C. R. Hall in 1909 it was not until 1921 that an adequate
plan of Bigo was prepared by A. D. Combe on behalf of the Uganda Protectorate
Geological Survey. Though only a compass traverse it needed little alteration
in 1960 when more precise surveying equipment was employed. During the
1950's E. C. Lanning and the Rev. Gervase Mathew indicated the significance
of the site for an understanding of the Iron Age of Uganda and in 1957 P. L.
Shinnie conducted excavations to "throw some light on the historical problems"
connected with the Bacwezi legends and "to reveal the material culture of the
builders of the earthworks". Further excavations were carried out from 19
August to 10 October 1960, in order to make as complete a survey as possible
of the site, to establish the cultural sequence and to obtain charcoal samples
for Carbon 14 dating. The earthworks are known from traditional sources and
are the subject of many articles, the most important of which are listed in the
bibliography. I have dealt at length elsewhere with the connections between
Bigo and the Bacwezi legends (Posnansky 1966 and 1969) and the purpose of
this paper is to describe the 1960 excavations, to summarize the traditional
evidence and to assess the historical significance of Uganda's largest and most
important ancient monument.
The Site (figure i)
The earthworks are situated on the south bank of the Katonga river,
seven kilometres upstream of the confluence with the Nabakazi river in Mawo-
gola county of the former Kingdom of Buganda. A track, a little over 16
kilometres long, runs from the main Masaka-Mubende road on a ridge to
the east of the Kakinga valley in which several dams are presently situated.
Bigo itself is at the southern apex of a bend of the river and is flanked on either
side by the smaller earthworks of Kagogo, three kilometres to the west, and
Kasonko, five kilometres to the east, described by Wayland (1934) and Lanning
The earthworks consist of a linear ditch running in an arc from the swamp
of the Kakinga to the Katonga. On the eastern side the outer trench bifur-
cates to enclose a hilly area over 4,000 feet high. In its two branches the outer
ditch system is over six kilometres long (3.75 miles). The earthworks include
within them a little under two square miles2 of undulating country, the western
part of which is low-lying and comprises a dry stream bed of a tributary of
the Kakinga. On high ground in the centre (figure 2) are situated a series
of ditches forming two lobate enclosures (Nos. 1 and 2) with smaller enclos-
sures on their eastern and southeastern sides. From the northwest corner of this
complex a ditch runs northeastwards to the Katonga river. Altogether there
are 10 kilometres (6.11 miles) of ditches. The outer ditch is irregular and has
126 UGANDA JOURNAL
at least 12 breaks in its inner circuit some of which are 20 to 30 metres
wide. The eastern bifurcation is even more irregular and has at least eight
openings and at one point is of negligible depth. The ditch in parts has a low
bank on the inner side though the bank is discontinuous and nowhere more
than two metres high. It would appear that the ditch was the important feature.
Its upcast was in many places thrown out in a haphazard manner, without
any attempt being made to construct a regular bank. Irregularities in the ditch
have been interpreted by Combe (1922) as being due to the need to avoid
masses of hard quartz. The ditch varies greatly in both width and depth, in
places it is at present over 10 metres wide and is between one and a half and
four metres deep. The whole of the ditch system is heavily overgrown and from
the form of the ditch faces where rock can still be seen obtruding it is prob-
able that the ditches were filled fairly rapidly with dense foliage and the sides
There is a ford across the Katonga river between the two outer ditches on
the eastern side. The river is now choked with papyrus and there is very little
perceptible flow of water. The ditches, except on the southeast of the outer
bifurcation, do not command the highest available ground and in most places
can be looked down upon, nevertheless from various parts of the central
enclosure system most of the outer ditches are visible.
The central enclosure system (Figure 2) is far more regular. The deepest
ditches are mostly on the southern and eastern sides and, in the case of Enclo-
sures 1 and 3, the inner banks are fairly continuous and in several cases quite
pronounced. At several places in Enclosure 3 the height from the bottom of
the ditch to the top of the inner bank, before excavation, was found to be over
seven metres, though it was normally of the order of four to five metres. Near
the entrances to Enclosures 1 and 3 outer banks were also found whilst En-
closure 4 has an outer bank on its southern half and no inner bank. Enclosures
1, 2, and 4 all have two entrances to the outside whilst 3 has a single entrance.
Three prominent mounds are situated within or around the central enclosure
area. Movnd III (Site C), a little over three metres high, lies outside the western
ditch of Enclosure 2 and the ditch is markedly indented to go around it. Mound
II (Mound C of Shinnie) is slightly arcuate and tapers to a point on the north-
east and at its maximum height is also a little over three metres whilst Mound
I (Mound A of Shinnie) is round, fairly regular and some two metres high.
Mound I lies to the east side of a break over 30 metres wide in the ditch separa-
ting Enclosures 1 and 2, whilst Mound II lies opposite this break some eighty
metres away from the ditch. Mound II commands a view over the outer ditches
and lies on a line towards the southwest facing both the major break in the
ditch of Enclosure I and the largest break in the outer ditch.
Shinnie described a further mound, Mound B, which was found after a
detailed contour survey (Figure 3), to be the inner bank of the ditch dividing
Enclosures 1 and 2. The earth of Mound I had spilled over onto it so that it
forms an extension of that mound to the west. Survey further demonstrated
that Mound I was remarkably flat and that a flat area existed to the east of
Mound II and that at one time Mound II probably had an arm enclosing this
flat area on its northern side. The complete absence of an embankment
for the ditch between Enclosures 1 and 2 east of the mound area, suggests
that the upcast from its digging may have been used for the erection
of Mounds I and II for which there is no sign of a surrounding ditch.
Some 40 metres to the south of Mound I in Enclosure I a series of irregular
.. I ...... .. : MAP 3 DITCH SYSTEM AT BIGO
3MUNSA .I "
C.i / KA............ C '.
\ "'M P I ''" I KATONGA RIVER
HA ..... .........:-.. ....
,.,.. "-.G -.0 .
... .... '" : E\ ... (, .... .-
-W ... ..
1. S::::::::::: ONK.O GO/,., MB-. A" .- ,
; ............. ^ ii i % i)i-(A-_^ -' "....- ..'.".
:: :::::::::: ::I:::::::::::::COR CTIO S
..-... .A : . : T
I C 40 o A0WE 10 1 S (b.. . .
KARA Wyon pr. .~eASEND KILOMETRES I
_- ,_~1__ I,
31., 1 N
Figure 2. Plan of central area of earthworks.
squarish mounds were located between three quarters and one and a half
metres high and some twenty metres across. Near them several grindstones
were found on the surface. They are in form entirely different from any of the
other features at Bigo and were interpreted as belonging to various habitations
of people said to have lived temporarily in the area during the 1930's by
lusufu Mukasa, a former Muluka chief of Bigo. Other significant features in
the area are two holes 120 metres east of the inner of the two outer ditches.
The largest is irregular in shape, about three metres across and some five to
six metres deep whilst the second is a little over a metre away and smaller and
shallower. It is impossible to determine their purpose, Gray (1935) considered
them waterholes and Mathew (1952) mine shafts, though as they scarcely
go deeper than the ditches themselves the latter view on present evidence seems
There are many game pits near the main breaks in the outer bank and
ditch (Figure 4, site B). These all appear to be of comparatively recent date
and were still being used in the 1950's before the game in the area was exter-
minated in an attempt to control the spread of tsetseflies.
Excavations were undertaken in four areas. In the central area (Site A,
figure 3) a series of 16 trenches were dug across Mounds I and II and the flat
area to the west of Mound II in order to establish the nature of these mounds
and the structural relationship between them. At the outer ditch (Site B,
Figure 4) a trench was dug across the bank and ditch and two trenches across
one of the widest breaks in the ditch system to ascertain if any structures
existed within the entrances to Bigo. A cutting was made across Mound III
(Site C, figure 4) to determine the structural relationship between the mound
and the Enclosure 2 ditch. At the ditch intersection, where Enclosures 1, 2
and 3 adjoin (Site D, Figure 4) three trenches were dug to indicate the rela-
tionship of the outer enclosures to the two main enclosures. In all, 781 feet3
(238 metres) of trenches, either three or five feet wide, were dug and, except for
three on the top of Mound I, all were taken down to the 'natural' undis-
turbed bedrock of decomposed gneiss or to the overlying subsoil.
Site A (Figure 3)
Shinnie's excavation4 had indicated a series of postholes on the buried
ground surface beneath Mound I. He also demonstrated that Mound I post-
dated the ditch between Enclosures 1 and 2 and that it had been filled in, partly
by erosion of material from Mound 1. The ditch had originally been 4 metres
deep and was only 60 cms. wide at the bottom.
The trench across Mound II (Trench 3, Figure 3 and section AB, Figure 5)
indicated that there was no ditch on either side of the mound and that the
mound soil had been thrown up from the east, which supports Combe's original
contention that the earth came from the ditch between Enclosures 1 and 2
where there is now no bank. A series of trenches east of Trench 3 (Section B-C)
revealed an old ground surface overlain by material similar to the make-up
of Mound II. This suggests that the level area to the east of Mound II was
built up artificially. Mound II is composed of compact broken subsoil and
bedrock, red or pink in colour, without any humus inclusions, indicating that
the material was taken directly from the ditch and not extracted from an
already existing bank. In the section exposed in Trench 5 a thin earth parting
130 UGANDA JOURNAL
between the material of the level area and the mound soil was revealed, demon-
strating that the area was levelled before the mound was thrown up.
On the original ground surface several features were noted. In Trenches 8
and 9 narrow ditches and a posthole were cut through the four to six inches
(10-15 cm.) of humus and the stony layer beneath, but were not dug into
the underlying subsoil. In Trench 4 two querns were found. Trench 5 revealed
a depression, or part of a ditch, some two and half feet (76 cm.) deep, filled
with a stony subsoil in which pottery was found throughout. Within a small
oval-shaped hollow, three feet by two, at the bottom of this depression a
cluster of stones, burnt bone and some pottery were found (Figure 5, 3).
None of the bone was diagnostic so that it is impossible to determine whether
the finds represent a burial. The pottery is crude, unlike that of any other Bigo
ware. Some of the pottery within the buried soil has a weathered surface.
Trenches 9, 10 and those on the top of Mound I revealed that the earth
of which Mound I is composed is different from that of Mound II. The mound
soil was less compact, had charcoal flecks and occasional lenses of dried dung and,
in contrast to the mound soil of Mound II, contained pottery, the occurrence
of which was also noted by Shinnie. It was apparent that the mound soil of
Mound I was not directly derived from a ditch dug into the subsoil and
weathered rock but had come from a weathered surface containing dung,
pottery and soil with charcoal.
From the shape of the level area east of Mound II it would appear that
Mound II was formerly crescentic in shape. If this was so, it is probable that
the earth for Mound I was in fact obtained from the partial destruction of
Mound II. The tailing off of Mound I to the north also suggests that the mater-
ial for it was derived from that direction. The existing part of Mound II is
some five feet (1.52 metres) higher than Mound I and though Mound I has
a far greater surface area than Mound II, the amount of earth in it is not in-
consistent with its being erected from a former eastern horn of Mound II.
Mound II may partly lie above an inner bank of the ditch between Enclosures
1 and 2, since the western extension of Mound I, called Mound B by Shinnie,
is clearly indicated from the contours (Figure 3) as being a ditch bank and not
a separate mound.
A trench was dug along the line of the ditch as projected from Shinnie's
two trenches but thirty feet to the east. No ditch was encountered which means
that the ditch ends between the two trenches. The sequence of event s indicated
by the excavations at Site A is, firstly a series of structures, some large as indi-
cated by postholes found by Shinnie under Mound I, others less substantial
as those found on the old ground surface beneath the level area and Mound II.
This can be termed Bigo Phase I. Later the ditch between Enclosures 1 and 2
was dug and the freshly dug material was used to level the area to the east of
the present Mound II. Immediately after, the crescentic mound, of which
Mound II is the remnant, was erected. Some time later, by which time silting
had taken place in the ditch, the flat-topped Mound I was thrown up making
use of material from the northeastern part of Mound II. Before the mound
surface had been stabilized by vegetation, pink mound earth had washed into
the ditch. Above this pink earth Shinnie found a darker soil with animal bones
and pottery which he interpreted as midden rubbish.5
Three trenches were dug across one of the low mounds to the south of the
ditch which Shinnie (1960, p. 18) interpreted as a possible source for the broken
pottery and bone in the ditch. The mounds were composed nearly entirely of
SITE A ,
*'' n ^ --'' -~' ~ ~- ^ '
/ / / T R CH S I
I / / I
it / / ,,.^ '
\ 1 .. -... .
\ "- : --- -'- \ .-- "' --
/ / \ \\
, \ \ i i
S , : -MOUND I /
S \ \fS 1 o
\ N \ I' \ /
'S > \ K \ i,., /S
I \ \ N
S-- -a- SECTIONS-
"' '' \ ~ ""
S- 50 -- E --
FI. '3. ll ) s g 1 7 a 1
...... ...... ..... -" t EC
I T Ci I:''" ,, --, w
"" ".\ -'" '-,\ ,
5 i -0 0 FE- t
\ S I -I S
T C -19 excavations
--- -'- 0 EET - -S
, TE \ '
I ,, '
aN'I ; ,,
a / ,a
/ f i
[TTTrrrTr T TTrr
0 C' H'
ILuJ.i.i U uli
,y' ;Ill I\
i r' I ii
- - 1" C
.. : ^ 0 I T H.
A SITE 0
B A N K
\/ff I r
Figure 4. Plan of sites B, C and D.
soft grey soil, clearly derived from cattle dung and midden rubbish. The pottery,
however, was largely undecorated and dissimilar from that in the ditch. The
mounds probably represent old cattle enclosures associated with houses,
the remains of the structures of which were not located.
Site B (Figure 4)
The trench dug across the outer ditch and bank (Plate 1) showed that the
ditch, which is 24 feet wide (7.32 metres) at the lip, was originally V-shaped
(Figure 5, Section A-B). There is now over five and a half feet (1.67 metres)
of natural infill. The bank is approximately five and a half feet high above
a buried ground surface. Within the primary ditch infill an almost complete pot
(Figure 8, 1) was found. As with most of the other entrances, the bank and
ditch on the two sides of the opening do not form a continuous line but provide
a slight overlap. A trench was dug across the area between the banks (Figure 5,
Section C-D) and a further one between the ditches. No structures were
located and it is evidence that if gates existed they were of the simplest
kind, being probably of thorn bush or other vegetation.
Site C (Figure 4)
The contour plan indicated that Mound III was of irregular form being
indented on the northern face. The section (Figure 5) showed that the mound
was overlain by earth upcast from a deepening of the ditch. This deepening
was evidenced by the cross-section of the ditch, a V form cutting through a
shallow U form. The mound is nearly ten feet high (3 metres) above a buried
ground surface on which pottery was found but no structures located. The
ditch varies in width from 20 feet (6.10 metres) on the eastern side to over 30
feet (9.15 metres) on the southeast and in its present state is nearly 10 feet
(3.05 metres) deep at its deepest. Excavation revealed over five feet (1.5 metres)
As the ditch goes around the mound, and some of its upcast overlies it,
it is evident that the mound predates the ditch. The time factor between the
two is uncertain but a turfline was found between the surface of the mound
and the the upcast of the ditch which suggests that the mound surface had
stabilized and vegetation grown on it before the main ditch was dug. The width
and depth of the ditch indicates that the source of the earth for the mound
came from a depression immediately to the east which was later deepened by
the excavation of the enclosure ditch.
Site D (Figure 4)
At Site D a large clearance operation was undertaken. In parts the ditch
was deeper than elsewhere at Bigo. A cutting was placed across the ditch
where it was nine feet (2.75 metres) deep. The bank on the northern side has
a fairly complex stratigraphy (Figure 5) indicating at least four building phases.
The sides of the ditch are steep,the lower portions having been dug through
weathered rock. An old ground surface occurs under the northern part
of the bank and to the south of the ditch, elsewhere it had been removed.
Some of the pottery from the old ground surface is the finest from the whole
of the Bigo sequence. There is nearly five feet (1.52 metres) of infill in the ditch.
The stratification of the bank would suggest that the ditch has been deepened
on several different occasions and the amount of infill, compared to the depth
of the ditch, was far less than at Site B. Two trenches were laid out at one of
5 2 2 d we rewr
ftSITE CCH r/ U/A r
COARSER ,NFILL >
91(HoED Qclt c ShTE B4cl
0 10 25 IS tp., Feet
S0 s Dt s
SITE d SITE B
0 "* 'Iorrorrr w I 5 "**b" ^ ^Wss_____________^
the intersections in order to determine whether there was a time interval
between the digging of the two ditches. Excavation indicated that both
were of contemporaneous construction.
The excavations at all four sites indicate that there was settlement at Bigo
prior to the construction of the ditch system. This occupation is referred to
below as Bigo I, the later developments as Bigo II. The ditch system postdates
Mound III but is contemporaneous with Mound II. After a short time interval
Mound II was destroyed and Mound I built and the ditch system near it
used for dumping rubbish. There is evidence that the ditch system at Site D
was deepened at various times. Mound II is interpreted as being the ekyikari
enclosure at the centre of a royal village, or orurembo.
Four charcoal samples were submitted to Geochron Laboratories. They
were chosen as representing a structural sequence which would be of value in
working out the time intervals in the structural history as outlined above.
Sample GXO 516 was derived from charcoal from the old ground level beneath
Mound II. A date of AD 1505+70 (based upon the Libby half-life of 5570
years) was obtained, which the laboratory suggests might be slightly young,
due to the presence of a few rootlets which may have remained after cleaning.
Sample GXO 517, AD. 1450+300, was from the depression noted in Trench
5, the sample was very small and consequently the counting error large. GXO
519, the largest sample, was obtained from the make-up of Mound II and
could conceivably have been included at the time of the construction, provided
a date of AD. 1570+90. GXO 518, was from the lowest infill of the ditch at
Site D. The date obtained was less than 200 years at two standard deviations.
The date could indicate that the sample was contaminated or that the sample
had washed down from above or that the ditch had been deepened relatively rece-
ntly. One of the first two possibilities is preferred as an explanation. It is significant
that the three acceptable dates have means very close to one another. Though
it was expected that GXO 519 would have provided the most recent date,
the dates are all within one standard deviation of each other. The only valid
conclusion that can be drawn from so few dates is that there is very little time
range between Bigo I and the erection of Mound I and that the main period of
Bigo's history can be assigned to a period around 1290-1575 AD. This is
a conclusion that agrees fairly well with previous estimates based on oral
traditions in which Bigo is ascribed to the Bacwezi dynasty as forbears of
Uganda's hereditary monarchies.
The Finds 8
A total of 4,287 sherds9 were recovered of which 3,532 were plain body
sherds, 35 % being from Bigo I contexts. The largest quantity of pottery came
from Site A (70%), though 713 sherds came from two trenches of Site B.
A total of 755 decorated and/or rim sherds were examined to determine the
pot form and decoration. In addition to reconstructable pots, 11 bases of other
pots were found. Of this collection more than 85% bore rouletted decoration,
normally in bands below or on the rim, executed in almost all cases with a
knotted grass roulette which is the most distinctive feature of Bigo ware. The
pottery can be divided into an abundant coarse ware and a rather fine ware.10
Variations on nine main forms of coarse ware were found, most of which are
136 UGANDA JOURNAL
illustrated in Figure 6 or in Shinnie (1960, Figures 10,12 and 13). The main
Type 1, large spherical pots between 30-40 cm. tall with everted or thickened
rims, mostly with a band of rouletted decoration below a plain rim
(Figure 6, 1), a rim diameter of 20-30 cm. and a rim breadth
normally more than 3 cm.; related to these are spherical pots with
straight decorated rims under 3 cm. broad (Shinnie 1960, Figure 12,
Nos 9, 11-12),
Type 2, slightly larger pots with thickened everted rims (Figure 6, 2) in
which the rouletting is normally on the rim and the diameter is
often greater than the height, a variation of this latter form has a
thickening of the internal face of the rim.
Type 3, large spherical pots with wide, straight or everted rims, 15-25 cm.
tall (Figure 6, 4),
Type 4, large jars up to 35 cm tall, with broad collars (5-8 cm.) and beaded
or plain everted rims (Figure 6, 3),
Type 5, large shallow dishes up to 65 cm., in diameter and 20 cm. deep with
thickened rolled over rims (Figure 6, 8), occasional internal handles
and decorated both on the rim and in some cases internally below
Type 6, hemispherical bowls with plain (Figure 8, 6), thickened, everted
(Figure 6, 5), or beaded rims,
Type 7, medium sized spherical pots with beaded or everted rims (Figure 6, 6)
which are largely smaller versions of Type 3.
Type 8, small spherical pots with narrow everted rims (Figure 6, 1-2) and
rim diameters of 9-15 cm.
Type 9, pot stands or pedestalled, wide-mouthed bowls. (Figure 6, 7; figure
Table 1 indicates the number of each type found, their provenance, and
decoration In addition a spherical pot with an inverted rim and narrow mouth,
decorated with a roulette pattern was found in a Bigo I context.
Pottery Type Bigo I Post-Bigo I Rouletted Painted Total
1 .. 6 33 37 16 39
2 .. 18 65 78 10 83
3 .. 2 11 11 6 13
4 .. 3 4 4 3 7
5 .. 6 3 9 7 9
6 .. 23 59 38 14 82
7 .. 9 7 11 3 16
8 .. 4 37 20 10 41
9 .. 6 2 6 5 14*
TOTAL .. 77 221 214 64 304
(* Several pot stands were collected during levelling away from the main sites and so this
total is not the sum total of excavated pieces).
cp: ~ : . ., -
Acrial view of Bigo showing layout of earthworks and papyrus choked Katonga
river. (By courtesy of the Air Ministry, Great Britain Crown Copyright).
,w ; f-.*^t ^
.-* ,* .,> ^ X~
!- ** .,
Site B. Section through outer ditch and bank.
As Post-Bigo I contexts include material from the make-up of the mounds,
which could have derived from Bigo I surfaces, as well as material in the upper-
most layers, it is difficult to give the exact proportions as to the use of various
types at different periods of time. It is evident that Types 4, 5, 7 and 9 are
distinctive to Bigo I and Types 1, 3 and 8, in particular, are probably more
distinctive to Bigo II, though all the types are found in all the different stra-
tigraphical contexts, which indicates the relatively short period of occupation.
Painting has been considered as one of the distinctive features of the
Bigo wares (Lanning 1957). Certainly painting is a feature of most of the sites
where the characteristic Bigo types are found, and is not found on later wares
in western Uganda, but there is evidence that at Bigo itself painting became
less important during the short period of occupation. Altogether 18% of the
total number of rim and decorated sherds bore traces of a red slip or paint,
but when the sherds from Bigo I were taken as a group they amounted to
30% whilst-if Types 4, 5, 7 and 9 are considered early--40% of them are
painted compared to a little over 30% of the forms which seem more common
after Bigo I. The paint in all cases has an ochre base and was applied as a
slip around the rim, particularly the inner side, and as vertical stripes on the
exterior of the rim (Figure 6, 3-6). Type 4 jars are normally of a particularly
fine fabric and the three decorated pieces all bear paint on the inside of the
rims as do six of the nine dishes of Type 5. Occasionally daubs of paint are
found on the bodies of the pots, particularly of Type 8. The pot stands and
Type 8 bowls are often decorated over the foot and over the body of the vessel.
In such instances the rims have been left plain.
In addition to the rouletting and the painted decoration, the rolling over
and thickening of the rims is a characteristic feature. Bevels are occasionally
found on the outward facing lips of pots and jars (Figure 8, 5). Other forms
of decoration are rare, only 11 examples being found of incised or impressed
decoration. The herring bone pattern on the only complete potstand (Figure
8, 7) would appear to have been made with a wooden roulette, which if it is
sc is the earliest example of the use of the carved wooden roulette in Uganda.
Two pots with nicked rims were found, one of which was the poorest made
pot in the collection (Figure 8, 3) and came from the pit dug into the old
ground surface in Trench 5 and thus may predate the whole Bigo sequence.
The pottery varies in colour from buff through to reds, greys and blacks
but this is largely an expression of the firing employed and of the clays used,
rather than an indicator of separate classes of pottery as was proposed by
Mathew (1953) and Lanning (1957). The fabric varies even within single
types and gritted fabrics with large quartz grains, broken pot used as temper
and a black swamp clay, apparently with neither grits nor comminuted pot
fragments, all occur.
The fine ware (Figure 7) which mostly consists of pots with a rim diameter
of less than 15 cm. is on the whole plainer than the coarse ware. There are
most frequently found in Bigo I contexts though sherds also come from the
mound soil. The finest pieces have plain beaded or everted rims and rouletting
is uncommon (Figure 7, 7). Some of the black pots are occasionally rough on
the interior, though a highly burnished, slipped, thin pottery is common
and reminiscent of the later fine wares found on Ankole capital village sites.
Several of the rims have fine incisions along their edge (Figure 7, 8) and in
one case a line of knobs runs along the rim lip (Figure 7, 6).
6. Main Fottery types, coarse ware,
1-2, 4, 4, 5-6 pots,
1. Type 1;
2. Type 2;
3. Type 4;
4. Type 3;
5. Type 6;
6. Type 7;
7. Type 9, pot stand;
8. Type 5, open dish.
Nos. 2-6, 1957 excavation.
Nos. 1, 7, 8, 1960 excavations.
(Reproduced by permission of the editors of the journal of African History.)
7. Main forms of fine ware. 1, 2, 5, 6, 8, black, no. 1 with internal red slip and burnished
exterior, 3, 4, 7, 9, 10, red wares with red slip. (reproduced by permission of the editors
of the Journal ofAfrican History).
1-2, Small spherical pots with everted rims;
3, Course bowl with nicked rim.
4, Elo gated pot with plain everted rim (?variation of Shinnie 1960 figure 10,1).
5, Rim with fingernail impressed herringbone motif.
6, Straight edged hemispherical bowl.
7-9 Pedestal bowls or Dot stands (No. 9, reconstructed in figure 6, 7).
Nos. 1-3, 6, 8, 9 black, 4 buff/red brick, 5 A. 60227/2; 8. A. 60307; 9. A. 60260).
It is of interest to note that none of the very large pots (over 40 cm. tall),
which later are so common at Mubende and areas near Lake Victoria, and
were probably made for the brewing and storing of beer have, been found.
The small number of jars in relation to the large number of spherical pots,
which in most cases have soot on their bases indicative of their usage as cooking
pots, suggests that water and milk was collected and stored in perishable wooden
pots, calabashes and skin containers.
A total of six pieces of hearth kerb were found though none was found in
its original context. From their thickness (over 10 cm.) and their curvature,
indicating a diameter of over a metre, and the absence of any fragments but
rims, it is apparent that they are similar to many pieces also found at Ntusi
which are occasionally decorated with roulette decoration, the finest examples
of which are on exhibition in the Uganda Museum. It is probable that they
belong to hearth kerbs similar to those in declining use by the Bahima today.
These latter are moulded into an oval form up to a metre across and with use
the clay becomes slowly fired. As the clay forming the base of the hearth is
plastered on to the ground, or the surround is placed around a cleared area
of ground, only the kerb is separate and only by finding a hearth in situ would
the whole structure be found.
One brown and rather crude pottery bead was found. (Uganda Museum
A.60312/11, 24 mm. long and 13 mm. at its maximum diameter).
Ironwork (Figure 9)
A limited number of iron objects were found. Some were shapeless agglo-
merations of rust. The best preserved pieces are all from the upper surface
of Mound I and consist of:-
(1) a tanged arrowhead whose original length was of the order of 17 cm.
The tang is square in section with six barbs on two edges but project-
ing from opposite faces. The blade is also asymmetrical in section
with each side of the central axis on a different plane (Uganda
(2) A broken bracelet, of oval shape and section, and originally around
5 cm. by 6 cm. This very similar to one found in the 1957 excavation
and was presumably for a woman or child.
(3) A spear ferrule, originally over 15 cm. long with a square tanged
point and a round section socket (Uganda Museum A.60279).
(4) A broken knife with a tang of quadrilateral section.
Of these pieces, the arrowhead, of which two more fragmentary specimens
have been found, including one from Shinnie's excavation (Figure x4,4), is the
most significant piece. Such tanged arrowheads are not found today amongst the
Interlacustrine Bantu but are common amongst the Madi in the northwest of
Uganda. It is of interest to note a tradition by Mrs. Fisher, in which the first
Babito King, Mpuga Rukidi, is said to have been accompanied by a "body-
guard armed with bows and barbed arrows"." The context in which these
arrowheads were found strengthens the suggestion previously made that they
indicate the Lwoo capture of Bigo12.
9. Iron work.
1. Barbed arrowhead.
3. Spear ferrule.
4. Knife; All from top soil of Mound 1.
Altogether 28 rubbing stones or riders were found, of which three measured
more than 15 cm. across their long axis and 25 less than 15 cm, across. All
but two are quartzites or quartzitic rocks derived from the Karagwe-Ankole
complex and one is of haematite. Fourteen of the stones are from the recent
occupation area in Enclosure I or from surface contexts. Though an extensive
surface collection was made, only 14 grindstones were found, an indication of
the lack of settlement by peoples using grain foods. Of these, nine were found
on the surface around the recent occupation area of Enclosure I and two as
surface finds on the west side of Mound I. Only three grindstones came from
excavations, two from the third layer of a trench across the denuded ekyikari
bank and the other from the mound soil of Mound III. All the grindstones were
broken and are made of granite. As with the stone rubbers etc. found in the 1960
excavations none are of rock which cannot be found within the general Bigo
area. They vary in form largely depending on the shape and size of the original
lump of rock, only one is double, having been used for grinding from two faces.
Dr. Brian Fagan, then of the Rhodes -Livingstone Museum, Northern
Rhodesia examined the bones and from diagnostic pieces identified the following
Domestic cow: teeth and mandibular fragments
Duiker: a mandible and a skull
Pig (species unidentified): phalanges.
Altogether from a collection of 219 limb bones, which excluded small frag-
mentary pieces, he was able to subdivide the 45 % of the collection (99 pieces),
which was capable of identification, into small bovidae (13%),
medium bovidae (12%) and large bovidae (74%). In addition
19 cow teeth were found, as well as fish (?Barbel) bones and an
unidentifiable bird cannon bone. He concluded that "the predominant
limb bones are of large bovidae. I have compared these bones with the limb
bones of domestic ox and am of the opinion that a substantial proportion of
the large bovid bones are those of domestic ox. I do not think that game animals
are common in this collection nor are small stock. No dog bones were identified.
It is noticeable that a number of the limb bones are of immature beasts
although not so commonly as at Bweyorere.13 No deductions can be made
as to the species or size of the cattle on this small sample".
The traditions about Bigo have been discussed by Sir John Gray (1935).
Though Bigo is mentioned in various traditions including those of Bunyoro
and Ankole and is connected with Bacwezi tales, there are no descriptive
accounts about the site itself and many of the traditions would appear to have
been invented later to explain the presence of such earthworks. It is to be hoped
that the new assessments being undertaken on the traditions of western Uganda
by such local scholars as Dr. M. S. M. Kiwanuka, Mr. B. Karugire and Mr. S.
Lunyiigo-Lwanga will go a long way to amplify what little information has
hitherto been collected.
Mugenyi certainly appears in the Bacwezi legends as a brother of Ndahura
and as a keeper of cattle. Bigo is mentioned in traditions associated with both
Ndahura and Wamara. Ndahura was the great warrior and it is not inconceiv-
144 UGANDA JOURNAL
able that he built the large defendable cattle enclosures on the Katonga river.
Gorju (1920, pp. 48-49) cites traditions of how the earthworks were a pro-
tection against the Banyankole who had taken the cattle of Kagoro, son of
Kyomya. Gray thinks that Mugenyi must be the name of the Mucwezi and not
the word for stranger, as many authorities have suggested, since Mugenyi is
only the word for stranger in Luganda and not in the other western lacus-
trine languages like Runyankore which was the more likely language in the
area before it become part of Buganda.
Kagwa's Basekabaka is quoted by Gray (1935, p. 230) as ascribing Bigo
to Cwa Nakaka, the second Kabaka of Buganda who mysteriously disappeared
from Buganda like his father. The fact that the Bacwezi tales are remarkably weak
in Buganda would largely rule this out though certain Baganda clans do claim
an origin from the Bwera area in which Bigo is situated.
When the Bacwezi finally disappeared local tradition 14 recounts how the
area around Bigo was entrusted to a Muiru smith called Kihesi. Kihesi was said
to have been of the Bunyoro Bamoli clan (bushbuck). He was given a drum
made of waterbuck skin called Lusama which was said to be the war drum of
Ndahura. Until recently the drum, or a later replica of it, was kept at Makole
near the Katonga a few miles away from Bigo. Makole was for a time the
Ssaza headquarters of Mawogola county. Kihesi is said to have acquired the
name Baralemwa and established the Kingdom of Bwera. Only the names of
eleven Bannabwera rulers are known which is eleven less than the dynasty
latterly ruling in Buganda.
Bwera is said to have remained quasi-independent throughout the nine-
teenth century, and in 1892 Lugard stipulated that it should so remain. It was
incorporated into Buganda in 1898 and confirmed as such in the 1900 Uganda
Agreement. Kabaka Junju is said to have tried to conquer Bwera but failed
and it was to Bigo that his brother and successor Semakokiro is said to have
taken flight. He is supposed to have rewarded one of the Bannabwera, called
Kehera, by giving him the district of Bulondaganyi in Kyagwe where his
successors remained as cattle keeping people. Kabaka Mwanga in 1898 also took
refuge in the vicinity of Bigo but there are conflicting reports as to whether he
actually spent any time at Bigo itself.
Bigo represents the largest of a group of sites in western Uganda (Figure i,
map 1) at which similar pottery has been found and which are thus assumed
to belong to the same period. At Bigo more than 200,000 cubic metres of earth
and rock were moved in the digging of ditches alone. Such an effort over a
relatively short period of time implies control over a labour force of perhaps
several thousand men for a period of years. The presence in the centre of the
site of an ekyikari or royal enclosure, as are later found on Ankole capital sites,
suggests that Bigo was a capital or orurembo. It is to be noted that similar
features are not found at the other earthwork sites, either those in the
immediate vicinity, like Kasonko and Kagogo, or those further afield like
Kibengo. From this can be inferred the existence of a degree of political
The evidence from Bigo clearly indicates a pastoral people and it is to be
noted that before the rinderpest epidemics of the 1890's and the spread of
the tsetsefly in the twentieth century, Bwera was good cattle country. Much of
the thorn bush, mostly of Acacia hostii, currently in the area, is of recent growth
s ' . . ",- -
'. . . ., .. . .. . . . ." .
Site D. after clearance. Figure in singlet is standing at junction of two ditches.
;+~~' L .'k
ci '-'-i .r'j
- U U U
Hearth with clay kerb in modern Bahima kraal near Bigo.
and when the first European explorers visited western Uganda in the second
half of the nineteenth century they described grassland rather than the bush
that now exists. The low ground towards the Kakinga is relatively well watered
and the Katonga river was less papyrus choked if the local traditions have any
reliability. The large area of Bigo must be related to the amount of pasturage
needed for royal herds.
The outer ditch of Bigo is hardly defensive having as it does many entrances.
The concentration on ditches rather than banks further indicates that it was
cattle they were keeping in, rather than building an earthwork impregnable to
outside assault. The inner enclosures would seem the more defensible. On the
south and east the ditches are backed by banks and are the deepest of the whole
Bigo system. There are only single entrances to Enclosures 3 and 4 and in each
case there are banks both internally and externally. The main entrance from
the south to Enclosure 1, and ultimately to the presumed ekyikari, also has
double banking at one side. The fact that the ekyikari faces the large break in
the ditch separating Enclosures 1 and 2 would suggest that the main approach
was from the south. The only other openings into Enclosure 2 are relatively insig-
nificant. It is thus possible to think of Enclosure 2 as being for the ruler with a
large enclosure to the east for cattle, whilst Enclosure 1 which is somewhat
smaller, and less advantageously placed on the slope, has the smaller Enclosure
3 attached. It is impossible to speculate further about the function of the enclo-
sures. No traces of buildings have been found except those revealed on the
ground surface buried beneath Mound I, the clearance of which would involve
a larger excavation than was feasible in 1960.
On the level area in front of the ekyikari and elsewhere around the site
very little in the way of surface pottery has been found, which confirms that
the occupation of the site was relatively short. The lack of strongly marked
differences between the pottery assemblages would support this conclusion
though future excavations will need to make a more exacting serialogical study
of the pottery from different stratigraphic units. It is to be assumed that the
habitations of the occupants of Bigo were of a flimsy character similar to those
of the present-day Bahima. The dearth of material finds or of structures
of any kind is in contrast to the immensity of the earthworks and would suggest
that the makers of Bigo regarded the earthworks as public works bestowing
prestige on their occupants.
The identity of the Bacwezi has been discussed elsewhere at length (Pos-
nansky 1966 and 1969). There are many factors to suggest that the makers
of Bigo were the ancestors of the Bahima who later established the Bahinda
dynasties of southern Uganda. The use of the roulette in pottery decoration
seems to have been an innovation coming from the north in the second mil-
lennium AD. Its use reaches into Rwanda, Burundi, Tanzania and Kenya
and so cannot be linked with the later Lwoo movements. The use of roulettes
is known from a wide area of the Sudanic belt and movements of peoples
may be assumed to have taken place into East Africa from the southern Sudan.
These movements were probably also responsible for the spread of the long-
horned humped Sanga cows and the development of predominantly pastoral
states, such as those belonging to the socially stratified societies of Ankole and
Rwanda. The movements were probably never large and successive groups
of new arrivals quickly adopted the Bantu language of the majority, possibly
through concubinage. In a recent work, Hiernaux 15 has suggested that the dif-
ferent physical appearance of the Bahima and Batusi from the Bairu and
146 UGANDA JOURNAL
Bahutu could be due to such movements from northeastern Africa. He uses the
term Ethiopide for these groups. Though it is possible to accept the presence
of such outside elements in the Bahima, the forces of social and nutritional
selection discussed in Posnansky 1966 should also be taken into account and
could have had some effect in producing taller, thinner people with relatively
If the main phase of Bigo is ascribed to the possibly ethnically heterogeneous
ancestors of the Bahima, a motivating force should perhaps be suggested to
explain why they threw up such earthworks at that particular time. It is possible
that movements from the north such as those of the Madi and later of the Lwoo
necessitated otherwise fragmented groups to come together for defence. The
situation of Bigo, flanked by Kasonko and Kagogo on the Katonga river or
Kibengo on the Nkusi river suggests that the river crossings were important
for defence from movement from the north.
The Bigo 'culture' extended over a wide area of the short grassland country
of western Uganda and elements of the pottery continued in the same area
till the present-day, particularly the rouletting and the rolled-over and
thickened rims. One element which disappears which may, however, be of
diagnostic significance, is the use of red paint. Red paint is found in the Sudanic
belt stretching as far west as Mauritania, and its limited use at Bigo could be an
indication of the northern origin of the maker of Bigo. Unfortunately until
extensive work is conducted in the southern Sudan, and areas in the Sudanic
belt immediately to the west, very little can be said about the possible ethnic
relationships of the makers of Bigo. The only items demonstrably of southern
Sudanese or northern Uganda origin are the tanged arrowheads found on the
surface of Mound I which could betoken the arrival of the Lwoo, the destroyers
of the power of the makers of Bigo. The lack of a startling contrast between
Mound I surface pottery at Bigo and pottery associated with the ditches
goes a long way to arguing cultural continuity. As the archaeological
evidence stands at present isolated from a detailed knowledge of the arch-
aeology of areas immediately to the west or to the north, it is conceivable
that the makers of Bigo and its destroyers were of similar origin and could even
have been Lwoo as originally suggested by Crazzolara.16 Nevertheless the
absence of strong traditions about the Bacwezi amongst the Lwoo and the
lack of survival amongst them of other elements of the Bigo 'culture', like the
ekyikari banks and the hearths, runs counter to this view.
Bigo has for long been associated in the literature with the nearby site of
Ntusi. Unfortunately no major excavation has been conducted at Ntusi either
to determine the exact nature of the 'dam' nor to work out the cultural sequence.
Examination of surface collections in the Uganda Museum indicates certain
marked contrasts in the pottery of the two sites. At Ntusi there are many
examples of roulettes other than of the simple knotted grass variety which is
dominant at Bigo. Nicked rims, insignificant at Bigo, are one of the charac-
teristic features of the Ntusi assemblages and develop into what have been termed,
for want of a better term, piecrustt' rims. Statistically from surface collections,
painting is rarer and yet painted rims have been selectively chosen. The pottery
perhaps compares as much with Mubende as with Bigo and the impression
gained is of a site occupied for a far longer period of time than Bigo. In addition
grindstones are very numerous.
It is evident that much still remains to be done both at Bigo and in the
surrounding areas if a vital phase of Uganda history is to be more clearly
understood. At Bigo itself a larger excavation is necessary to expose the various
postholes and other features on the buried ground surface in order to describe
more adequately the Bigo I occupation. More radiocarbon dates from many
more stratigraphic features are necessary to appreciate the structural history.
None of the entrances to the inner enclosures have been tested to see if any
wooden structures existed. At Ntusi it is vital that large excavations are required
to determine its exact relationship to Bigo, with which its name is so often
connected. What the 1960 excavations did achieve was the elaboration of the
physical achievement of the makers of Bigo, its dating to the Bacwezi era and the
demonstration that though short-lived, and possibly never finally completed,
at least three phases of development are represented in the Bigo Phase I build-
ings, the ditch system and the ekyikari bank and the destruction of that bank
and the erection of Mound II.
The site was excavated with the financial help of the Uganda Museum,
the Uganda Protectorate Government and the British Academy. Equipment
was loaned by the Game and Fisheries Department, the Geological Survey,
the Boy Scouts Association, the Ministry of Community Development and
East African Railways and Harbours. Research on the finds and radiocarbon
dating were made possible by grants from UNESCO and the Wenner-Gren
Foundation for Anthropological Research. The East African Tobacco Com-
pany made a grant to Makerere College to enable undergraduate participation.
The excavator was assisted in the field direction by Messrs. Charles Sekintu,
Patrick Bulenzi, Michael Harlow and Joshua Muthama. Professor Roland
Oliver and Miss P. Tangye assisted in a volunteer capacity for varying periods
and seven students of Makerere College assisted for a fortnight. An average of 26
men was employed for most of the period for the excavations and for clearing
dense undergrowth in the ditches. Dr. Brian Fagan examined the faunal
remains. Miss Kathleen Lyons and Mr. Peter Schmidt, then graduate students
of the Makerere African Studies Programme, assisted with the analysis of the
pottery. Pot drawings are by Mr. George Kakooza, Mr. Francis Serumaga and
Mr. Godfrey Nyerwanire. Figure 1 was drawn by Mr. M. A. Torgbor of the
University of Ghana whilst all the plans, iron work and plates are by the
writer. Mr. E. C. Lanning very kindly made available notes on Bigo that he had
collected whilst he was with the District Administration at Masaka.
ANNOTATED BIBLIOGRAPHY OF THE PRINCIPAL REFERENCES TO
1909 Uganda Official Gazette, 15 May, pp. 137-38, description of earth-
works by D. L. Baines, at the time no habitation within the area.
Ascribed to Mugenyi who came from north soon after Kintu's time.
1910 Uganda Official Gazette, 15 April, report of a view of an unnamed
White Father who ascribed the earthworks to an Abyssinian called
Zimbo who passed through Uganda in the sixteenth century. From
the context of the note it is apparent that Zimbo is a mistake for the
Zimba whose origin is unknown but who ravaged the coast of East
Africa in the late sixteenth century AD. There is no evidence though
of Zimba movements as far into the interior as Uganda.
1920 Gorju, P. J., Entre le Victoria, L'Albert et L'Edouard, Rennes, pp. 53-54.
148 UGANDA JOURNAL
Gorju visited Bigo in 1919 and ascribed the Bigo earthworks to the
Mucwezi, Mugenyi, and considered that they were built to defend
the area from the south. The Bacwezi he thinks were Bahima, the
the Bahima were Galla and he suggested correspondences between the
Galla word for moat (bijo) and Bigo.
1922 Combe, A. D., The ancient trenches at and near Biggo, unpublished
report in Geological Survey Files, Entebbe, besides making an ex-
tensive survey using prismatic compass, he found large circular
"dishes" up to 1.25 metres in diameter. It is possible that these dishes
(now lost) might have been complete hearths. He suggested that the
"fort" was unfinished.
1934 Wayland, E. J., Notes on the Biggo bya Mugenyi: some ancient earth-
works in northern Buddu, Uganda J., 2 pp. 21-32. Incorporates Baines'
and Combe's descriptions and Combe's plans. Compares Bigo to
Zimbabwe both of which he considered Bantu and their culture
sprung from a "common root".
1935 Gray, J. M., The riddle of Biggo, UgandaJ., 2 pp. 226-33. Provides an
account of the oral traditions relating to Bigo and considers the earth-
works of Kasonko and Kagogo to be flank defences for Bigo.
1952 Western Province "Crewe", history of Makole, Uganda Herald, 5
February a description of the history of the Bigo region on the occasion
of the opening of the railway station at Makole, nine miles from Bigo.
Mathew, Rev. A. G., Report on the archaeological sites in western
Uganda, unpublished report to H.E. The Governor (Paper No.
15 in MIS 15, Resident of Buganda Files).
1953 Mathew, Rev. A. G. Recent discoveries in East African archaeology,
Antiquity, 27 pp. 212-19. Emphasizes the importance of Bigo in rela-
tion to the Iron Age of the interior of East Africa and relates it
firmly to Ntusi whose size he somewhat exaggerated. The first publi-
cation of a Bigo airphotograph.
Uganda Oficial Bulletin, No. 43, item 26, (April, Vol. 4, pt. 1, pp. 13-
15). Archaeological sites on the Katonga River, with special refer-
ence to the site known as BIGO BYA MUGENYI, an account by
K. Marshall, Government Archaeologist attached to the Geological
Survey of Uganda, who recorded visits to Bigo prior to 1953 together
with a summary description of the site and a plan for extensive ex-
Lanning, E. C., Ancient earthworks in western Uganda', Uganda J. 17,
pp. 51-62, a description and map of all the earthworks thought to be
related to Bigo and tentative classification. Ascribes the sites to the
1954 Cole, S., The prehistory of East Africa, Penguin, London, pp. 280-81,
a description of the earthworks largely based on work by E. C. Lanning.
Relates movements from north to Hamitic invaders.
Wachsmann, K. P., Ancient earthworks in western Uganda; notes on
finds, Uganda J., 18, pp. 190-92, a record of surface finds of a green
glass bead and stone pendant from Bigo made by Isufu Mukasa.
f'"7 Lanning, E. C., Proto-historic pottery in Uganda, Proc. Third Pan-
African Congress on Prehistory, Livingstone 1955, London, pp. 313-318, a
description of Uganda's Iron Age pottery with particular attention
given to the Bigo wares.
1957 Lanning, E. C., World's most extensive system of earthworks, East
Africa and Rhodesia, 14 November p. 337.
1958 Lanning, E. C., Forts of the stranger, East African Annual 1958-59
(published by East African Standard, Nairobi), pp. 79-83, a descrip-
tion and survey of attitudes to Bigo.
Wrigley, C. C., Some thoughts on the Bacwezi, UgandaJ. 22, pp. 15-17
refutes the possibility that the Bahima were the ancestors of the,
Bacwezi or that they built Bigo which he ascribes to a loose-knit
1959 Posnansky, M., Progress and prospects in historical archaeology in
Uganda, Uganda Museum Occasional Paper No. 4, pp. 31-40, contains
a survey of the Iron Age of Uganda and the significance of Bigo for
-- Shinnie, P. L., Excavations at Bigo, Uganda, Antiquity, 33, pp.54-57.
A preliminary account of the 1957 excavations.
1960 Shinnie, P. L. Excavations at Bigo 1957, Uganda J. 24, pp. 16-29,
a definitive account of the 1957 excavations,
1961a. Posnansky, M., Pottery types from archaeological sites in East Africa,
Journal of African History, 2, pp. 181-191, contains an initial typology
of Bigo ware and a preliminary account of the 1960 excavations.
b. Posnansky, M., Iron Age in East and Central Africa-points of compari-
son, South African Archaeological Bulletin, x6, pp. 134-136, a refutation
of the similarities between the form of the Bigo earthworks and Zim-
babwe proposed by Wayland and later writers.
1963 Posnansky, M., Towards an historical geography of Uganda, East
African Geographical Review, I, pp. 7-20, contains a discussion of the
Bigo culture in relation to the historical geography of western Uganda.
Cole, S., The prehistory of East Africa, New York, Macmillan, pp. 319-
24, largely an amplification of the section on Bigo contained in her
1963 Shinnie, M., Ancient African Kingdoms, London, Arnold, pp. 93-97,
provides a brief description of the 1957 excavations in context of the
history of the interlacustrine area.
1966 Posnansky, M., Kingship, archaeology and historical myth, UgandaJ. 3o,
pp. 1-12, an evaluation of the conflicting views concerning the
first centralized states of western Uganda and the relationship between
the Bacwezi and the Bigo culture. Demonstrates the cultural continuity.
existing between the Bigo culture and the later Bahima kingdoms
1967 Posnansky, M., The Iron Age in East Africa, in Background to evolution
in Africa, ed. W. W. Bishop and J. D. Clark, pp. 629-649, places the
Bigo culture within the context of the East African Iron Age.
Davidson, B., The growth of African civilization: East and Central Africa
to the late Nineteenth Century, London and Nairobi, Longmans, pp. 52-54,
advocates the prestige significance of Bigo as Wamara's capital.
1969 Posnansky, M., Introduction to second edition of Twilight tales of the
black Baganda by Mrs. A. B. Fisher, London, Cass, a survey of the Bacwezi
traditions of western Uganda with an evaluation of the historical
significance of the builders of Bigo.
150 UGANDA JOURNAL
1. Collector was the original term given in Uganda to a District Commissioner.
2. In Posnansky 1961a p. 187 following Shinnie 1960 p. 16 the area is inadvertently given
as five square miles.
3. The excavation was conducted using feet and inch measurements though metric equiva-
lents are given for international reference.
4. The plan depicting the layout of the trenches in Shinnie 1960 (p. 17, Figure 1) was shown
to be somewhat inaccurate. Trench I was 180 feet on the ground though only 160 feet
on the plan, whilst Trench II should be 10 feet further north than it is depicted. His Trench
I as indicated on his Figure 9 is only the southern 60 feet of the trench indicated on Figure
7, the northern 110 feet was only taken down two feet.
5. Interpretation of the ditch sequence based on Shinnie 1960, pp. 17-18.
6. Interpretation of the radiocarbon dates is based on comments provided by Geochron
7. Posnansky, 1966 pp. 4-5.
8. All the finds are in the Uganda Museum collections.
9. This is a larger figure than that given in the report on Bweyorere (Uganda, J. 32 1968, p.
170 the earlier total being based on an incomplete field count.
10. Shinnie, 1960, p. 18.
11. Fisher, 1911, p. 119.
12. Posnansky, 1966, p. 6.
13. Posnansky, 1968, Uganda J. 32, 1968, p. 172
14. Contained in historical files collected by the Masaka District Administration from 1908-62,
presented to the author and now deposited with Makerere University College Library.
15. Hiernaux, J., La diversity humaine en Afrique Subsaharienne, Brussels. 1968.
16. Crazzolara, J.P., The Lwoo, Uganda J., 5, 1937, pp. 1-21.
Uganda Journal, 33, 2(1969) pp. 151-162.
THE BAKOPI IN THE KINGDOM OF BUGANDA,
AN ANALYSIS OF THE CONDITION OF
THE PEASANT CLASS IN THE EARLY COLONIAL PERIOD
By J. A. ATANDA
Writers on the colonial history of the Kingdom of Buganda seem to have
shared, unconsciously, the prejudice of the Ganda traditional society against
the class of people known as the bakopi in that society. The prejudice is that
of neglect. In the traditional society, the bakopi were people who "did not
matter". Professor Fallers defines a mukopi (pl. bakopi) as "simply a person of
no particular distinction."2 Fallers goes on to say that to be distinguished from
a mukopi is either "to be a mulangira, prince or descendant of the Kabaka" or
"to be a mwami-a person who rules". By this definition, therefore, the bakopi
in the period under survey were the group of "undistinguished" persons in
Buganda, including, as Fallers points out, any person who was neither a
mulangira nor a mwami and who was not important in any other way.
Yet, there was much that happened to the bakopi during the early part
of the colonial period worthy of record. If the bakopi were "undistinguished"
as far as getting mailo lands and ruling sazas, gombololas, milukas, were concerned,
they were distinguished in the manner of their "suffering without bitterness".
An analysis of their plight is necessary for two reasons. First, it will throw some
light on the social history of the period. Secondly, it will show another picture
of the colonial rule in Buganda which has hitherto been painted largely with
focus on the benefits derived from that rule by the Kabaka and the Bakungu
In this regard Sir Harry Johnston's implied optimism that his 1900
Agreement would also confer benefits on the bakopi seems not to have been
justified. In a mood of triumph recommending the Buganda Agreement to
Lord Salisbury, Harry Johnston wrote, "A good deal of this land handed over
to Her Majesty's Government is under cultivation, but it is settled by the
'bakopi', or peasants who up to the present have been practically serfs. These
people will become, therefore, the tenants of the British Government as re-
presented by the Uganda Administration."3 Subsequent events did not justify
Johnston's two hopes that the bakopi would become "tenants of the British
Government" and would cease to be practically serfs. The land settlement
effected after the signing of the Agreement left very little in the hands of the
government that was valuable for peasant agriculture,4 and the bakopi willy-
nilly, became, or remained, to a large extent tenants of the chiefs. The bakopi
discovered that they had more masters to serve during the colonial period than
before. Not only did they continue to serve the Kabaka and the chiefs, but
they also had to serve the new government and the Christian Missions. To a
large extent this problem of serving four masters remained true up to the end
of 1927 when the "Busuhl and Envujo Law" was passed.
152 UGANDA JOURNAL
During the period 1900 to 1912 the bakopi faced many dilemmas in meeting
their obligations to their various masters. In order of priority, at least from the
point of view of the administration, the obligations to the government came
first. These comprised largely the payment of taxes and the provision of
labour, which more often than not was unpaid. On occasions, both taxes and
labour were related.
As Professor Low has pointed out,5 one of the main duties which the chiefs
of Buganda had to perform in consequence of the 1900 Agreement was the
collection of hut tax. This was the anchor sheet on which the Agreement was
to rest, judging from its Clause 20 which hinted that the Agreement would
lapse if the chiefs should fail to meet the hut tax obligation.6 It may be argued
that, since some form of taxation had been imposed on the people prior to the
colonial period, the imposition of the hut tax in 1900 may not have been an
unusual burden. But, as has been pointed out by Low, the new hut tax "was
clearly a heavier imposition than had been levied previously . ."7
The difficulty of the bakopi in getting taxes paid increased greatly in 1900
when, as much as possible, farm produce and livestock, would no longer be
accepted as taxes. Instead, cash payment in rupees was wanted, and only
where this was not possible, a month's labour for the government was accepted.
A number of bakopi who could not get rupees locally and who did not
want to do a month's labour were obliged to seek employment elsewhere. The
burden in 1900 was considerable, and it has been reported that some bakopi
had to pawn their children for one or two years and one or two were said to
have committed suicide.8
This solution by emigration was typified by what became known in
official records as the "Baganda colony at Kisumu". In July 1901, the Com-
missioner reported about this "colony" that:
"At Kisumu there was, two months ago, a colony of about sixty Baganda;
probably there are many more now. It is not quite clear why these men
have gone there. At first I was told they had arrived with the meritorious in-
tention of working for rupees in order to pay their hut tax, and I thought
they were worthy of encouragement... I now hear that these Baganda
went to Kisumu to evade hut tax in their own country, and not, as I
was at first led to believe, to enable them to earn money with which to
pay the tax . "9 Although the Government eventually stretched its
dragnet over this "colony" the experiment of the Kisumu "colonists"
set the pattern for other emigrations in the futures.
The government itself realized that the year 1901 was a hard one for the
tax payers, the majority of whom were bakopi. This, in itself was largely a
direct result of official pressure on the chiefs to force their subjects to pay.
For example, the Acting Commissioner in charge of Buganda decided some
times in that year to withhold the salaries of the chiefs "till they prove to me
that they are doing the work [of collecting taxes] for which they are paid."'0
The situation with regards to taxation was worse for the bakopi the following
year; for Low's assertion that "by 1902 the crisis was over""1 was not borne
out by subsequent events. As a matter of fact, in that year the burden was
increased as a result of a decision taken by the Lukiko in late 1901 and accepted
by the government, to allow the chiefs to charge a rent of two rupees from
each of their bakopi tenants.12 This was in addition to the three rupees payable
to government as hut tax.
Inability on the part of the bakopi to produce cash to meet these obligations
meant that they must pay for both tax and rent in labour. But the payment of
THE BAKOPI 153
tax by labour was in itself a hard alternative, judging from the conditions
under which the labour was to be rendered. These conditions were laid down
in Johnston's memorandum of 11 October, 1900, thus:
"Where people are unable to pay their taxes in cash, ivory, rubber, or
wild animals, they may do so in labour, the labour taking the form of a
month's work in payment of each hut tax or gun tax. Whilst working for
the Government on those terms, the labourers must arrange to find their
own food if they are working off their taxes by settled labour at any one
place and not by porterage. If they are working to pay off tax by por-
terage, as that system of labour is paid for at a higher rate than others,
they will be given their posho at the expense of the Administration during
the month they are employed at carrying. If natives prefer however, they
can work on settled works for the Government for a period of two months
and receive their posho at the expense of the Administration; or they can
work one month and find their posho themselves."13
From whichever way it is considered, any of the conditions stated above
meant hardship for the bakopi. In the case of labour by porterage, the food
was paid for by the excess of the wages paid for this type of labour over wages
for ordinary labour, but most bakopi disliked labour by porterage because it
involved moving from place to place. In the case of "settled works", labourers
had to work for two months instead of one, if they were fed, which meant that
one month's labour was for the tax and the other month's labour for the cost
of feeding. This meant that labourers so engaged had to be away from home
for two months.
Thus it was the first condition, labouring for a month with self-feeding,
that most people opted for. Even this involved untold hardship, since the
labourers had to work far away from their own homes and in places where
food could not be got easily. The place where this situation obtained at its
worst was Entebbe, and it was to Entebbe that most of the labourers were
sent. As early as 1901, there was the report to the effect that: "There is an
outcry at present by the labourers against working off their taxes at Entebbe,
as they suffer so from hunger."14 The Administration tried to remedy this by
a plan to build shops which would be let to traders who could provide food
for sale. But it is not certain whether this plan was effected, and if so, whether
it was effective. Whatever the case, the labourers at Entebbe experienced other
hardships. In August 1902, the Deputy Commissioner in charge of Buganda,
George Wilson, reported from Kampala that:
"the main difficulty at present is an alleged dislike for labour in Entebbe
that has become national. I am told that some months ago a thousand
men left Bugangadzi to work out the tax. Within two days from here they
heard they were to work at Entebbe, they at once bolted back. Since then
orders were issued not to send people from that country in here or for
more than four days' journey. The difficulty has re-appeared and only
as late as this morning 100 men were brought in from Butunzi-when
they were ordered to Entebbe 70 asked leave so as to try and get rupees".s1
Three reasons were given by Wilson for this general dislike by the bakopi
of labour at Entebbe. There was "difficulties in getting supplies", there was
"beating by overseers, (particularly one Juta)" and the labourers had to work
for "long hours-from dawn, when they are counted till late afternoon when
they were counted out again-and this without a midday interval, and with a
distant camp". In some circumstances the labourers had "no time for getting
154 UGANDA JOURNAL
rations and cooking them". The labourers were reported to have said that they
had "not even time to get the jiggers out of their toes". In the circumstances,
it would not be surprising that the belief became current "that several bakopi
have died on returning to their shambas after doing a month's work at
Entebbe". In spite of these hardships suffered by the bakopi, the Deputy
Commissioner, George Wilson, did not seem to sympathize with them. His
lack of sympathy was largely due to the fact that he did not believe that the
bakopi were really suffering. Although, as he himself said, he had been "aware
of a feeling antagonistic to Entebbe for some time", he did not believe that
there was any just cause for such a feeling. In fact, he only made a report on
the bakopi grievances to the Commissioner in August 1902 because the bakopi
case was "strongly represented" to him by the Regents. Even then, his view
was that "there may be some truth" in the Regents' reports, in which case,
calling the attention of the administration to the situation might remove any
such grievances that might exist. Wilson remained generally unsympathetic
and concluded with the comment "I am perplexed as to whether the bakopi
are in need of the unremitting and tender care we're naturally rather inclined
to give them."
The attitude, that the bakopi in Buganda were not really suffering, some-
times served as an impediment to any attempt by the Protectorate Govern-
ment to mitigate the hardship suffered by the bakopi. There was a further
example of this in 1902. In August of that year, Mr. Spire reported to the
Commissioner the difficulties which the bakopi were facing in Buddu as a result
of tough measures taken by the chiefs with respect to tax collection. On their
return from Kampala 16 in June that year, the chiefs "appear to have given
out that the people must pay in rupees," and when the people asked where
they were to get the rupees, the chiefs replied that "they must produce them
from somewhere."17 However, the Commissioner was moved with compassion
by Mr. Spire's assurance that "there are practically no rupees in Buddu, the
large majority cannot pay in cash." Consequently, the Commissioner coun-
selled caution. He declared:
"Although I am anxious that those who can pay in cash should do so,
it is quite contrary to my wishes or any instructions I have issued that
pressure should be brought to bear on the people to produce rupees
when they have not got them, as appears to be the case with the majority
of the people in Buddu."l8
Unfortunately for the bakopi, the investigation which the Commissioner
ordered into Mr. Spire's allegation was conducted under Wilsonian influence.
George Wilson influenced the investigation in a memorandum which he
issued in his capacity as the Deputy Commissioner for Buganda to all the
Collectors in the Province who were to assist in the investigation. In that
memorandum, not only did he dismiss Spire's allegation (yet to be investigated)
as unfounded, but he also warned all the Collectors both in Buddu and in
other parts of Buganda "not to do anything", in conducting their enquiries,
"that will discourage a payment of taxes in cash when such payment can be
expected with reason."]9 In the circumstances, it should not be surprising that
the Collector who investigated the allegations in Buddu reported that:
"I have made all possible enquiries from chiefs, sub-chiefs and people,
and as a result find no ground on which to base any charge of undue
oppression in the matter of a cash payment... Their chiefs have in no
case that I could discern put any undue pressure on the 'bakopi' to
THE BAKOPI 155
This "favourable" report was quickly sent by Wilson to the Commissioner
who expressed his happiness that "no undue pressure has been employed by
the chiefs".21 It was clear that no salvation was immediately forthcoming to
the bakopi from the government. Nor could the chiefs afford to show any
sympathy. On their part, not only were they demanding their rents, but in
view of the "no tax, no salary" ultimatum given them in 1901, they also could
not exercise leniency in collecting taxes.
In the circumstances, the bakopi were obliged to work out their own sal-
vation. The bakopi devised a number of ruses to evade paying hut tax either
in cash or in labour or they resorted to emigration. One means of evading
taxation was an offspring of a regulation that if a man could produce two
rupees for inspection as evidence of his intention to pay the tax of three rupees
in full, he could be allowed time to work for himself to get the remaining one
rupee. The bakopi, who could not individually get two rupees each in the first
instance and who had no hope at all of getting rupees to pay the tax in full,
devised a method of collectively getting only two rupees which they made
available for inspection from hut to hut. The report of how this was done, as
given by the administration in August 1902, is worth quoting:
"When hut to hut visitation was quite recently made, some of the chiefs
were delighted with the promise of a big cash return and wondered where
all the rupees came from. It has now transpired that in some villages the
same two rupees have done the duty for the whole settlement, these
having been passed on from hut to hut ahead of the inspectors, mostly
However, this "rather an amusing incident", as Wilson described it, was soon
checked by the saza chiefs who directed that prospective taxpayers be massed
in their compounds to prevent the ubiquitous two rupees circulating.
Another method of evasion was resorted to by arguing that a hut tax could
only be rightly imposed on married men, as only such men were obliged to
have huts. Thus, during the tax season, bachelors who had huts often aban-
doned them to live with a married couple, posing as members of the couple's
family. According to one report in 1903, as "many as four or five able-bodied
young men will live in this way, as perhaps the 'family' of an old couple for
whom they will pay the tax."23 This same report also records that "if a young
man lost his wife, even though he might have had a home and garden of his
own he, in order to avoid paying hut tax, would leave the place and let it go
to ruin, take up his abode with some married couple whom he will probably
help to pay the tax-one tax again." In cases where death did not intervene,
men were known to send their wives away so that they became "bachelors"
during the period of tax collection. They would then live in two or three as a
"family" to pay one tax or "become practically vagrant if they could place
their wives in permanent quarters."24 Other married bakopi, whose wives did
not die, and who could not become "bachelors" by sending their wives away
on a visit, tried to evade hut tax by "absenting themselves from their huts"
whenever a tax collector was sent to them.25 But the saza chiefs in Buyaga
found such a drastic solution to this trick that it did not last long. "All the
women, whose husbands are absent have been brought together in a large
camp and are kept until their husbands either pay cash or go to Kakumiro as
The evasion of hut tax was combated by legislation. There was, for example,
"The Hut Tax Evasion Law" passed by the Lukiko in June 1904 to the effect
that any person "found assisting another person to hide upon his land ....
156 UGANDA JOURNAL
shall be guilty of an offence and shall be liable to a fine of three rupees or to
one month's imprisonment."27 At the same time, there was "The Tax Collec-
tion Law", which made it incumbent on each chief to see that his subjects
"pay the Government Hut Tax, before they do any work for him."28
By far the most effective means of checking tax evasion was the imposition
of a poll tax of two rupees introduced in September 1904 and which was paid
by male "adults not liable to Hut Tax".29 Poll tax must have been found not
only an effective check against tax evasion but also a better method of raising
revenue so that it completely superseded hut tax in 1909. In that year hut
tax was abolished and an annual poll tax of five rupees was imposed on every
male adult.30 For the bakopi, this meant a heavier burden than hitherto. The
chiefs were very vigilant in collecting the tax, twenty per cent of which went
back to them as rebate, so that evasion was either impossible or very difficult.
From 1909 onwards, it appeared that the only way open to the bakopi to avoid
the burden or to raise the necessary rupees was to leave Buganda.
Emigration itself was not a new method to the bakopi in 1909. as has been
shown earlier in the case of the Baganda "colonists" at Kisumu. Again, as
early as April 1902, the Acting Sub-Commissioner for Buganda, Stanley Tom-
kins, learnt during a tour of the Province that some of the "poor people ....
have gone to German territory to evade taxes . "31 Although he was opti-
mistic that "they will return ere long", it is not known how long it took them
to return or whether, in fact, they ever returned. What is certain is that there
were more serious reports of emigration from elsewhere in Buganda before that
year ended. There was, for example, a report of mass emigration from Mawo-
gola county. This was so serious that in reporting to the Commissioner, an
official, Mr. Spire, said that the word to use in describing the phenomenon
was not "emigration" but "Exodus". He reported in October 1902 that "In
passing through the country my course led me from one deserted village to
another. The place is becoming a wilderness."32 Most of the people involved
in it went to Toro, where the collection of hut tax was reported to be more
It will be wrong to think that emigration was caused only by the desire
of the bakopi to evade tax. There were other reasons. Of these, the desire to
avoid endless labour for the chiefs was by far the greatest. This led also to
internal migration. On wanting to know "how it was that the bakopi could
not remain in one saza", an Administrative Officer, F. H. Leakey, was
informed by the chiefs that "the minute they heard that Msaza was building
a house, making a road, or cultivating, they promptly left."33 Indeed, there
can be no doubt that the bakopi had genuine reason to run away from endless
labour. As an official, Stanley Tomkins, remarked after a tour of Buganda
Province in 1903, there was "a good deal of discontent owing to the amount
of work they are called upon to do."34 The depth of the discontent can be
judged from the nature of the grievances which the bakopi expressed to Tom-
kins during his tour:
"They say they have to work for a month or produce Rs. 3 every year
for Government; then they have to work for a month or provide Rs. 2
for their chiefs; then they have to cultivate a patch of ground for Govern-
ment, and at all times they are called on to make roads, schools, churches,
etc. and they really don't know what is going to happen next. There is
free talk amongst the people that they are all going to be taken presently
to bridge the Nile."
THE BAKOPI 157
It is curious that the official to whom all these grievances were expressed
and who admitted that the people had a "good deal of discontent" was not
moved to consider the plight of the bakopi with sympathy. His view was that
"the amount of work they are called upon to do is really very little, but to
people like themselves, who have lived all their lives in almost idleness, the
task seems very hard," and that the bakopi should be worked so hard that they
would get used to "doing double of what is asked of them now."
It was in order to achieve this objective that during the talk he subse-
quently had with the Regents in November 1903, Tomkins threatened to
"hold every individual chief, great and small, personally responsible for the
cultivation on his milos"35. By this threat he hoped to stop the "running away
of people from one district to another" and also check chiefs "who are actually
avoiding making their people work, to induce crowds to settle on their lands."
One other possible reason for this threat was to boost the "produce scheme".
By this scheme, each chief was expected to put his mailo land under cultivation
in order to produce one or more of the cash crops, notably cotton, which were
in demand by British traders.
From whichever way it is considered, the threat issued to the chiefs by
Tomkins was to have the effect of increasing the amount of labour hitherto
done by the bakopi. A common response to this increased pressure was to
emigrate. Another example of emigration was from Buyaga, where it was
reported in early 1904 that about eight hundred families left that district
towards the end of the previous year for Bunyoro and Toro.36 There was also
a report that at Kakumiro, over 2,000 men deserted to emigrate to Toro.S7
These reports of emigration must have forced the administration to
inquire more formally into the bakopi discontent. Even Tomkins, had to under-
take a tour of the Kingdom to investigate the nature of the grievances which
the bakopi had against the chiefs in particular and the administration in general.
Thus, in March 1905, Tomkins, in his capacity as the Sub-Commissioner for
Buganda Province and accompanied by the Katikiro, Sir Apolo Kagwa, toured
a number of counties in the province with a view to inquiring into bakopi
grievances. The r6sum6 of the grievances as given by Tomkins is worth quoting:
"In the first place, many accused the chiefs of refusing to give up the
tax tickets until they paid so many shells (I have some well-proved cases
of this), and in some cases till they had paid them the two rupees rent
due from them to the landowner. Then again, there was an outcry about
the amount of work they had to do which took them away from their
homes. If they could not pay cash for hut-tax, they had to go to Kampala
or elsewhere, to work a month for Government; then they had to work
so long on the Kabaka's Baraza, kisakati, houses, etc.; and among other
things, they had to build a road for His Highness to cycle on. Then they
are ordered off to clear roads, then to build a church, or school, and in
addition to all this probably have to work a few months for the chiefs."38
As a result of his experience during this tour, Tomkins was convinced that
there "is no doubt about it the oppression by the chiefs in some places, is
This oppression was not experienced by the bakopi in labour and tax
collection alone. It featured in the judicial aspect as well. First, according to
the report by Tomkins, chiefs were in the habit of illegally demanding two
rupees before they heard cases from the bakopi. Secondly, the bakopi were
often punished with fines out of proportion to the offence committed. Most
158 UGANDA JOURNAL
of the minor chiefs had no regular salaries so, they levied fines "for the slightest
offence, and the unfortunate Mukopi has to pay, and hold his tongue."
Thirdly, the bakopi sometimes had to pay fines indirectly for offences they did
not commit. In this respect, one case was brought to Tomkins during his tour.
"A minor chief was fined 'so many goats' for not doing certain road
cleaning; he, in turn, fined the next man to him the same number of
goats, for not doing this work; he again fined the next man, till it came
down to the lowest chief who fined the unfortunate bakopi, and they
could do nothing."
Perhaps, more startling is a report to the effect that the bakopi seemed not
to have been free from oppression in the hands of even Christian missionaries
who, as humanitarians, were expected to be the champions of the oppressed.
Tomkins was positive in his report that "there were bitter complaints about
the work they had to do for the Missionary Society." He cited a case "Where
a man had paid his hut-tax, and his rent to the chief, and he was then working
on a church building, where he had been kept for two months"- The victim had
to appeal to Tomkins "to know how much longer he would have to work
before being allowed to return to his home." Amazed by these revelations,
Tomkins had to confess that a "Sub-Commissioner learns very little of the
working of his province by sitting at Kampala hearing the version of what is
going on in the counties from the Regents." It was probably in order to
correct this that more inspection tours were organised with a view to ascertain-
ing bakopi grievances.
One such tour undertaken in July 1905 by an official, Mr. V. Manara,
revealed grievances which covered the grounds already included in an earlier
report just cited, and they need not be repeated39. There was, however, an
additional point in the report of Manara's tour undertaken in August the
same year. It was that some chiefs were in the habit of fining their men "for
choosing to pay rent in cash rather than work."40 The bakopi were also reported
to be cheated by the chiefs whenever the former was employed for hired
labour which was supposed to be paid labour. In this connection,
Manara put the case thus: "The chiefs want labour, they agree to pay a certain
sum but when the work is done they always find faults with it and aft er no end
of trouble and delay, the most the bakopi get is half the amount agreed upon."41
As a result of the experiences in this and other officially sponsored ins-
pection tours, the Government professed a change of heart as far as the bakopi
were concerned. It came to be recognized that they had genuine grievances
which should be redressed. As Manara expressed, "Government aim is to
protect the natives and see that justice is done to them".42 To remove some of
these grievances, it was recommended that receipts be given for all fines and
the bakopi be specially told to demand receipts for fines imposed on them;
that a law book be prepared to show what punishments might be inflicted
for stated offences; that receipts be issued for rents by the chiefs to the bakopi
tenants in order to avoid double payment.43
How far this liberal policy was pursued after 1905 and to what extent the
grievances were really redressed is difficult to say. The official records are
silent on the bakopi after 1905. This silence was only indirectly broken in 1912
when there was an attempt to effect a reform in gombolola divisions. The
incident itself showed that even if the 1905 liberal proposals were effected
there were other forms of oppression. An examination of the policy pursued
up to 1912 as regards gombolola divisions will bear this out.
THE BAKOPI 159
Up to that year, as an official D. W. Cooper pointed out, the practice
existed to the effect that when chiefs or other landowners purchased a new
piece of land, they could choose under which 1st class sub-chief it was tobe
administered quite irrespective of its locality. Consequently, it often happened
that "shambas situated in the centre of one division are not ruled by the
chief of the division in which they are situated but arbitrarily placed under the
administration of the chief chosen by the owner of the land."44 This arrange-
ment caused hardship to the bakopi in a number of ways, of which Cooper
noted two. First, "old and sick men living in isolated parts are obligated to
go miles to be examined for exemption from poll tax." Secondly, men who
wished to have their cases tried in the gombolola courts nearest to their plots
could not do so "but must walk miles to their own Gombolola court merely
because the owner of the land does not like his land to be in the division in
which it is really situated". These problems were only resolved in 1912 when
reforms were effected to have coterminous gombolola boundaries.
However, this did not necessarily mean that the ills of the bakopi were
over. In fact, the passing of the "Busulu and Envujo Law" in 1927 showed that
the bakopi still laboured under some hardship after 1912.45
Those who are familiar with the reaction to authority by oppressed peoples
in other areas in colonial Africa will certainly be surprised that for all their
suffering for a decade at least, the bakopi in Buganda did not contemplate
rising in rebellion against those who governed them. Not even the inhabi-
tants of the Buvuma Islands who showed initial signs of resistance to the pay-
ment of hut tax4 can be seriously regarded as an exception to this general
passivity of the bakopi. For, they were quickly frightened into submission by a
clever ruse of threatening to use "imaginary soldiers" for tax collection.47
This apparent lethargy of the bakopi in the face of oppressive rule can be
accounted for by a number of factors at work in Ganda society. First, a general
attitude of docility to authority seemed to have been ingrained in the bakopi.
Social anthropological studies on Ganda society have emphasized the charac-
teristic of that society by which every man of necessity had to obey those above
him in the hierarchy of authority.48 This obedience appeared to be most
marked in the bakopi who, by virtue of having no one to dominate (except
their wives and children) were constantly on the receiving end as far as orders
were concerned. To them order came from the many overlords they had-
from the mwami w'ekyalo (village chief) to the Kabaka. The result was that the
bakopi were accustomed to accepting authority without question, no matter
how harsh. As early as 1905 an administrative officer, S. Ormsby, illustrated
this point by showing the contrast between the attitudes to authority of the
Basoga and the Baganda.
"... the natives here are less servile to their chiefs and are not afraid of
charging one to his face in the presence of an European: such an act
would be almost impossible to a Muganda ...... in Buganda... the
natives are more subject to the authority of anyone placed over
Another administrative officer, S. Tomkins, made a similar observation with
respect to the bakopi of Buganda. In one of his reports, he declared: "it is
seldom you will find one of them has the pluck to stand up and complain
against his chief".50 In fact, it required his own prodding in 1905 tomake the
bakopi voice out their grievances against their chiefs.
There was a second factor at work. Ganda traditional society was such
that chieftaincy was not hereditary, except, in a qualified sense, the Kabaka-
160 UGANDA JOURNAL
ship. All men of ability were, in theory at least, eligible for chieftaincy. Without
doubt, the bakopi cherished the hope of becoming chiefs. All that was necessary
was to show capability and to be so known to the Kabaka or to anyone who
could make recommendations to him. Rising against authority would certainly
spoil a mukopi's chances of ever becoming a chief. It was possible that in the
colonial period, the bakopi still cherished this hope. Unfortunately such a hope
was an illusion. In the colonial period, the acquisition of western education
was, to a large extent, a sine qua non for chieftaincy titles; and the chances of
the bakopi acquiring this qualification were thin, as has been shown in a recent
study.51 As long as the bakopi cherished this hope of becoming chiefs, they could
not contemplate organising a rebellion against authority. Finally, even if some
of the bakopi had wanted to rebel, they would have found it difficult to establish
a leadership. Only a man of ability could organise a rebellion. Given the
structure of Ganda society, any person would rather direct his energies to
becoming a mwami and acquiring kitibwa (honour) than lead a rebellion. In
the absence of leadership, therefore, the bakopi could neither organise nor
carry out a rebellion.
1. Professor Fallers, like other scholars of Ganda society, has warned against any belie
"that the Baganda are (or were traditionally) divided into self-conscious classes or status
groups". (See L. A. Fallers, Social stratification in traditional Buganda, in The King's Men
edited by L. A. Fallers, p. 70). The reason for this warning being the concept that Ganda
society was fluid and that equal opportunity existed for all men of ability to move from
a lower class to a higher one just as misused opportunities could lead to a fall from a
higher class to a lower one. True as this was, the fact is that at a given point in time the
distinction between one class and another could be made. In fact, in the early colonial
period the factors making for the fluidity of the traditional society were considerably
2. Fallers, Ibid, pp. 68-69.
3. Johnston to the Marquis of Salisbury, 12 March, 1900, copy marked confidential 7620 in
Entebbe Secretariat Archives (hereinafter referred to as E.S.A.) of the covering letter,
forwarding the 1900 Agreement to the Colonial Secretary in London.
4. See D. A. Low's account of the Post-Agreement land settlement in D. A. Low and C.
Pratt, Buganda and British overrule, 1960, pp. 107-128.
5. Low and Pratt, Ibid., p. 98.
6. Ibid., p. 98.
7. Ibid., p. 99.
8. Mengo Notes, Dec. 1900, cited in Low and Pratt, Ibid. p. 100.
9. Her Majesty's Commissioner and Consul General (hereinafter referred to as Commis-
sioner) Entebbe to Collector, Kampala, 17July 1901, in E.S.A. A9/1.
10. Deputy Commissioner, Kampala to Commissioner Entebbe, 3 Oct. 1901, in E.S.A.,
11. Low an d Pratt, op. cit., p. 104. Low based his conclusion on Tomkins' report in his letter
to Sadler dated 7 July 1902.
12. Government accepted this resolution in order to encourage the sub-chiefs who had no
regular salaries, to continue to render their services in tax collection.
13. Harry Johnston's memorandum dated 11 October, 1900, Enclosure in Johnston to
Tomkins, 20 Oct. 1900, in E.S.A., A9/1. Rubber, ivory and wild animals were not
acceptable in lieu of cash after 1902.
14. Uganda (Buganda) Province Report for November, 1901 by Stanley C. Tomkins dated
11 December, 1901, in E.S.A. A8/1.
15. George Wilson to Commissioner, Entebbe, 22 August, 1902, in E.S.A. A8/2.The following
references are to this report.
THE BAKOPI 161
16. They went to Kampala primarily for a ceremony to commemorate the Coronation of
King George V, but they also took the occasion to discuss administrative matters with the
Deputy Commissioner for Buganda, George Wilson, who must have warned them of the
consequences of any negligence in tax collection.
17. Commissioner, Entebbe, to Deputy Commissioner Kampala, 25 August 1902, in E.S.A.,
19. Wilson to Collectors in Buganda, 30 September, 1902, in E.S.A. A8/2.
20. Wyndham to Deputy Commissioner, Kampala, 24 October, 1902, in E.S.A., A8/2.
21. Commissioner, Entebbe, to Deputy Commissioner, Kampala, 21 November, 1902,
in E.S.A., A9/2.
22. George Wilson to Commissioner, Entebbe, 22 August, 1902, in E.S.A., A8/2.
23. Stanley Tomkins' Report of a tour through Mawokota, Busuju, Gomba, Buddu, Kabula, Singo;
dated 12 November, 1903; Enclosure in Tomkins to Commissioner, Entebbe, 13 Nove-
mber 1903, in E.S.A. A8/3.
24. George Wilson to Commissioner, Entebbe, 22 August, 1903, in E.S.A., A8/2.
25. Report for February 1904 by Ormsby, Enclosure in Tomkins to Commissioner, Entebbe,
3 March 1904, in E.S.A., A8/4.
27. Buganda Agreement and Laws of Buganda, (Prepared and published by the Lukiko), p. 72.
Copy in File No. 1569 in E.S.A.
28. Ibid. p. 73.
29. Ibid. pp. 76-77.
30. Ibid. pp. 119-120
31. Tomkins to Commissioner, Entebbe, 26 April, 1902, in E.S.A., A8/1.
32. Spire (?) to Commissioner (?), Entebbe, 24, October 1902.
33. F. H. Leakey to Col. Sadler, 23 October 1903, in E.S.A., A8/3.
34. Stanley Tomkins' Report of tour through Mawokota, Busuju, etc. op. cit. The following re-
ferences are to the same report. 'Kisakaties' are fences.
35. Tomkins to Sadler, 18 November 1903, in E.S.A., A8/4.
36. Ormsby to Sub-Commissioner, Kampala, 29 February 1904, Enclosure in Tomkins to
Commissioner, Entebbe, 3 March, 1904, in E.S.A. A8/4.
37. V. M. Manara, Kakumiro Report for January 1904 dated 5 February, 1904, Enclosure
in Tomkins to Commissioner, Entebbe, 17 February 1904, in E.S.A. A8/4.
38. Stanley Tomkins' Report of a tour through the counties of Chagwe, Bugerere, Mruli and Bulamwezi
(Bulemezi), Enclosure in Tomkins to Ag. Commissioner, 23 March 1905, in E.S.A.,
A8/6. The following references are to the same report.
39. See details in Mr. Manara's inspection report for July 1905, dated 1 August 1905, Enclo-
sure in Sub-Commissioner, Kampala to Commissioner Entebbe, 14 August, 1905, in
40. Mr. Manara's Report for August 1905, dated 1 September, 1905, Enclosure in Sub-
Commissioner, Kampala, to Commissioner, Entebbe, 4 September 1905, in E.S.A.,
A8/7. See also his Report for December 1905 in Manara to Sub-Commissioner, Kampala,
6 January 1906, in E.S.A., A8/7.
41. Manara's Report for October 1905 with respect to Bulemezi and Buruli, dated 3 Nove-
mber 1905, Enclosure in Tomkins to Commissioner, Entebbe, 24 November 1905, in
42. Manara's inspection report forJuly 1905, op. cit.
43. Ibid. and Tomkins' Report of a Tour through Chagwe, Bugerere, etc. op. cit.
44. D. W. Cooper to Provincial Commissioner, Buganda, on the subject of Gombolola
Divisions, 2 January 1912, in E.S.A., S.M.P. (Secretariat Minute Paper) No. 2349.
45. The events leading to the enactment of this law are yet to be fully investigated. But one
of the results of the enactment was that it put an end to the hardship of summary evic-
162 UGANDA JOURNAL
tion which the bakopi had suffered hitherto. See the clauses of the "Busulu and Envujo Law"
in Native Laws of Buganda, (Revised Edition, Uganda Printing and Publishing Co.
Ltd., 1941), pp. 48-52.
46. See instances in Report on Uganda (Buganda) for September, 1901 by Stanley C. Tomkins,
dated 3 October 1901 in E.S.A., A8/1 and District Report for October 1901, in E.S.A.,
47. District Reportfor 1901, op. cit. The ruse was this: While in Buvuma Islands to see that the
people paid their tax, Tomkins, according to his own account, "turned on a gramophone
which reproduced the voice of one drilling men, and the bugle calls, which sounds some
of them were familiar with", and when he "suggested sending those imaginary soldiers to
collect the taxes if they did not start bringing them at once, they cried out "No, don't
send them, we'll bring it". The result, according to Tomkins, was that the people began
48. See The King's Men, op. cit., particularly chapter 5, Leadership, authority and the village
community by M. Southwold and chapter 6, Authority patterns in traditional Buganda
by A. I. Richards.
49. S. Ormsby to Sub-Commissioner, Jinja, Central Province, 2 February 1905, Copy in
50. Stanley Tomkins' Report of a tour through Chagwe, Bugerere, etc. op. cit.
51. T. Watson, A history of Church Missionary Society high schools in Uganda, 1900-1924: The
education of a Protestant elite, Ph. D. Thesis, University of East Africa, 1969.
Uganda Journal, 33, 2 (1969). pp. 163-174.
CONTINUITY THROUGH CHANGE IN THE SOCIAL HISTORY
By REGINA M. SOLZBACHER
Directly south of Kampala's industrial area, east of the police lines and
stretching up the slopes of a low hill lies the community of Kibuli. Kibuli
covers two hundred and forty three acres, eighty of which are taken up by a
large mosque at the crown of the hill and four schools directly below. The
remaining one hundred and sixty three acres on the lower slopes and sur-
rounding low lying area are the site of an unplanned residential area housing
approximately 13,000 Africans. Kibuli is one of Kampala's most rapidy ex-
panding areas with an annual growth rate of 13%. The population of Kibuli
is as diverse as any to be found in Kampala. There are representatives of over
eighty tribes from eleven countries (not including the expatriate school staff).
They speak languages belonging to every major language family in East Africa.
The residents belong to over eighteen different religious denominations and
work at over one hundred different occupations covering a wide spectrum,
although most tend to be manual workers. Although most housing consists of
wattle and daub there are a number of substantial homes scattered throughout
the area. Yet, in spite of all this diversity residents refer to Kibuli as a "village".
The purpose of this essay is to explain, at least in part, how an area with such
an exploding and diverse population can maintain a sense of community.
It is proposed to do this by tracing the history of the community through the
past seventy years showing how the interplay of ecological and socio-cultural
pressures from outside the community and constant structural features and shif-
ting bases of identification within, have allowed for the development and
maintenance of a unique, integrated whole. It is a social history because it deals
not so much with individual events and specific persons as with social processes
of change and identification.
The Founding of Kibuli:
The history of Kibuli begins in 1900. As a result of the Uganda Agreement,
Prince Mbogo, leader of the Baganda Muslim community, was granted
twenty-four square miles of land "for himself and his adherents."I Three hundred
and twenty acres of this consisted of the whole of Kibuli hill. After initial
boundary adjustments, one minor deletion, and one addition, this tract forms
the site of Kibuli today. A land grant of this size within the Kibuga was ex-
ceptional. The Kibuga estates averaged nine acres and many were as small as
three acres.2 Grants were principally made to prominent chiefs whose major
land holdings and principal homes were elsewhere. Their grant in the Kibuga
was for the building of a house near the court, but Mbogo and the Muslims
at that time held a special position in Buganda which made this particular site
appropriate. At the time of the Agreement the earlier power of the Muslims
was on the wane. In 1890 Kalema, the only fully Muslim Kabaka, had been
driven out by the Christians who had reinstated Mwanga. Kalema after
164 UGANDA JOURNAL
fleeing with his followers towards Bunyoro died of smallpox and the Muslims
chose Mbogo to be their Kabaka.
The son of Kabaka Suna, Mbogo (b.1835) had managed to survive both
the slaughter of Suna's sons by Mutesa I and the subsequent slaughter of
princes by Kalema. A man with remarkable ability to maintain peaceful
relations with dissenting others, Mbogo had always been a favourite of Mutesa
and highly respected by Mwanga3. In 1891 the Muslims under Mbogo were
defeated by Mwanga's army, reinforced with Lugard's troops. In the settlement
with Lugard, Mbogo renounced his claim to the Kabakaship and wanted to
stay with his people in the predominantly Muslim county of Butambala. Lugard
though insisted that he must come to live nearby in Kampala4. Mbogo moved
therefore with his followers to Nakasero Hill near where Draper's Department
Store is located today. The Christian missionaries at this time encouraged the
impression that Muslims were more loyal to Mbogo than to the Kabaka and
were distinctly second-class Baganda and not to be trusted.5 Muslim Baganda in
fact were often referred to as distinct from or foreign to other Baganda6 7.
Although Mbogo was not involved in the Muslim uprising of 1893,
the British still felt it wise to deport him to Zanzibar where he stayed for three
years growing to appreciate much of the coastal culture. The year after his
return he showed his continuing loyalty to the British by refusing to co-operate
with the mutinying Sudanese troops who offered him the Kabakaship for his
support. This loyalty to the British never wavered and he in turn was recognized
by them as the temporal and religious leader of the Muslim community.
The high esteem in which the British held him was shown in the central role
Mbogo played in the negotiation of the Uganda Agreement. His mark appears
directly after that of the two regents of Daudi Chwa. In summary, although
Mbogo was respected, it was still felt wise to keep the Muslim leaders under
surveillance but out of the mainstream of power.
After 1900 land at Nakasero became scarce as the European and Asian
communities continued to expand. Mbogo and his retinue moved soon after the
Agreement to Kibuli. What advantages the allotment had in size it lacked in
location. Although the easternmost part of the Kibuga, it was cut off
from Kampala by the Nakivubo swamp and from Mengo by the Kayunga
swamp. To go to Mengo from Kibuli one had to go all the way around the
south of Nsambya. Kibuli was therefore to a large degree physically isolated
from both urban centres. It was not until the swamps were drained in the 1930's
that Nsambya Road, which connects Kibuli with Mengo, could be built and not
until 1950 that Press House Road, connecting Kibuli with the industrial area
and directly with the centre of Kampala, was put in.
A large, traditional Kiganda thatch house was first built on the southern
slope of Kibuli hill, to be replaced by 1914 by a large, wooden two-storeyed
house, with a carved door reminiscent of the coast. This served as Mbogo's
palace, from which, assisted by his Katikiro, Masudi Kisasa, he ran the affairs of
the Muslim community. Basically the house is the same house as that where
his son Badru Kakungulu lives today. Other persons settled round Mbogo
along the southern portion of the area.
The total population of Kibuli at that time could not have exceeded
several hundred for, except for Mbogo's immediate compound, the settlement
pattern was that of the typical rural kiganda village. Isolated houses were
surrounded by substantial shambas. The descendents of many of these early
residents, in fact some of the original residents themselves, live scattered around
the different sections of Kibuli today forming a skeletal, stable, home-owning
Map of Kibuli
(By Courtesy of the Town Planning Department)
166 UGANDA JOURNAL
network among the more recent non-Baganda arrivals.
Although the majority of the persons who gathered about Mbogo in
Kibuli were Baganda there were also some Swahili peoples from the Coast.
Kibuli therefore from the beginning has had a tribally and linguistically
heterogeneous population; but it was a village united around one central
Unity in Land and Politics:
For tax collecting purposes, as a part of the Kibuga, Kibuli was under the
jurisdiction of Apolo Kagwa, the Katikiro of Buganda,8 but since tax collection
was largely done in co-operation with the landlord, this duty was carried out in
fact by one of Mbogo's men. When in 1922 the administrative machinery
for the Kibuga was changed and the capital was divided into mulukas, Kibuli
Estate plus the twenty-five acre allotment to the northeast made to the Katikiro
of Buganda, were made into one muluka and designated Mutuba VI.
Parish chiefs were selected by the ministers and the Omukulu we Kibuga
and presented to the Kabaka for approval. The assent of the landlord was
necessary before a name could be put forth. In Kibuli one of Mbogo's men,
Asman Kikirengoma, was appointed and remained chief until 1939. He was
then replaced by chief Kavuma, a native of Kibuli, raised at the palace, who had
been working for the Muslim community. Kavuma remained chief until 1968
when on his retirement he was succeeded by his former clerk. Although the
headmen who assisted the chief were changed with greater frequency and select-
ed to represent what was to become a much more diversified population, the
political continuity of Kibuli is truly remarkable for an area within the Kibuga.
This continuity of political leadership was largely due to the fact that
Kibuli Estate has remained for the most part in the hands of one owner.
Up to 1931 the original land grant made to Mbogo remained intact. But in 1931
forty acres at the northwest corner were transferred to the Governor of Uganda
to be used largely for the extension of the Police Lines and the building of
Public Works Department housing. Aside from this transfer and the gift of
eighty acres by Kakungulu for the site of a mosque and schools, less than twenty
acres of the original allotment has since been transferred to only six individuals.
One plot was transferred in 1938, one in 1940, one in 1942, one in 1953, and
two in 19579. The retention of such a large block of the Kibuga Estate which is
still over one hundred acres is highly atypical and contrasts sharply with
the situation found by Southall in Kisenyi and Gutkind in Mulago.lo It has
resulted in a high degree of stability of land usage and has allowed the stabili-
zation of a growing population. Except for the clearing of land for the extension
of the schools in the early 1960's and the building of oil tanks in Wabigalo
in the 1950's, large segments of the population have never been forced to move
and even in the above mentioned cases many were relocated, through choice,
The importance of the relationship between landownership and political
continuity is clearly seen in the question of boundaries. The boundaries of the
Kibuga became stabilized with the Uganda Agreement; but inconsistencies
between maps and administrative records, especially with regard to the eastern
boundary, appeared into the 1930's. In spite of the fact that the survey of the
Kibuga in 1907-1908 showed the Kibuga to cover an area of twenty square
miles, the same size that it was when dissolved, the Land and Survey Depart-
ment maps do not include Kibuga Block 15 (Nsambya and Kibuli-Wabigalo)
until the 1940's. Yet administrative records show Kibuli to be part of the
When the Kibuga was divided into mulukas, Kibuli Estate and Kisugu
Estate, which cover a major portion of what is now called Wabigalo, were
designated as one muluka; but the chiefs of Kibuli have never collected taxes in
Wabigalo. Rather they have used the old boundaries of Kibuli Estate as a
marker. Taxes in Wabigalo have always been paid to the neighouring gom-
bolola. Residents of Kibuli Estate and Wabigalo do not consider Wabigalo
part of Kibuli.
In 1931, when the forty acres at the northwest corner of Kibuli became
part of Kampala, an Acholi residential section at the top of the ridge became
part of the Kampala municipality. Yet this section has always paid taxes in
Kibuli and therefore to Mengo. The fact that technically this is not part of
Kibuli or of the Kibuga, has always been recognized and the quarter has always
been largely administered by its own headman working under the chief of
Kibuli. Kibuli Estate and Kibuli muluka have become synonymous in practice
if not officially.
Kibuli as an Educational Centre:
In 1921 Prince Mbogo died and was succeeded both in Kibuli and in his
leadership in the Muslim community by his then fourteen year old son, Badru
The following year brought a change to Kibuli which was to have a major
impact on the community. This was the founding of the primary school.
When young Badru was selected to succeed his father as leader of the Muslim
community the British administrators, as well as the powers in Mengo, were
concerned that his education had consisted almost entirely of Arabic and Islamic
studies. Up to that time Muslim schools offering Western-style curricula
were non-existant, for all the early schools were Mission supported and since
the Muslims had no missionaries in the formal sense they had no one to found
schools for them. Koranic schools of course could be found through Buganda
of which the first had been founded in Kibuli already before 1914.
In 1922 the Lukiko founded a primary school in Kibuli thereby inaugu-
rating Muslim secular education not only in Kibuli but in Uganda. There were
only three students at first, Badru Kakungulu, H. Goloba and A. W. Ssimbwa
under the tutelage of Luka Sajjabi. They met in a small mud structure. More
students were admitted and soon there were over 40 pupils. The school was
known as Kiwotoka schoollI. By 1925, when Prince Badru went on to King's
College Budo, the school had developed enough for a burnt-brick building to be
erected. The students at this time were by no means all children. Ages ranged
from those under ten to men over forty. Although all the students were not
Muslim, preference was given to Muslim applicants. Students came in daily
from outside Kibuli as well as from the community. The founding of the Pri-
mary School marked the beginning of Kibuli as a centre for Muslim education.
Several events helped bring this about. In 1937, on the occasion of his Golden
Jubilee, the Aga Khan, leader of the Ismaili community, gathered together the
prominent Muslims of East Africa and formed the East African Muslim Wel-
fare Associationl2 and Kakungulu became the Vice-President of the Association.
Largely through funds from the prosperous Ismaili community, this organi-
zation has been able to contribute financially to the development of
Muslim education. The same year Rahmadan Gova, a former classmate of
168 UGANDA JOURNAL
Kakungulu at Kibuli Primary School, returned to Kibuli School to teach.
Seeing that the government would help sponsor Muslim schools only when the
religious community formed a registered educational association, and encourag-
ed by a number of Muslim leaders, he formed the Uganda Muslim Education
Association. Seven prominent Muslim leaders served as trustees. It was regi-
stered in 1940.
Later Kakungulu gave eighty acres at Kibuli as the site for a large Juma
mosque and for the location of schools. This grant consisted of the top and the
upper slopes of the hill. Thus Kibuli was not an allotment to the Muslim
community such as Nsambya and Rubaga had been to the Catholics and
Namirembe to the Protestants. Rather it was a grant to an individual. The
Muslims had not been allocated land in 1900 and this deficiency was not
remedied until 1913 when Mbogo requested such land on which to build
six hundred mosques throughout Uganda. Therefore Kibuli was really at first
a residential centre of the Muslim community. Seeing the need for land for
schools, and the desirability of the site being located in or near town, prompted
Kakungulu to make the grant.
The new mosque was started almost at once. Its construction was largely
financed by the Ismaili community with the Aga Khan laying the foundation
stone in 1941. Its completion was delayed by the war and it was finished only in
1950. It serves now as a marker in Kampala comparable to the cathedrals at
Namirembe, Rubaga and Nsambya. To the residents of Kibuli, both Muslim
and non-Muslim, it serves as a source of great pride.
1945 saw the inauguration of the first year of secondary school education at
Kibuli. With the subsequent opening of a hostel, Muslim students from different
parts of Uganda came to board at the Junior Secondary School. A Grade C
Teacher Training College at Kasawa, begun by the Buganda Government as a
model school in 1932-3, was given to the Muslim community who moved it to its
present site in Kibuli where it became the Kibuli Muslim Teacher Training
Collegel3. By 1960 Kibuli Junior Secondary School became a Senior Secondary
School. Soon after independence, a USAID grant was given to build a new
school plant and staff housing for the secondary school and these were finally
completed and first utilized in 1966. Further building has recently been started
to accommodate the Higher School Certificate programme started in 1968.
The staff buildings for the secondary school and the TTC staff, house the only
non-African residents of Kibuli. In 1966 Lubiri Secondary School, whose
premises had been destroyed in the fire in the Lubiri, moved temporarily into
the former Kibuli Secondary School buildings.
The last fifty years has seen Kibuli become an educational centre with
two secondary schools, a teacher training college and a full primary school.
The total enrolment of these schools is 1,721 (1,156 without the Lubiri School).
Looking up the slopes from the village the schools have served as a source of
great pride to the residents of Kibuli, most of whom unfortunately cannot
afford to send their children to any but the primary school. Even here since the
language of instruction in the early primary years is Luganda, the student body
is principally Baganda and western Ugandan.
With the overcrowding of the Secondary School hostel a number of
students have been forced to take up lodging in the village where an even larger
number of students attending various Kampala day schools have also taken
advantage of cheap local housing. Together these young people form a very
prominent minority of the Kibuli population and along with the growing
Africanization of the school staffs in the past few years, is closing the gap
somewhat that has existed between the villagers and the residents of the school
The Growth of Kibuli's Population:
During the early years, Kibuli village grew slowly. Because of limited
educational opportunities for Muslims, there was little inducement for Buganda
Muslims to come to town to settle. The uneducated Muslim farmer could do
better in the rural area growing cotton as a cash crop than he could as a manual
labourer in Kampala. Considerable numbers of migrants from Toro, Bunyoro
and Ankole had been coming to Kampala as early as 1900 in search of work.14
A few of these migrants started filtering into Kibuli. Allowing migrants to settle
amongst them was not unusual for the Baganda. Traditionally the Baganda
absorbed members of conquered tribes and have never objected to migrants
settling near their rural homes, especially if the immigrants were willing to
work for them. The western Ugandans who settled in Kibuli furthermore were
easily integrated into the community since they spoke languages mutually
intelligible with Luganda and shared many cultural patterns. Most, though,
were not Muslim. Although the newcomers settled scattered throughout the area
the densest settlement in Kibuli at the time was along the road running past
Mbogo's home and on to Nsambya.
During these early years Kibuli was principally still an area devoted to
agriculture; but after 1920 as Kampala continued to grow and its ecological
pattern gained in complexity, Kibuli began to attract a new type of resident.
Several events occurred at this time just outside Kibuli which later were to be of
great importance for the development of the community. In 1915 the Port Bell-
Kampala railway was opened. It passed directly north of Kibuli. The next
year the Kampala Municipality purchased a stretch of land directly west of the
Kayunga swamp in Nsambya for the building of police lines.
In 1930 the town planner Mirams presented a plan for the development of
Kampala which has since been largely implemented. Foreseeing the coming of
growing industry he suggested locating the industrial estates directly north of
Kibuli. This was a logical site since the Kampala-Jinja railway was to be open-
ed there, parallel to the Port Bell line, the following year. He further men-
tioned that Kibuli, because of its location, (although he did not mention Kibuli
by name) should be used for building houses for working class people.15
During the period between 1920 and 1940 population growth continued to
be slow. In the 1920's the founders of what has since come to be known as the
Acholi quarter arrived in Kibuli. Locating themselves at the northwestern
point of Kibuli, the point nearest to downtown Kampala, they formed at that
time an isolated, alien pocket in the more sparsely settled section of the village.
There were two men originally who had come to Kampala looking for work.
They were joined three years later by three others who had been transferred
for the location of the railway from Jinja. It was only in the 1930's when that
particular strip of land became part of Kampala municipality that the sector
really began to grow large enough to be considered a quarter and only in the
past ten years that it has become so incredibly overcrowded. Today it is still
inhabited completely by Acholi and two of the original five settlers are still
In the 1930's the Wabigalo section of Kibuli was also beginning to build up.
First farmers from western Uganda, Banyankole and Banyoro principally and a
few Batoro, set out shambas. They were soon joined by several Abasamia from
southeastern Uganda who came with the railway. The population of Kibuli
was definitely becoming tribally heterogeneous, but still was dominated by
peoples from the interlacustrine kingdoms.
Along with the growth of population came markets, shops and bars to meet
the needs of the people The antecedents of both Kibuli and Wabigalo markets
were founded by Kibuli residents by the 1940's although both have since been
moved from their original sites. Some of the founding vendors are still there.
At first principally matoke was sold but as the population became more diver-
sified other commodities were added. There were also several dukas and bars
operating in Kibuli then which are still functioning today. By 1948 business
was brisk enough for Kibuli to be mentioned along with Katwe, Wandegeya
and Kisenyi in a complaint made by the Asian businessmen of Kampala mu-
nicipality to the town clerk, that the African merchants of the Kibuga were
In 1948 the census noted Kibuli to have a population of 1,026 of whom
less than 100 were of foreign origin. It was the year 1949 that opened the period
of most rapid growth in Kibuli's history. Once again the European and Asian
communities needed room to expand and it was decided to develop Kololo
hill.17 Kololo hill had become a part of Kampala Township already in 1916
but it had never been developed. Originally the site of a Nubi camp, its southern
slope was later the base of labour camps and temporary squatters' homes.
With the clearing of the hill in the late 1940's a large number of migrant labour-
ers, especially Banyankole, were rendered homeless and many shifted residence
to Kibuli. Kakungulu's sister offered them land in the southern section of Kibuli
and a very crowded colony of tiny thatched houses was built. This was named
Mululembo, the Lunyankole word for a small town, and formed an island
of extreme density, in sharp contrast to the much more spacious settlement
pattern around it. Other migrants scattered throughout other sections of
Kibuli. A number squatted on land which had been set aside for the schools.
Their homes were generally distinguishable by the very poor state of construc-
tion. For these residents were mostly porters who looked on Kampala as a
temporary place of residence. Almost none of the Banyankole migrants of the
1940's live in Kibuli today. They have been replaced, though by a constant
stream of Banyankole of similar economic status.
Another ethnic group that started to move to Kibuli at this time, but with
much more sense of community commitment, were the Nubi. Although Nubi
settlements could be found in the Kampala area since the turn of the century,
they did not start to filter into Kibuli until the 4th Battalion of the King's
African Rifles was moved from Bombo to Jinja in 1939. The number of Nubi
migrants from various locations increased significantly in the 1950's when
the mosque was built. For the Nubi are the only immigrant group in Kibuli of
any size that are Muslim. Originally Malikis, since Islam reached the Sudan
from the north, the Nubis, in years preceding their coming to Kibuli, became
Shafiis in common with the coastal influenced Baganda. A number of these
Nilotes have intermarried with Muslim Baganda and become very Gandanized.
Yet Baganda women marrying into the Nubi community are in turn strongly
influenced by Nubi ways. Although the Nubi often moved next door to other
Nubi, they have never formed large ethnic clusters in Kibuli and can be found
today scattered throughout the community just as the Baganda. They are the
only immigrants who in significant number own their own homes. In several
cases where owned by prosperous businessmen, these are substantial block
In 1950 a murram road was opened between the expanding industrial
area and Kibuli, serving as a by-pass for traffic from Jinja going to Entebbe or
Mengo. Three years later Press House Road was opened, and the Public
Works Department put up staff housing at the bottom of the hill on the
stretch owned by the municipality since 1931. These houses face the town, in
contrast to the houses of the village at the top of the slope which all face into
The increase in traffic coupled with the building of the Police lines right up
to Kibuli's western boundary led to the rapid expansion during the late 1950's
of the depot area. Shops, both dukas and service facilities, began to line both
sides of the road. The market was moved south where there was more room.
Bars, serving both bottled and native brews, multiplied as the police came to
Kibuli in their off hours. The Acholi quarter grew, blending, at least geograph-
ically at this point, with the surrounding areas to become one of Kibuli's two
largest drinking centres (the other being Wabigalo).
By 1959 the census reported Kibuli's population at 3,767. From 1948 the
rate of growth had been 12.8% per annum. The area of greatest growth
during the 1950's had shifted from the southern portion to the northern section
near to Kampala as new migrants moving to Kibuli in increasing numbers
were looking for housing close to the industrial area.
The late 1950's saw the plots in the industrial estates filled and in 1962,
as independence neared, demands for further plots resulted in the extension
of the estates eastwards in the direction of Bugolobi. Kibuli, by reason of its
location, became the factory workers' residential area that Mirams had men-
tioned. But rather than the construction of neat, permanent houses built by the
municipality or industry, the quarters to be rented were built almost entirely by
the house-owners of Kibuli. Where shambas of matoke, cassava and coffee had
formerly surrounded individual dwellings, blocks of mud houses were built,
often only several feet from the house-owners own residence. The symbiotic
relationship between owner and renter was, and still is, a face to face relation-
ship in most cases.
This period was marked by the entrance of a large Kenyan population,
principally Luo at first, later Abaluhia also, into Kibuli. The Luo had already
formed a colony in neighboring Namuwongo in the 1940's and have always
favoured the section east of the Kibuga and towards Port Bell as their place of
residence. Slowly in the early fifties, they began to inch their way up towards
Wabigalo. By 1952 there was a Luo malwa bar found there although many of
the patrons came from Namuwongo. By the late 1950's the trickle of Kenyans
joining the Abasamia in Wabigalo had turned to a torrent and by the 1960's
into a flood making Wabigalo into one of the most densely populated areas of
Kampala. In the past five years they have started to spread into every part of
Kibuli and up into Kisugu.
Although the Luo, Abaluhia, Acholi, Nubi, Banyankole and Baganda make
up the largest portion of Kibuli's population, simultaneously with this influx
was the filtering in of smaller numbers of representatives of a wide variety of
other tribes from all over East Africa. This diminished somewhat the impact of
each of the more numerous groups.
The exploding population of Kibuli has brought changes not only in
density, but also in the ecological distribution of the various ethnic groupings
that comprise the population. With the clearing of land in the early 1960's for
the building of the Secondary School, about one hundred thatched houses
were torn down. They were inhabited mostly by Banyankole. A number of
172 UGANDA JOURNAL
them shifted over to the north, just a few hundred feet from their former homes.
Here they are mixed with members of other tribes especially Banyaruanda
Mulelembe, originally a Banyankole settlement, was infiltrated by Nubi in
the 1950's as some of the Banyankole moved to the less crowded area around it.
The Nubi have stayed, but in the last three years most of the remaining Banyan-
kole have departed and been replaced by Abaluhia from Kenya. In the depot
irea many of the Baganda, who did not own businesses situated there, have also
chosen to move to the now less crowded southern portion of the community.
They have been replaced principally by peoples from northern Uganda
such as Alur and Acholi. Since the Acholi quarter is now contiguous with the
depot area this tapering off of Nilotes from the north into Kibuli has helped to
bind the Acholi quarter somewhat more into the mainstream of Kibuli. Popu-
lation movement in Kibuli through the years has not consisted only of immigra-
tion and emigration from the community, but has also involved considerable
movement within the community as residents try to find optimum conditions'
within the setting of Kibuli.
The picture though should not be held of ethnic hordes roaming through
successive areas of the community. It is individual residents or nuclear families
wishing to live closer to relatives or friends that move when a room becomes
vacant. Fifty-four per cent of the present residents have relatives living in
the community and the vast majority of residents state that their best friend
lives in the community. Therefore residents often perceive themselves as living
near friends and relatives, not necessarily near fellow tribesmen.
Although some ethnic groups can be found living in clusters, there are
always other members of such groups living scattered elsewhere in the com-
munity and only the Acholi quarter is tribally homogeneous. Consequently it
is the only neighbourhood in Kibuli known by an ethnic name. All other
neighborhoods are named either after one of its more prominent residents,
after the landlord, with reference to something it is located near (such as
schools, police lines, boundary) or the name is of unknown origin.
Kibuli residents today consider it poor manners to ask someone of his tribe,
or religion or nationality. These are considered potentially divisive factors which
are to be played down. Furthermore religion, nationality, linguistic grouping,
and tribe each divide the population differently and often the range of possible
affiliation within each criterion has become so wide that confusion as to a neigh-
bour's ethnic affiliations is often quite real. Muslims though are somewhat of an
exception and are given a degree of deference within the community because
Kibuli is considered "theirs" as the schools are considered Muslim schools
although a significant number of non-Muslims attend. No-one though would
state that the majority of Kibuli's population are Muslim. As a unit they
typify themselves on a status criterion referring to themselves as "poor people",
"working people", or "the common man". But since Muslims in the larger
Ugandan context are considered less educated and therefore less successful than
their Christian brothers there is a considerable overlap in reality in these analy-
tically distinct evaluations.
In April, 1968 Kibuli became part of Kampala Municipality. It had in
1969 a population of about 13,000. In the past decade it has grown at a rate of
13.2% per annum. In the strict usage of sociological terminology it can no
onger, if it ever could, be considered a "village". Yet it has been stated, this is
the term used by the residents in describing their home. Thus, in spite of the
exploding population and incredible heterogeneity, a sense of community has
been sustained. Through Kibuli's seventy year history certain pressures from
outside the community can be traced and structural constants and shifting
bases of identification from within can be identified. This has created condi-
tions favourable for the development and maintenance of a sense of community
Ecological factors, external to the community have exerted strong pres-
sures for population development and change. The expansion of early Kampala
as a European and Asian community gave the push that brought the original
inhabitants to Kibuli and attracted the first large wave of migrants in the late
forties. The existence of the large swamps prior to the 1930's isolated the early
community from the Ganda capital reducing its value for the building of
expensive homes. Its early isolation from Kampala meant it was not desirable
for early expansion of the town and allowed the community an introverted exi-
stence in its founding years. The drainage of the swamps and subsequent loca-
tion of the industrial estates made it an ideal location for the homes of workers
who could afford little for transport. Kibuli's location at the easternmost bound-
ary of the Kibuga furthermore made it a more attractive area for non-Baganda
migrants. Sociocultural factors outside the community have been instrumental
in shaping Kibuli into the community that exists today. Hostile, or at best suspic-
ious attitudes, of the non-Muslim Baganda, Christian missionaries, and British
officials towards the Muslim population of Buganda brought about pressure to
keep the Muslim leaders in the Kampala area but out of harm's way. Kibuli hill
was a highly suitable site. Neglecting to give the Muslim community an urban
based grant, as those given to the Christian denominations, made Kibuli a
likely location to serve both as a Muslim residential centre and as an admini-
strative and subsequently educational centre.
Land tenure and continuity of political leadership have had a major
impact on the history and style of the community and served as structural
constants through the years. Kibuli is found almost entirely on one land grant-
a grant which has remained virtually intact in the hands of one man and his
successor. Land usage patterns, except for the clearing of homes from the eighty
acres on which the mosque and schools were built, have shifted little. Because of
its size and early ethnic identity Kibuli was designated as a political unit admini-
stered for almost its entire existence by two men who worked closely with the
Although Kibuli has always been tribally and linguistically heterogeneous
the bases of acceptance and absorption into the community have shifted over
time. At first tribal and linguistic diversity were outweighed by religious simi-
larity and loyalty to one man. The condescending attitudes of other Baganda to
the Muslim may have made this religious identification that much more mean-
ingful. The later acceptance of the Nilotic Nubi by the Baganda has also been
largely on this religious basis. The first waves of migrants into Kibuli were
largely from the interlacustrine kingdoms. These migrants spoke languages
mutually intelligible with Luganda and were culturally similar in many respects
with their hosts. Their absorption into the then sparsely settled area was
accomplished with relative ease.
The past educational deficiencies of the Muslims has meant that most of the
Baganda of Kibuli are both educationally and occupationally handicapped and
therefore with reference to the larger Kampala community share with the
newer migrants from the north and Kenya similar economic problems and
174 UGANDA JOURNAL
status. Although status may serve as a new basis of identification for the com-
munity as a whole, the criterion of religion operates not only as a criterion of
elite status within the community but also by its visibility in the mosque and
schools as a symbol of community uniqueness.
Research in Kibuli was conducted by the author while serving as a Research
Fellow of the Makerere Institute of Social Research from December, 1966 to
The major portion of the material on which this essay is based was derived
from oral sources. I would like to thank Badru Kakungulu, Rahmadan Gova,
Chief Kavuma, the Headmasters of the schools of Kibuli and the numerous
long-term residents of Kibuli for their patient assistance. I would
also like to thank the Land and Surveys Department in Entebbe and Kampala,
the Uganda Electricity Board, and the Department of Town and Regional
Planning for their co-operation.
Virtually nothing has been published on the history of the Muslim com-
munity in Uganda. Tracing down those few fugitive references that exist was
greatly facilitated by the co-operation of many of my colleagues at Makerere
University College. Special thanks are due to my assistants, Boniface Kayondo,
Edward Mukanga, Fred Muwonge-Ibanda and Yahaya Simwegerere who
assisted both in collecting information in the field and in the tedious search of
I. Gee, T. W., A century of Muhammadan influence in Buganda, 1852-1951. (Uganda
Journal, 22, 1958, p. 141).
2. Thomas, H. B. and Spencer, A. E., A history of Uganda land and surveys, Entebbe, 1938,
3. Gee, T. W., op. cit., p. 140.
4. Perham, M., Lugard; theyears of adventure 1888-1898, London, Collins, 1956, pp. 316-319.
5. Ashe, R. P., Chronicles of Uganda, London, Hodder and Stoughton, 1894, p. 327.
6. Gutkind, P. C. W., The royal capital of Buganda, The Hague, Mouton, 1963, p. 18.
7. Katumba, A. and Welbourn, F. B., Muslim martyrs in Buganda, (Uganda Journal, 28,
1964, p. 163).
8. Gutkind, P. C. W., op. cit. p. 56.
9. In the past three years long-term leases on a number of small plots along Nsambya and
Mbogo road have been given. Substantial block house are being built on these plots.
A large number of other plots have been surveyed principally along the main roads.
The effects of this on the community cannot be determined.
10. Southall, A. W. and Gutkind, P. C. W., Townsmen in the making. Kampala E.A.I.S.R.,
1957, pp. 31 and 118.
11. Uganda Herald, 23 February 1944, p. 3.
12. Ingrams, H., Uganda: a crisis of nationhood. London, H.M.S.O., 1960, p. 78.
13. Carter, F. The education of African muslims. (Uganda Journal, 29, 1965).
14. Powesland, P. G., A history of migration in Uganda, essay in Economic development and
tribal change, edited by A. I. Richards, Cambridge, Heffer, p. 20.
15. Mirams, A. E., Report on the town planning and development of Kampala. Vol. 1. Entebbe,
G.P., 1930, pp. 92 and 106-110.
16. Gutkind, P. C. W., op. cit., p. 43.
17. May, E., Report on the extension scheme: Kololo-Naguru. Entebbe, G.P., 1947.
Uganda Journal, 33, 2, (1969). pp. 175-199.
THE VEGETATION OF SOUTHWEST KIGEZI
By A. HAMILTON
Southwest Kigezi is an area of great attraction to botanists, both because
of the diversity of the vegetation and because of the magnificence of the scenery.
The area dealt with in this paper is bounded to the west by the Congo border;
to the south, with the exception of the Bufumbira volcanoes, by Rwanda;
to the east by the Ishasha-Kiruruma valley (but including Mafuga Forest)
and to the north by the Kayonza road passing through the Impenetrable
Forest (See map). The Bufumbira volcanoes are omitted from this examination.
Contained within this area is the lava plain of Kisoro, Echuya bamboo forest,
lakes Bunyonyi and Mutanda and the Impenetrable (Bwindi) Forest, to
mention only a few of the more interesting features. The Impenetrable Forest,
being the only large extent of natural forest stretching between 5,000 and
8,000 ft. (1,500-2,450 m) in the area, is unique and of extreme importance
both botanically and zoologically. The present account of the vegetation is
based on comparatively few observations and must be regarded as very pro-
visional. It is hoped that it will provide a focus for criticism and a basis for
Geology and geomorphology. Geologically the part of Kigezi dealt with here
falls into two areas, the Precambrian and the volcanic. In the former, which is
the most extensive, the rocks are NNW-folded mudstones, shales, phyllites
and occasional bands of quartzite belonging to the Karagwe-Ankolean System.
Intrusive granite is exposed beneath Ruhuhuma swamp (Combe and Sim-
mons 1933). Upwarping associated with the western rift valley has lifted these
rocks to form the Rukiga highlands. In the southeast of the area, river-systems
have developed along the strike of the rocks and drain predominantly NNW
towards lake Edward. In some places cross-strike rivers connect adjacent valleys,
as between Ruhuhuma swamp and lake Mutanda. The valleys usually lie
between 6,000 and 6,500 ft. (1,830-1,980 m) and the ridges between 7,500
and 8,000 ft. (2,290-2,450 m), connected to the valleys by steep-sided slopes,
often of over 15 degrees. The two main ridges in the southeast are the Echuya
ridge, between the Bufumbira bay (which will be described later) and the
Bunyonyi-Ruhuhuma valley, and the Kackwekano ridge between Bunyonyi
and the Ishasha valley.
Further north, in the area covered by the Impenetrable Forest, the topo-
graphy is more rugged and much more dissected, presumably because it has
been subjected to more recent uplift than other parts of the Rukiga highlands.
In the south of the Impenetrable Forest the highest hills are over 8,000 ft.
(2,450 m) and the valleys about 6-7,000 ft. (1,830-2,150 m), but in the north
the country is lower and the hills somewhat less steep. The drainage is mainly
by the Mbwa and Iri rivers and their tributaries towards the northwest.
The volcanic area of Kigezi lies mainly in the extreme southwest, with
an outlier at Muko near the northern end of lake Bunyonyi. The geology has
been dealt with in detail by Combe and Simmons (1933). It is part of the large
00 54' S
S -f142* kA 7
k BUFUMBIRA BAY Echuya For
+ Bufumbira X V +
(177)* Volcanoes X -
- + -, ++ ,,
Figures give mean annual rainfall in cms. (those in brackets for less than six year's readings).
VEGETATION IN KIGEZI 177
Virunga volcanic area, most of which lies in adjacent parts of the Congo and
Rwanda. Some of the volcanoes in the Congo are still active, but, in Uganda,
the only indications of the former volcanism are hot springs and occasional
earth tremors. The most striking features of the Kigezi volcanic area are the
three big Bufumbira volcanoes, Muhavura (13,541 ft.) (4,128 m) Mgahinga
(11,400 ft) (3,475 m) and Sabinio (11,690ft) (3,654 m), which lie on an east-
west line along the Rwanda-Uganda border. These mountains, which rise
very steeply from a comparatively flat plain, known as the Bufumbira bay,
consist largely of lavas, which have flown down and filled the plain. After the
main volcanic activity had subsided, there appeared numerous subsidiary
vents, releasing large quantities of ash, and this has blanketed many of the
older lava flows (as well as some of the surrounding Precambrian rocks) so
that today the Bufumbira bay consists of an intricate pattern of lavas and
Soils. The soils of Kigezi have been described in a memoir of the re-
search division of the Department of Agriculture (Harrop 1960). As far as the
vegetation is concerned, the soil properties of greatest importance are pro-
bably depth, soil moisture, nutrient status and erodibility. The soils are unduly
shallow only over the slow-weathering lavas of the Bufumbira bay.
In Kigezi, soil-water availability seems to depend mainly on three factors,
the nature of the parent material, the quantity of humus in the soil, and the
topography. The best example of the effect of parent materials is in the Bufu-
mbira bay where the high porosity of the ashes and lavas has resulted in a
virtual absence of permanent streams. Humus is able to hold much more
water in an easily available form than clays. In southwest Kigezi it is every
where high, but even so differences in the quantity of organic matter in the soil
have been found to be important in the distribution of different vegetation
types at Kackwekano Farm (Chenery 1951). Topography is also important in
controlling the distribution of soil water, lower slope soils being not only
moister but in some cases also richer in exchangeable bases. Such ionic in-
creases have been recorded from Kackwekano but not from the Impenetrable
Forest (Leggat and Osmaston 1961). They are not considered to be as impor-
tant as soil-water, as far as the vegetation is concerned either in Kigezi (Che-
nery 1951) or Ankole (Johannesen et al. 1967).
In Uganda, the base-status of soils is usually correlated with pH: the
lower the pH, the lower the concentration of cations. All soils developed over
Precambrian rocks in Kigezi are either moderately of very acid. The lowest
pH recorded (2.9) and indeed the lowest ever encountered in montane soils
by Chenery during wide experience in the tropics, was found in the highest
soil horizon of the Impenetrable Forest. Another unusual feature of the Impe-
netrable Forest soils is their clay fraction, which consists of dioctahedral mica,
a mineral which has not been recorded as the dominant clay fraction in soils
from anywhere else in the world. Both the base status and pH of soils on the
volcanic rocks are higher, but these are sometimes deficient in available
All the soils in the Precambrian area, with the exception of some found
near Kabale, are very susceptible to erosion when the natural vegetation cover
is removed. The necessity for soil conservation has led to the construction of
terraces, which now form a characteristic feature of the Kigezi highlands.
Climate. Measurements of temperature over a long period of time are
only available from Kabale meteorological station (East African Meteorolo-
gical Department 1967). Here, at 6,138 ft (1,871 m), the mean annual tem-
J F M A M J J A S N D
Fig. 1 Mean monthly rainfall:
Kabale (1919-1967), Bunyoro (1951-1967) and Kisoro (1940-1967).
Note the low figures for June, July and August.
VEGETATION IN KIGEZI 179
perature is 16.70 C. There is little deviation from this value through the year;
the daily temperature variation from about 100 to 230C being much more
marked. The lowest temperature ever recorded is 2.80C. Although no other
figures are available, there is no reason to suppose that mean annual tempera-
tures do not follow the normal East African temperature/altitude relationship,
which gives a lapse rate of 1.980C for every thousand feet (J. A. Channon-
personal communication). However, on top of this overall regime, marked
local variation is caused by the topography, in particular by cold air accu-
mulating in valleys at night. This cold air is usually saturated with water
vapour and the mists formed in this manner are not generally dispelled until
sunrise. The blanketing effect of this air is not felt on ridges, where the
level some hours after of out-going radiation at night must be considerable.
Rainfall stations in southwest Kigezi are rather too widely scattered to
give more than a general picture in an area of such diverse topography. Mean
annual values from these stations are shown on the map, on which it can be
seen that rainfall increases both towards the southwest and with altitude. As
far as plants are concerned, total rainfall is often not as important as its dis-
tribution. Generally speaking more water accumulates in the soil if the rain is
lighter and more protracted, whilst run-off and evaporation are both greater
if the same amount of rain falls in short sharp showers. Kigezi, where the former
conditions are more frequent, is fortunate in this respect, but this is more than
offset by the marked seasonality of the rain. There are two dry seasons, one
in December-January and the other in June-July (Fig. 1) and the latter in
particular can sometimes develop into a pronounced drought (Atlas of Uganda
1962). The effect of this dry season on the vegetation must be considered in
the light of the soil-moisture properties in the different areas. Thus, in volcanic
areas and on upper slope sites, even in the highest rainfall areas, water-stress
becomes seasonally critical.
Precipitation and temperature are certainly the climatic variable of
greatest importance for plants in the area. Wind speeds are usually low.
Wind-roses and measurements of evaporation and sunshine are only available
for Kabale (East African Meteorological Department 1967).
Forested Regions on Precambrian Rocks
In the area in question, there are three large blocks of forested country,
Impenetrable, Echuya and Mafuga forests, all of which are Central Forest
Reserves. In addition there are several small patches of secondary forest,
particularly close to the Impenetrable.
(a) The Impenetrable Forest
The rugged topography of the Impenetrable Forest, combined with its
comparative inaccessibility, makes it one of the least known of Uganda's
forests, but it has been examined from aerial photographs and mapped by
Cahusac (1958). The most widespread types south of the Kayonza road are
Types 2, 4a, 9, 10, 13a, 14, 16 and 18.
180 UGANDA JOURNAL
Forest types in the Impenetrable Forest (After Cahusac).
Number Area Rough altitudinal range
2 North and west .. 5,000 6,000 ft (1,500 1,830 m)
4a North 5,000 6,000 ft (1,500 1,830 m)
9 .. Centre 6,000 7,500 ft (1,830 2,290 m)
10 South 7,000 8,000 ft (2,150 2,450 m)
13a South and west 6,500 7,000 ft (1,980 2,150 m)
14 .. Centre .. 7,000 7,500 ft (2,150 2,290 m)
16 .. Southeast .. 6,500 7,250 ft (1,980 2,220 m)
18 Southeast .. 7,000 8,000 ft (2,150 2,450 m)
21a Southeast .. About 8,000 (about 2,450 m)
In this analysis Type 2 is Poor hill forest; Types 4a, 9, and 13a are Rich mixed
forest; Types 10 and 18 Medium mixed forest; Type 14 Piptadenia mixed
forest and Type 18 Chrysophyllum dominant forest. The species recorded from
those later visited by Cahusac on the ground are given in Table 1, where
Type 21a, the uncommon but important bamboo association, is also included.
Most of my own observations have been made near Ruhiza, on the eastern
side of the forest. The area falls mainly into Type 18 on Cahusac's map. Here
the ridges lie at about 7,500 8,000 ft (2,290 2,450 m) and the valleys
between them at about 7,000 ft (2,150 m). Many rivers drain into a large
swamp, Mubwende. For classifying the vegetation the most natural system
appears to be one related to topography. Hilltops, ridges, hillslopes, gullies
and valleys all have different vegetation, each with characteristic structure
and species-composition. The preferences of certain tree species for these sites
is shown in Table 2 and sketches of some forest profiles in Fig. 2. Photos 1 3
are views of three fairly typical forest types.
In hilltop forests, which are rare, the trees are stunted and frequently
multi-boled and the canopy, which is closed and even, is only about 10 m tall.
The most common tree species is Macaranga kilimandscharica with Olea hochs-
tetteri also abundant. There are few understorey species, those present being
mainly in the family Rubiaceae. The ground flora is open and species-poor,
with much Mimulopsis solmsii and several species of ferns.
Ridgetop forests are common. They are taller than hilltop forest, have a
more luxuriant understorey and a greater number of species. The canopy is
usually at about 21 m and closed. In places, Faurea saligna forms almost pure
stands, and amongst the taller trees, Podocarpus milanjianus. Polysciasfulva, Olinia
usambarensis and Rapanea rhododendroides are also common. Macaranga kilima-
ndscharica, Maesa lanceolata, Psychotria megistosticta and Rytigynia sp. are some of
the more frequent understorey trees and there is a rich shrub layer up to 3 m
tall, in which the most abundant species are Cyathula cf. uncinulata, Mimulopsis
solmsii and other Acanthaceae and, less commonly, Pteridium aquilinum and
Cluytea sp. In ridge-top forests, as well as in most other forest types, the climber
Urera hypselodendron is abundant.
In hillslope forests the canopy is higher, the tree trunks closer together
and the number of tree strata greater and better defined. Chrysophyllumfulvum
is often the most abundant upper storey tree. In one forest near Mubwende,
the tallest trees are Chrysophyllum and Olea hochstetteri, both reaching 27 m,
with a dense smaller tree layer about 18 m tall of Cassipourea ruwensorensis,
Allophylus macrobotrys and Drypetes aff. gerrardii. There are numerous under-
storey trees, mostly Rubiaceae. In the thick shrub layer, about 2 m tall,
Cyathula cf. uncinulata is the most abundant species, becoming semi-scandent on
Albizia gummifera X
Alchornea sp. .. X
Arundinaria alpine ..
Bersama abyssinica ..
Bosqueia phoberos ..
Carapa grandiflora ..
Cassipourea sp. ..
Chrysophllum spp. .. X
Fagara spp X
Faurea saligna .. X
Ficalhoa laurifolia X
Hagenia abyssinica ..
Harungana madagascariensis X
Ilex mitis ..
SPECIES-DISTRIBUTION IN THE IMPENETRABLE FORESTS
(FROM CAHUSAC 1958]
4a 9 10 13 14
16 18 21a
X Myrianthus arboreus
Myrica salicifolia ..
X X N.eoboutonia macrocalyx
X Olea sp.
X X Olinia usambarensis
X Pittosporum spathicalyx
X Podocarpus milanjianus
Pygeum africanum ..
X X Rapanea rhododendroides
X Sakersia laurentii
Syzygium guineense ..
X X Tetrorchidium didymostemon
9 10 13a 14
16 18 21a
X X X X X
FOREST COMPOSITION NEAR RUHIZA, IMPENETRABLE FOREST
Drypetes aff. gerrardii
qf. Maytenus undatus
XXX = abundant;
Key to symbols
XX = common;
X = present
VEGETATION IN KIGEZI 183
h ( 2 0
nh s SCALE
no M) 10
Fig. 2. Sketches of forest profiles in the Impenetrable Forest. (Captions page 184).
Captions to Figure 2
Fig. 2. Sketches of forest profiles in the Impenetrable Forest. 1-hilltop
forest near Ruhiza, 8,100 ft. (2,470 m.); 2-ridgetop forest on slightly sloping
ground, near Ruhiza, 7,600 ft. (2,320 m); 3-hillslope forest near Mubwende,
7,500 ft. (2,290 m.); 4-valley and lower hillslope forests near Itama River,
5,700 ft. (1,740 m.).
Key to symbols: a-Allophylus macrobotrys; b-Neoboutonia macrocalyx; c--
Croton macrostachys; d-Drypetes aff. gerrardii; e-Ekebergia rueppeliana; f-
Chrysophyllum fulvum; g-Galiniera coffeoides; h-Conopharyngia holstii; j-Psy-
chotria megistosticta; 1-Alangium chinense; m-Macaranga kilimandscharica;
n-Newtonia buchananii; o-Olea hochstetteri; p-Podocarpus milangianus; r--
Rapanea rhododendroides; s-Strombosia scheffieri; t-Maesa lanceolata; u-Olinia
usambarensis ;v-Polyscias fulva; w-Cassipourea ruwensorensis; x-Rytigynia sp.
Captions to Photographs
1. Ridge forest near Ruhiza 8,000 ft. (2450 m.). The trees with the thin
straight trunks in the foreground are mainly Rapanea rhododendroides.
Faurea saligna is the dark coloured tree in the centre background, and on
the extreme right is Maesa lanceolata.
The shrubs are mainly Cyathula cf uncinulata and Mimulopsis solmsii.
2. Hillslope forest near Mubwende, 7,500 ft. (2290 m.).
The straight white trunks in the foreground belong to Cassipourea ruwen-
sorensis. Behind are some large trees of Chrysophllum fidvum.
3. Gulley forest near Ruhiza, 7,500 ft. (2290 m.). A large spreading tree of
Polysciasfulva can be seen in the centre of the picture.
4. A clearing in the gulley forest caused by elephants digging for roots of
Note the climber tangles and the invasion of the site by Pteridum aquilinum.
VEGETATION IN KIGEZI 185
tree trunks and scrambling up to a height of 6 m. Ferns are abundant on the
forest floor beneath the Cyathula.
Gully forests vary in species-composition and structure from their upper
to lower parts. They are usually very open, this being partly attributable to
the instability of the soils and partly to elephants which can in places cause
great damage by trampling the ground and uprooting trees and digging for
their roots (Photo 4). The main species appear to be Macaranga kilimands-
charica and Polysciasfulva. The usual appearance of the upper part of these
gullies is of a dense tangle of shrubs and climbers, standing about 2 m tall,
the main species being Cyathula cf. uncinulata, with associated Rubus sp., Impa-
tiens sp., Urera hypselodendron and Pteridium aquilinum. From this tangle scattered
trees, mainly Neoboutonia macrocalyx, Dombeya goetzenii and Polysciasfulva, emerge,
often bearing a heavy load of climbers, amongst which Urera and species of
Clematis are common. The vegetation is similar in the lower parts of the valleys,
but here tall trees of Pygeum africanum, Croton macrostachys and Strombosia sche-
fleri are also found.
In the valleys the canopy is taller, reaching 30 m, with well developed
understorey and shrub layers. Amongst the common trees Symphonia globulifera
and Newtonia buchananii are characteristic. Croton megalocarpus is said to form
almost pure stands in some parts (J. Ball-personal communication). Parinari
excelsa occurs but is probably not as common as is sometimes supposed, being
mistaken for other species.
Further north, in the valley of the Itama river, where the land lies at
about 5,200 to 5,800 ft. (1,580-1,770 m), the vegetation has essentially the
same structure as at Ruhiza. The topography is less rugged and the vegetation
considerably more luxuriant, with a greater number of species, taller trees
and denser undergrowth. Although situated in Types 2a and 4c of Cahusac's
map, the species composition is more or less similar to that given for Type 4 a
(see Table 1).
Ridgetop forests contain numerous species, including Olea hochstetteri,
Harungana madagascariensis, Polyscias fulva and Strombosia scheffleri. The canopy is
about 20 m tall. There is little understorey or undergrowth vegetation. On
hillslopes, the most common of the taller trees is Strombosia scheffleri. Newtonia
buchananii and Ekebergia rueppeliana are also abundant in the upperstorey, with
the attractive Meliaceous trees Carapa grandiflora and Guarea mayombensis com-
mon in the second tree stratum. On lower hillslopes, as well as in gullies and
valleys, the tree fern, Cyathea manniana, grows luxuriantly. It constitutes one of
the main items of food in the diet of the mountain gorilla, which is quite
common in this part of the forest. The undergrowth on hillslopes consists
largely of ferns and Acanthaceae, with Brachystephanus africanus amongst the
latter. In the lower parts of gullies and in valleys the most common tree species
are Croton macrostachys, Myrianthus holstii, Neoboutonia macrocalyx and Anthocleista
zambesiaca, but in many places are widely spaced, with large patches of shrubs,
climber tangles and tree ferns in between.
Two of the commonest climbers at Itama are Sericostachys tomentosa and
Urera hypselodendron. Although the former species can usually be found with a
few flowering shoots, it also exhibits the phenomenon of periodic mass-flowering.
Such is the abundance of Sericostachys in the canopy that, when this happens,
the forest appears from above like a billowy white sea (T. Spiropoulos-
personal communication). As far as I know, the periodicity is not definitely
known. The mountain bamboo (but not in Kigezi) and several species of
Acanthaceae on Elgon (Tweedie 1965) are also known to be mass-flowerers.
186 UGANDA JOURNAL
The differences in floristic composition between Ruhiza and Itama are
presumably mainly due to changing environmental conditions associated
with altitude. Clair Thompson (in Leggat and Beaton 1961) considers that the
forest can be roughly divided into two altitudinal zones, above and below
about 6,800 ft (2,070 m), but this classification is too simple, since, amongst
other things, it overlooks the topographical pattern superimposed on the
altitudinal. In the present state of knowledge it is sufficient to say that the
forest becomes less tall and poorer in species at higher altitudes. Amongst
others, the following taxa become rarer or absent in the higher parts of the
forest:- Parinari excelsa, Guarea mayombensis, Carapa grandflora, Entandrophragma
excelsum, Aningeria adolfi-friederici, and the following taxa become more common-
Podocarpus milanjianus, Rapanea rhododendroides, Macaranga kilimandscharica.
The overall frequency of all trees in the Impenetrable Forest has not yet
been established, but an enumeration of timber trees in the eastern part
(in Leggat and Osmaston 1961), showed that the most abundant timber
trees were species of Chrysophyllum; followed by Olea welwitschii (= Olea hochs-
tetteri), Podocarpus spp., Strombosia scheffleri and Ficalhoa laurifolia, and by less
common Entandrophragma excelsum, Fagara spp., Pygeum africanum, JVewtonia
buchananii, Symphonia globulifera and Ocotea usambarensis.
(b) Echuya Forest
Mountain bamboo, Arundinaria alpina, is rare in the Impenetrable Forest,
being found predominantly in a few sites over 8,000 ft (2,450 m) in the south-
east. This contrasts strongly with the second large area of natural vege-
tation on Precambrian rocks, Echuya Forest. In this forest, which is situated
on the Echuya ridge between the Bufumbira bay and the Bunyonyi valley,
bamboo is dominant over the greater part of the reserve, from below 7,400 ft
(2,260 m) to 8,000 ft (2,450 m). This altitudinal range is unusual for Uganda,
where bamboo normally only grows gregariously above 8,000 ft (2,450 m), as,
for instance, on Elgon, the Bufumbira volcanoes and Ruwenzori. Another
unusual feature of bamboo in Kigezi is the rarity of flowering shoots. Mass-
flowering has never been reported.
The bamboo at Echuya grows on ridges, hillslopes and in gullies. It
varies in height from 10-15 m. The shoots are spaced well apart and not
clumped. The type and abundance of associated species varies with the amount
of cover. Where the canopy is lighter, a fair number of woody and herbaceous
plants are found beneath the bamboo. Bersama abyssinica is common, reaching
a height of about 5 m. Small shrubs, such as Rubus sp., Piper capense and Rhamnus
prinoides, are scattered over a dense cover of Panicum sp., which is associated
with Parochetis communis, several species of Impatiens, including I. eminii, Dro-
guetia iners, and the ferns, Asplenium friesiorum and Dryopteris kilmensis. Where
the bamboo is taller and the shade denser, the undergrowth is much poorer.
The most common shrubby species is Mimulpsis solmsii, forming bushes about
2 m tall often over an incomplete layer of Panicum.
In many gullies at Echuya, patches of broad-leaved forest occur. The
most common species are Macaranga kilimandscharica and .Neoboutonia macrocalyx,
standing about 15 m tall. Other trees present are Rytigynia sp., Bersama abyssinica,
Maesa lanceolata, Erythrococca sp., Myrica salicifolia, Polysciasfulva and Nuxia sp.
Widely spaced shoots of bamboo occur with the trees. There is a shrubby layer
of Mimulopsis solmsii, about 1 m tall, beneath which are found several species
of Urticaceae, including Urtica massaica, and ferns.
VEGETATION IN KIGEZI 187
(c) Mafuga Forest
Mafuga, on the eastern side of the Ishasha valley is not a climax forest,
being thought to be only about sixty years old. However, it is of particular
interest because of the light it throws on forest regeneration and succession.
Stretching from the Ishasha valley to Mafuga ridge, from 6 to 8,000 ft (1,830
-2,450 m), the vegetation of Mafuga is variable and patchy with areas of
grassland and woodland contained within the reserve. As in the Impenetrable
Forest, the most convenient way to classify the vegetation is on a topographical
In forests on the ridge at 8,000 ft (2,450 m) the commonest trees are Maca-
ranga kilimandscharica, Hagenia abyssinica, Rapanea rhododendroides, Agauria sali-
cifolia, Polyscias fulva, Maesa lanceolata, Psychotria megistosticta, Pittosporum spathi-
calyx and Nuxia floribunda. Forest dominated by Agauria are rather open, but
otherwise they are closed and about 15 m tall. There is usually a very open
shrubby layer containing Xymalos monospora and Psychotria sp., with numerous
herbaceous species present on the forest floor, including Asplenium friesiorum.
A. pseudoserra, Thalictrum rhynchocarpum, Droguetia iners and Peperomia sp., but
under Agauria the undergrowth is taller and consists of a dense tangle of Rubus
sp., Urtica massaica, Pteridium aquilinum and grasses. Neither Agauria nor Hagenia
regenerate beneath closed forests at Mafuga and an earlier stage of the succession
on ridges is provided by patches of Myrica salicifolia and Nuxia congesta, wood-
The hillslope vegetation is a complex mosaic of forests, woodlands and
grass-covered areas. The commonest trees in forests on upper hillslopes are
Macaranga kilimandscharica, Polyscias fulva and smaller Xymalos monospora, with
less frequent Maesa, Ficalhoa and Psychotria, and, on lower slopes, Albizia
gummifera, Polysciasfulva and Pygeum africanum, with smaller Maesa and Allophylus
macrobotrys. The undergrowth is denser on the lower slopes, where it consists
of a thick herbaceous tangle containing much Phytolacca dodecandra. In the
lower slope forests, climbers such as Acacia sp., Gouania longispicata and Clematis
simensis are common. The patches of woodland and bracken-grassland contain
abundant Myrica salicifolia as well as the trees Hagenia abyssinica, Albizia gum-
mifera, Agauria salicifolia, Erythrina abyssinica, Dombeya mukole and Croton oxypetalus.
In the upper parts of gullies Dombeya goetzenii and Cyathea manniana grow
over a dense shrub layer of Acanthaceae. In the lower parts Polyscias, Albizia
and Pygeum are the most common trees, growing with Alangium chinense, Pittos-
porum spathicalyx and Dombeya goetzenii and there is a thick herbaceous tangle
0.5-1.5 m tall, with abundant Urtica massazca. In the more open gullies, the
wild banana, Ensete sp., is abundant.
An enumeration carried out by the Forest Department (unpublished,
1947) showed that numerous seedlings and small trees of species now charac-
teristic of the more mature forests of the Impenetrable are present under the
tree canopy. The species recorded include Podocarpus milanjianus, Entandrophragma
excelsum, Chrysophyllum albidum, C. fulvum, Fagara macrophylla, Aningeria sp., Faga-
ropsis angolensis, Ekebergia rueppeliana, Symphonia globulifera, Cassipourea ruwenzo-
rensis and Strombosia scheffleri. It is apparent that the present vegetation consists
of several early stages in the regeneration of forest, ranging from Pteridium-
grassland through woodland to secondary forest.
One of the commonest trees at Mafuga at the present time is Albizia
gummifera. This species has an interesting distribution in the highland parts of
western Uganda. It is common on the Ruwenzori at c. 7-8,000 ft (2,150-
2,450 m) on hillslopes, but does not occur, at least commonly, in the Impene-
188 UGANDA JOURNAL
trable Forest. It is, however, one of the most abundant trees in patches of
secondary forest just outside that reserve near Rushasha, where it is associated
with Polysciasfulva, Carapa grandiflora, Maesa lanceolata, Croton macrostachys, Maca-
ranga kilimanscharica, Schefflera sp., Alangium chinense and Neoboutonia macrocalyx,
all species found in the forest and Tremaguineensis, Nuxia sp. Hagenia abyssinica and
Erythrina abyssinica, species of woodland and early secondary forest. Its abundance
on the Ruwenzori is probably due to the greater instability of soils of the very
(d) Secondary Forest Patches
Patches of secondary forest increase in density towards the Impenetrable
Forest, being entirely absent from the Bufumbira bay. These forest patches
are found mainly in gullies, but, in less intensely cultivated areas, also on
hillslopes and in valleys. North from lake Mutanda towards the edge of the
Impenetrable Forest at Rushasha, patches of open forest increase in number,
size and density. Albizia gummifera and Dombeya goetzenii are two of the commonest
species on hillslopes and amongst the first to appear, with Rumex usambarensis
and Acanthus arboreus in the shrub layer. Northwards, Erythrina abyssinica and
Markhamia platycalyx, which are found frequently in the Bufumbira embay-
ment decrease in frequency, whilst Ensete sp. and Lobelia giberroa increase, and
many forest trees appear. These include Myrianthus holstii, Trema guineensis,
Maesa lanceolata and Alangium chinense. Myrica salicifolia, Nuxia floribunda, Dodonaea
viscosa and Vernonia sp. are also found, but these have greater affinities to wood-
land, which will be described later. On the hillslopes at 6,000 ft (1,830 m),
the first Hagenia appears, together with the decorative Sakersia laurentii, and
these are associated with Nuxia sp., Maesa lanceolata, Polyscias fulva, Albizia
gummifera and rarer Anthocleista zambesiaca, with Lobelia giberroa and Pteridium
beneath. In the valleys a similar increase in number of forest tree species
occurs towards the Impenetrable Forest. Croton macrostachys, Anthocleista zambe-
siaca, Mitragyna rubrostipulata and Pygeum africanum become common. Pygeum
is incidentally, one of the commonest tall indigenous trees found in valleys
amongst densely cultivated land, where it is protected for its valuable timber.
Just outside the Impenetrable Forest boundary, the number of forest tree
species rises sharply and includes typical closed forest species such as Carapa
grandiflora, Macaranga kilimandscharica and Schefflera sp.
There are two large softwood plantations in southwest Kigezi, at Mafuga
and Muko. The chief species are Cupressus lusitanica and species of Pinus, P. patula
and P. radiata, with fire-breaks of Eucalyptus. A great deal of research on soft-
woods has been carried out at Mafuga. In addition to these large and extre-
mely important schemes, there are numerous smaller patches of Black Wattle
and Eucalyptus, which have been planted for local fuel and pole production.
Non-Forested Regions on Precambrian Rocks
Kigezi is today one of the most densely populated parts of eastern Africa.
The pattern of settlement has varied somewhat during the past fifty years,
at least around Kabale, which is said to have been only thinly populated in
1910 (Purseglove 1950). There is no doubt that, at one time, either all or
nearly all of the Precambrian area was covered with forest. There is evidence
from pollen-analysis that forest clearance started very early in certain
parts of Kigezi. Continued cutting and burning has steadily encroached on
VEGETATION IN KIGEZI 189
the forests, which have been replaced by new plant communities associated
with agricultural and pastoral practices. The vegetation of these disturbed
areas has been described by Burtt (1934) and Snowden (1933, 1953). In the
present description it is arbitrarily classified into three categories, based on
species-composition, structure and land use.
(a) Woodland and Dry Pasture
On ridges and upper hillslopes, in places not used for agriculture, a
common type of vegetation is an open woodland, which, except where very
dense, is subjected to occasional fires. This woodland is called "Subtropical
bush" by Snowden (1933) and "Montane Thicket" by Langdale-Brown (1960).
According to Langdale-Brown, it is one of the most widespread types of vege-
tation in the study area, being particularly common from the Echuya ridge to
the Ishasha valley and northwards as far as Mafuga. Both authors consider it
to be a climax vegetation type, but this seems unlikely since at Mafuga and
elsewhere, where left comparatively undisturbed, it can be seen to be changing
into secondary forest.
The trees are usually widely spaced and from 4-12 m tall. The most
abundant is probably Nuxia floribunda, accompanied by Myrica salicifolia,
Erythrina abyssinica (mainly below 7,000 ft (2,150 m), Hagenia abyssinica (mainly
above 7,000 ft (2,150 m), Dodonaea viscosa, Agauria salicifolia and Philippia
benguelensis, all of which are fire-resistant. There is an understorey layer of
Pteridium, in which Rumex usambarensis, Polygala ruwenzoriensis and Pycnostachys
erici-rosenii are abundant. Numerous other shrubs and herbs, including Acanthus
arboreus, Eriosema montanum, Rumex abyssinicus, Triumfetta sp., Vernonia sp. and
Crotolaria sp., are also present. Woodland in Echuya forest differs from those
elsewhere in containing abundant Hypericum lanceolatum, a tree common in other
parts of Kigezi only above 9,000 ft (2,750 m) on the Bufumbira volcanoes.
The following grasses have been recorded from this type of vegetation by
Snowden (1933): Hyparrhenia cymbaria, H. diplandra, (both common), Cymbopo.gon
afronardus, Digitaria uniglumis, Melinis minutiflora and Eragrostis olivacea. According
to Snowden, with the further destruction of this vegetation by burning and
clearing, it gives way to a "Subtropical Drier Pastureland", in which the dominant
grasses are Hyparrhenia filipendula, H. cymbaria, Exotheca abyssinica and Setaria
sphacelata. Langdale-Brown (1960) gives a similar list, but also records two
other species of Hyparrhenia as abundant, H. lecomtei and H. pilgeriana.
(b) Thicket and Moist Pasture
A widespread vegetation type is pastureland of a low sward of Pennisetum
clandestinum. This is found chiefly in valleys and lower hillslopes where there is
grazing but no burning, In the Bufumbira bay, it is common on volcanic
ashes and on those areas which have been cleared of lava boulders to make
pastures. Amongst the many associated plants are the grasses Andropogon
abyssinicus, Cynodon plectostachys, Digitaria scalarum, Eragrostis tenuifolia, and Sporo-
bolos sp. (Snowden 1933), and Pycnostachys erici-rosenii, Pentas zanzibarica and
species of Hibiscus, Vernonia, Commelina, Gnaphalium, Dissotis and Cerastium.
Isolated bushes and thickets are very common in these meadows. On the
east of the swamp behind Muko, for instance, there is an extensive patch of
Acanthus arboreus, about 3 m tall. Maesa lanceolata is abundant and the rich
ground flora contains Dipsacus pinnatifidus, and species of Geranium, Galium,
Impatiens and Hibiscus. Elsewhere in pastureland, small thickets or isolated
trees of Dodonaea, Clerodendrum johnstonii, Erythrina abyssinica, Lantana sp., Rumex
190 UGANDA JOURNAL
usambarensis, Dracaena afromontana (in gullies), Vernonia sp. and Solanum aculeas-
trum are common.
Cultivation in the past was mainly restricted to the valleys and lower
hillslopes, but, with the increasing population, this is now extending to upper
hillslopes and ridges (Purseglove 1950). In addition, swamps are being drained
to supply extra land, particularly for growing vegetables. On the steeper
slopes terrace cultivation is practised to prevent soil erosion. The main crops
are sorghum, finger millet, sweet potatoes, Irish potatoes, peas, beans, tobacco
and, below 6,400 (1,950 m), bananas. Weeds are common in the fields. Members
of the Compositae, for instance Conyza floribunda, Crassocephalum montuosum,
Bidenspilosa and B. grantii, are particularly abundant, together with Achyranthes
aspera, Rumex bequaertii, Galium sp., Commelina benghalensis and various grasses.
The Vegetation of the Volcanic Areas
In the Bufumbira bay there is little woody vegetation apart from small
plantations of exotic trees. Much of the ash-covered area is either Pennisetum
clandestinum meadow or is under cultivation. The woody plants are found on
the lava flows, but these do not reach a great size, nor do they form extensive
woodlands or thickets. The general aspect of the vegetation is rather "xero-
The woody plants that occur on lava flows include Ficus sp., Hagenia
abyssinica (rare), Cussonia holstii, Erythrina abyssinica and Solanum aculeastrum, the
last three also being used for live hedges.
One of the commonest shrubs is Rumex usambarensis. Other shrubs and
herbs present include the grasses Exotheca abyssinica and Hyparrhenia filipendula
(Langdale-Brown 1960) and Acanthus arboreus, Crassocephalum sp., Galium sp.,
Justicia unyorensis, Kalanchoe sp., Pentas lanceolata, Stachys aculeolata, and Vernonia
sp. On lava flows in the Muko volcanic outlier at the north end of lake Bunyonyi,
a fairly well developed thicket is sometimes present. Dodonaea and Maytenus
senegalensis are abundant, and the other species present include Bridelia sp.,
Erythina abyssinica, Lantana camera, Nuxia sp., Pteridium aquilinum, Pycnostachys sp.,
Rumex sp., Syzygium cordatum (an unusual habitat for this plant) and Triumfetta
Robyns (1948) gives an account of the stages of colonization of lava
flows in the Congo near Kigezi. After the first colonization by lichens and
mosses, loose earth accumulates in the crevices where many ferns, Rumex
usambarensis, Kalanchoe crenata, Celsia brevipedicellata and other herbs start to
grow. Shrubs then become established and this is succeeded by a sclerophyllous
woodland about 10-15 m high with a discontinuous tree stratum of Myrica
salicifolia, Erythina abyssinica, Cussonia holstii, Pittosporum spathicalyx, Olea africana
and Maesa lanceolota. Presumably lava flows in Kigezi also once bore such a
woodland before clearance.
The swamps found in many valleys in Kigezi contain several vegetation
types and a great diversity of species. There is some degree of change with
altitude, seen, for instance, in changes in the dominant species and in the res-
triction of certain taxa to certain altitudinal ranges. Examples of the latter case
VEGETATION IN KIGEZI 191
are Eriocaulon schimperi, Erica sp., and most species of Helichrysum; all found only
above 6,500 ft (1,980 m). The largest break occurs between Muchoya swamp
(7,400 ft (2,260 m) ) in Echuya Forest and swamps at lower altitudes. This
division was also recognized by Snowden (1963), who classified this swamp into
his "Temperate Zone", the others into his "Subtropical".
The vegetation of Muchoya swamp has been described by Morrison
(1968). The dominant over most of its area is Pycreus cf. nigricans, a tussock-
forming sedge, not recorded from other Kigezi swamps. Two other tussock-
forming species occur in restricted areas, Cyperus platycaulis var. lucentinigricus
and Pennisetum mildbraedii, the latter usually rooted on the suppressed tussocks
of Pycreus. The Pycreus tussocks are about 30 cm tall and 75 cm across, and their
structure dominates the distribution of other species on the swamp surface.
In the gaps between the tussocks, the ground is fairly flat and standing water
sometimes occurs. The other species on the swamp fall into two groups, those
that are directly influenced by the edaphic conditions of the swamp surface
and those that are largely independent by being rooted on the tussocks (Fig. 3).
The first group contains Hypericum peplidifolium, Carex spp., Sphagnum, Hydrocotyle
sibthorpiodes, Panicum eikii and Alchemilla ellenbeckii, some of which are also some-
times found rooted on the tussocks, and the second Erica sp., Myrica kandtiana,
Helichrysum spp., Lobelia cf. giberroa, Kniphofia thomsonii and others. Many of the
species of this second group also occur in nearby grassland. The swamp vege-
tation is occasionally burnt and it is probable that if this were not so, Erica
and Myrica would grow larger and cover more extensive areas of the swamp.
Seven of the most important herbaceous species dominant on swamps at
lower altitudes in southwest Kigezi are Cyperus latifolius, C. cf. papyrus, C. platycaulis
var. lucentinigricus, C. cf. denudatus, Miscanthidium voilaceum, Sphagnum sp., and
Typha australis. Cyperus papryus is usually found growing towards the margins of
swamps. Papyrus is a large sedge, growing to over 3 m and most abundant
below 6,500 ft (1,980 m). It has few associated species. In the Ishasha valley,
for instance, where Papyrus covers large areas, the only common associated
species are Cyperus platycaulis var. lucentinigricus and Thelypteris squamigera. Mis-
canthidium voilaceum is a common swamp grass in Uganda, growing to a height
of 2.5 m or so and generally found inside a Papyrus fringe. Unlike Papyrus,
it has numerous associated species, especially Thelypteris, Cyperus platycaulis var.
lucentinigricus, Pycreus mundtii, Dissotis canescens and Impatiens sp., Miscanthidium
often grows with Sphagnum, which may suppress the Miscanthidium, itself giving
way at a later date to swamp forest (Lind 1956). Sphagnum carpets can be seen
at Butongo and Muko. Apart from poorly grown Miscanthidium shoots, its
associates are Drosera madagascariensis, Pycreus mundtii, Cyperus nudicaulis, Hyperi-
cum scioanum and Lobelia cf. kiwuensis. Typha australis, covers comparatively small
areas towards the margins of swamps. Its distribution in Kashambya valley
suggests that its growth is favoured by flooding and heavy silting (Lind 1956).
At Muko it has few associated species, the only common one being Scirpus
inclinatus. C. cf. platycaulis var. lucentinigricus (already noted as an associate of
Miscanthidium and Papyrus) is dominant over the largest part of Muko swamp,
where it is found with Hypericum sp., Alchemilla ellenbeckii and Thelypteris squa-
migera. C. cf. denudatus is dominant over much of Mubwende swamp in the
Impenetrable Forest, where it occurs with Alchemilla, Thelypteris, Helichrysum
spp., Lobelia cf. gibberoa, Epilobium, sp., Crasocephalum sp., Gnaphalium luteo-album
and Eriocaulon schimperi.
Whilst these herbaceous communities cover the greater part of Kigezi
swamps, the climax vegetation is probably often swamp forest dominated by
b a cad a
0 50 cm
Fig 3. Line transect in Muchoya Swamp showing tussocks of Pycreus ofnigricans (a), on which
are growing Kniphofia thomsonii (b), Pennisetum mildbraedii (c) and Helichrysumformosissimum (d),
and intervening hollows with Panicum eikii (e) and Sphagnum sp. (f(.
VEGETATION IN KIGEZI 193
Syzygium cordatum. This is a vegetation type at present rapidly diminishing
in extent due to cutting and draining of the swamps for agriculture. The
species itself is recorded from many swamps, but only forms extensive patches
of closed forests in the Ishasha valley and in Rubanda swamp. In the Ishasha
valley, the canopy is about 6-9 m tall and consists almost entirely of Syzygium.
The trees are gnarled and covered with epiphytic bryophytes and lichens.
Climbers, particularly Begonia cf. meyeri-johannis and Culcasia scandens, reach
up into the crowns. A lower tree layer, about 3-6 m tall, is present beneath
the Syzygium, in which the most common trees are Rapanea pulchra, Myrica
kandtiana and Grumilea sp., with rarer Ilex mitis. The forest floor is uneven,
consisting of areas that are slightly raised and areas slightly sunken. On the
raised parts, there is a solid mat of Selaginella sp., growing with Galium sp.,
Impatiens eminii, and Asplenium pseudoserra. In the hollows are Hydrocotyler of
verticillata, Zehneria minutiflora, Polygonum sp., Crassoscephalum montuosum, Cypeus,
cf. platycaulis and several ferns, including Osmunda regalis.
Apart from those woody species already mentioned as found on swamps,
species of Allophylus, Erica and Rhus are found at Muko swamp. Except Erica,
these are not normally considered to be swamp genera, and they may be using
tussocks to free themselves from direct contact with the swamp surface. The
only other Kigezi swamp from which the genus Erica has been recorded is
Muchuya, but it is interesting to note that on the Ruwenzori, Erica bequaertii
forms tall swamp forests at about 9,000 ft (2,750 m).
Around the margins of swamps in Kigezi there is a change in the vegeta-
tion consequent to the wetter and richer soil. Cyperus latifolius is very common,
being found with many other plants, such as Polygonum spp., Ranunculus multi-
fidus, Crassocephalum picridifolium, Lythrum rotundifolium and Alchemilla ellenbeckii.
The marginal vegetation at Muchoya is different from elsewhere and here,
where bamboo or Hypericum thicket directly borders the swamp, Alchemilla
ellenbeckii forms tangles up to 1 m tall, with associated Stachys aculeolata, Mariscus
tomaiophyllus and Lobelia cf. gibberoa; and, where grassland borders the swamp,
there are large tussocks ofPennisetum mildbraedii. In some bays, Gunnera perspensa,
a plant much relished by elephants, is abundant.
Factors Controlling the Distribution of the Vegetation
With increasing altitude, the climate becomes colder, the in-coming and
out-going radiation during clear periods higher, and the annual rainfall
greater. The net result of all the environmental changes associated with altitude,
at least as regards forest trees, is to lower the number of species, to increase
uniformity in structure and physiognomy and to decrease the average leaf
size (unpublished data). Apart from Clair Thompson (in Leggat and Osmaston
1961), already mentioned in connection with the Impenetrable Forest, several
other authors have attempted to divide the vegetation of Kigezi into altitu-
dinal zones (Hedberg 1951; Snowden 1933, 1953). The complexity of zonal
distribution is best seen on the Bufumbira volcanoes (Kingston 1967), but is
also apparent at Echuya forest, many of whose species, such as Arundinaria
alpina and Hypericum lanceolatum, are more usually found at higher altitudes.
The vegetation patterns associated with topography described for the
Impenetrable Forest are almost certainly related to variations in soil-moisture
availability. The distribution of many of the trees elsewhere in Uganda helps
to confirm this hypothesis. For instance, Faurea saligna is generally found on
drier sites, particularly ridges, and is also very common on the highly porous
soils of the lower slopes of the Bufumbira volcanoes.
194 UGANDA JOURNAL
Similar patterns are apparent not only in the other forests but also in the
disturbed areas. This can be seen, for instance, in the distribution of agricul-
ture, which tends to be concentrated on lower slope and valley soils, where the
soil-moisture regime is more stable. A suggested scheme showing the inter-
relationships of the various vegetation types found in the Precambrian area
is shown in Table 3.
Water-stress, particularly during the June-July dry season, must also
be responsible for the xeromorphic appearance of the vegetation on the highly
porous volcanic soils. Although outside the area, this is most obvious when the
bamboo zones of the Bufumbira volcanoes and Echuya are compared. On
the volcanoes the shoots are thin and wiry and only reach a height of about
3 m, whilst at Echuya they are thick and up to 15 m tall.
The openness of gully vegetation in the Impenetrable Forest seems to be
due to soil instability. Many of the trees found, such as Polyscias fulva and
Pygeum africanum, are known to be secondary forest components at lower alti-
tudes, and it is likely that their presence in gullys in the Impenetrable is a
reflection of similar ecology. However some species are found to occupy different
niches at different altitudes. Pygeum is, for instance, a member of the climax
vegetation at 10,500 ft (3,200 m) on Elgon, and the same is true of Hagenia,
which in the Rukiga highlands is almost invariably a member of secondary
It is obvious from this account of the vegetation of southwest Kigezi
that much more research into the ecology of this and other highland regions
of Uganda needs to be undertaken. Such research would undoubtedly greatly
modify the tentative conclusions reached here and would be of great value both
for increasing knowledge of tropical ecosystems and for providing a basis for
the best utilization of the potential of the land. The Impenetrable Forest
provides an irreplaceable open-air laboratory for future studies.
Numerous individuals have assisted in this work. I would particularly
like to record the help given by Dr. G. H. Lucas and Dr. B. Verdcourt at
Kew, the staff of the Botany Department at Makerere, the officers of the
Uganda Forestry Department and Mr. E. Balyetagara and Mr. J. Tumwe-
sigye in Kigezi. Dr. M. E. S. Morrison and Miss N. Masembe kindly read
the manuscript and supplied useful criticism.
SUGGESTED INTERRELATIONSHIPS OF VEGETATION TYPES ON SOILS DERIVED FROM PRECAMBRIAN ROCKS
CLIMAX RIDGE FOREST WOODLAND
Podocarpw milanjianus Nuxia congesta
Macaranga kilimandscharita \ Hagenia abyssinica
Olea hochstetteri Aauria salicifolia --
Maesa lanceolata- Myria salicifolia
Rapanea rhododendroids AB Dodonaea viscosa AB
CLIMAX SLOPE rORESTY SECONDARY FOREST ABY
Chrysophyllumfulvtu Albizia gummifera
Olia hocksetteri Pelyscas fulva
Dryptes aff.' rrardii A
Allophylus macrobotrys X
CLIMAX VALLEY FOREST
Symphonia globulifera A
Pygeum africanum A
Woodland trees and
Rumex spp. ---
Hyparrhenia spp. ABC
Dodonwea viscosa AC
A, Cutting and clearing; B, Burning; C, Grazing; X, Better sites; Y, Poorer sites.
BEADLE, L. C. and LIND, E. M. (1960). Research on the swamps of Uganda. Uganda J. a4,
Burrr, B. D. (1934). A botanical reconnaissance in the Virunga Volcanoes of Kigezi, Ruanda,
Kivu. (Kew Bulletin, pp. 145-165).
CAHUSAC, A. B. (1958). Forest type map of the Impenetrable C.F.R. Forestry Dept., Entebbe.
CHENERY, E. M. (1951). Unpublished report, mentioned in Harrop (1960).
COMBE, A. D. and SIMMONS, W. C. (1933). The volcanic area of Bufumbira. Geological Survey
of Uganda, Memoir No. 3. Entebbe, Government Printer.
East African Meteorological Department (1967). Climatological statistics for East Africa and
Seychelles, Part 2. East African Meteorological Dept., Nairobi.
EGGELING, W. J. and DALE, I. R. (1951). The indigenous trees of ihe Uganda Protectorate. Glasgow,
HARROP, J. F. (1960). The soils of the Western Province of Uganda. Agricultural Dept., Uganda.
HEDBER G, O (1951). Vegetation belts of the East African mountains. (Svensk. Botanisk. Tidskrif.
45, pp. 140-202).
JOHANNESEN, I., KROSSHAUG, L. and KINGSTON, B. (1967). Soil type classification in Rwoho
Forestry Dept., Entebbe.
KINGSTON, B. (1967). Working plan for Mgahinga C.F.R. Forestry Dept., Entebbe.
LANGDALE-BROWN, I (1960). The vegetation of the Western Province of Uganda. Agricultural Dept.,
LEGGAT, G. J. and OSMASTON, H. A. (1961). Working plan for Impenetrable C.F.R. Forestry
LIND, E. M. (1956). Studies in Uganda swamps. (UgandaJ. 2o, pp. 166-176).
LIND, E. M. and TALLANTIRE, A. C. (1962). Some common flowering plants of Uganda. London.
Oxford University Press.
MORRISON, M. E. S. (1968) Vegetation and climate in the uplands of south-western Uganda
during the later Pleistocene period. 1. Muchoya Swamp, Kigezi District. (Journal ofEcology,
56, pp. 363-384).
PURSEGLOVE, J. W. (1950), Kigezi resettlement. (Uganda J. 14, pp. 139-152).
ROBYNs, W. (1948) Flore des spermatophytes du Pare National Albert, Volume 1, Bruxelles.
SNOWDEN, J. D. (1933), A study in altitudinal zonation in South Kigezi and on Mounts Muha-
vura and Mgahinga, Uganda. (Journal of Ecology, ua, pp. 7-27).
SNOWDEN, J. D. (1953). The grass communities and mountain vegetation of Uganda. London, Crown
TWEEDIE, E. M. (1965). Periodic flowering of some Acanthaceae on Mt. Elgon. (Journal of the
East Africa Natura I History Society, 25, pp. 92-94).
U ganda. Atlas of Uganda (1962) Kampala, Lands and Survey Dept.
This list contains the scientific names of all species mentioned in the text. The most up-
to-date authorities have been consulted where available, but undoubtedly some of the names
given will by now have been superseded. The Rukiga names are taken partly from my own
notes, and partly from Eggeling and Dale (1951), Leggat and Osmaston (1961) and Lind and
Tallantire (1962). Miss F. Kabeere has corrected the spelling and substantially augmented
VEGETATION IN KIGEZI
Acanthus arboreus Forssk.
Achyranthes aspera Linn.
Agauria salicifolia (Comm. ex Lam.) Hook.f. ex Oliv.
Alangium chinense (Lour.) Rehder
Albizia gummifera (J. F. Gmel.) C.A.Sm.
Alchemilla ellenbeckii Engl.
Allophylus macrobotrys Gilg
Andropogon abyssinicus R. Br. ex Fresen.
Aningeria adolfi-friederici (Engl.) Robyns et Gilbert
Anthocleista zambesiaca Baker
Arundinaria alpina K. Schum.
Aspleniumfriesiorum C. Chr.
A. pseudoserra Domin.
Begonia meyeri--johannis Engl.
Beilschmeidia ugandensis Rendle
Bersama abyssinica Fresen.
Bidens grantii (Oliv.) Scherff.
B. pilosa L.
Brachystephanus africanus S. Moore.
Carapa grandiflora Sprague
Cassipourea ruwensorensis (Engl.) Alston
Celsia brevipedicellata Engl.
Chassalia subochreata (De Wild.) Robyns
Chrysophyllum albidum G. Don
C.fulvum S. Moore
Clematis simensis Fresen.
Commelina benghalensis L.
Conopharyngia holstii (Engl.) Stapf
Conyza floribunda H.B.K.
Crassocephalum montuosum (S. Moore) MAilne-Redhead
C. picridifolium (DC.) S. Moore
Croton macrostachys Hochst. ex Del.
C. megalocarpus Hutch.
C. oxypetalus Muell. Arg.
Culcasia scandens P. Beauv.
Cupressus lusitanica Mill,
Cussonia holstii Harms.
Cyathula uncinulata (Shrad.) Schinz
Cyathea manniana Hook.
Cymbopogon afronardus Stapf
Cynodon plectostachys (K. Schum.) Pilger
Cyperus denudatus L. f
C. latifolius Poir.
C. nudicaulis Poir.
C. papyrus L.
C. platycaulis Bak. var. lucentinigricus (K. Schum.) Kukenth.
Digitaria scalarum (Schweinf.) Chiov.
D. uniglumis (A.Rich.) Stapf
Dipsacu pinnatifidus A. Rich.
Dissotis canescens (E. Mey. ex Grah.) Hook.f.
Dombeya goetzenii K. Schum.
D. mukole Sprague
Dodonaea viscosa (L.) Jacq.
Dracaena afromontana Mildbr.
Droguetia iners (Forsk.) Schweinf.
Drosera madagascariensis DC.
Dryopteris kilmensis (Kuhn) O. Ktze.
Drypetes gerrardii Hutch.
Ekebergia rueppeliana (Fresen.) A. Rich.
Entandrophragma excelsum (Dawe et Sprague) Sprague
Eragrostis olivacea K. Schum.
E. tenuifolia (A.Rich.) Steud.
Erica bequaertii De Wild.
Eriocaulon schimperi Korn
Eriosema montanum Bak.f.
Erythrina abyssinica Lam.
Exotheca abyssinica (A. Rich.) Anderss.
Fagara macrophylla (Oliv.) Engl.
Fagaropsis angolensis (Engl.) Dale
Faurea saligna Harv.
Ficalhoa laurifolia Hiern
Galiniera cofeoides Del.
Ganaphalium luteo-album L.
Gouania longispicata Engl.
Guarea mayombensis Pellegrin
Gurmera perpensa L.
Hagenia abyssinica (Bruce) J. F. Gmel.
Harungana madagascariensis Lam.
Helichrysumformosissimum (Sch. Bip.) Sch. Bip.
Hydrocotyle sibthorpioides Hook. f
H. verticillata Thunb.
Hyparrhenia cymbaria (Linn.) Stapf
H. diplandra (Hack.) Stapf
H. filipendula (Hack.) Stapf
H. lecomtei (Franch.) Stapf
H. pilgeriana Hubbard
Hypericum lanceolatum Lam.
H. peplidifolium A. Rich.
H. scioanum Chiov.
Ilex mitis (L.) Radlk.
Impatiens eminii Warb.
Justicia unyorensis S. Moore
Kalanchoe crenata Haw.
Kniphofia thomsonii Baker
Lantana camera L.
Lobelia giberroa Hemsl.
L. kiwuensis Engl.
Lythrum rotundifolium Hochst.
Macaranga kilimandscharica Pax
Maesa lanceolata Forssk.
Mariscus tomaiophyllus C.B. Cl.
Markhamia platycalyx (Bak.) Sprague
Maytenus senegalensis (Lam.) Exell
M. undatus (Thunb.) Blakelock
Melinis minutiflora Beauv.
Mimulopsis solmsii Shweinf.
Miscanthidium voilaceum (K.(Schum.) Robyns
Mitragyna rubrostipulata (K. Schum.) Harv.
Musanga cecropoides R. Br.
Myrianthus hotstzi Engl.
Myrica kandtiana Engl.
M. salicifolia Hochst. ex A. Rich.
Neoboutonia macrocalyx Pax
Newtonia buchananii (Baker) Gilb. et Boui.
Nuxia congesta Fresen.
N. floribunda Benth.
Ocotea usambarensis Engl.
Olea africana Mill.
0. hochstetteri Baker
0. welwitschii (Knobl.) Gilg et Schellenb.
Olsinia usambarensis Gilg
Oamunda regalis L.
Panicum eickii Mez
Parinari excels Sabine
Prochetis communis Buch.-Ham. ex G. Don.
Pennisetum clandestinum Chiov.
P. mi!dbraedii Mez
Pentas lanceolata (Forsk.) K. Schum.
- Omushaga, Omuremankobe
- Hyparrhenia = Emburara.
- Echuvu, Mufe.
VEGETATION IN KIGEZI
P. zanzibarica (Klotzsch) Vatke
Philippia benguelensis (Engt.) Alm et Th. Fries Jr.
Phytolacca dodecandra L'Herit.
Pinus patula Schlecht. et Cham.
P. radiate D. Don
Piper capense L.f.
Pittosporum spathicalyx De Wild.
Podocarpus milanjianus Rendle
Polygala ruwenzoriensis Chodat
Polysciasfulva (Hiern) Harms
Psychotria megistosticta (S. Moore) Petit
Pteridium aquilinum (Linn.) Kuhn.
Pycnostachys erici-rosenii R. E. Fries
Pycreus mundtii Nees
P. nigricans (Steud.) C.B. Cl.
Pygeum africanum Hk. f
Ranunculus multifidus Forsk.
Rapanea pulchra Gilg et Shcellenb
R. rhododendroides (Gilg) Mez
Rhamnus prinoides L' Herit.
Ritchiea albersii Gilg
Rumex abyssinicus Jacq.
R. bequaertii De Wild.
R. usamba.ensis (Dammer) Dammer
Sakersia laurentii Cogn. ex De Wild et Dur.
Scirpus inclinatus (Del.) Aschers. et Schweinf. ex Boiss.
Sericostachys tomentosa Lopr.
Setaria sphacelata Stapfet Hubbard
Solanum aculeastrum Dunal
Stachvs aculeolata Hook.f
Strombosia scheffleri Engl.
Symphonia globuifera Linn.f.
Syzygium cordatum Hochst. ex Sond.
S. guineense (Wild.) D.C.
Tetrorchidium didymostemon Pax et K. Hoffm.
Thalictrum rhynchocarpum Dillon et A. Rich.
Thelypteris squamigera (Schlecht.) Ching.
Trema guineensis Ficalho
Typha australis Schumach.
Urera hypselodendron (Hochst.) Wedd.
Urtica massaica Mildbr.
Xymalos monospora (Harv.) Baill.
Zehneria minutiflora (Cogn.) C. Jefrey
NOTES ON CONTRIBUTORS
John Atanda is a lecturer in history of the University of Ibadan and has spent
a year on secondment to Makerere.
William Banage is the Professor of Zoology at Makerere and a past President
of the Uganda Society. The work for this article was done when he was in the
Faculty of Agriculture and was first presented to the East African Academy
Dr. R. L. Brahmachary works for the Institute of Statistics, of the Republic of
India in Calcutta. The work presented here arises from occasional vacations
spent in East Africa's game parks.
Alan Hamilton is a research student in the Department of Botany at Makerere
and has contributed to the Uganda Journal previously.
Merrick Posnansky is a regular contributor to the Uganda Journal and a past
President of the Uganda Society. At present he is the Professor of Archaeology
at the University of Ghana.
Regina Solzbacher has recently completed a Ph.D. thesis for the University of
Pennsylvania on the social structure of Kibuli. She spent two years doing
intensive research in this "village". The article published here arises from
a talk given to the Uganda Society.
Uganda Journal, 33, 2, (1969) pp. 201-208
TERRITORIAL BEHAVIOR AND POPULATION
IN THE GREY-BACKED FISCAL SHRIKE
By W. B. BANAGE
The Grey-backed Fiscal Shrike (Lanius excubitorius Pr6v. & Des Murs)
is a large and elegant shrike measuring about 10 inches long. Of several good
descriptions of it, that by Bannerman, quoted here, is perhaps the best. "The
upper parts are grey including the rump, underparts white; these areas being
divided by a broad black band continued from the back forehead through
the eye and on to the wings which are black with a broad white patch most
conspicuous in flight. The basal half of the tail is white, the terminal half
black." There is a heavy, black bill, typically hooked as in all Laniid shrikes,
and the legs are powerful and blackish-brown in colour (Bannerman, 1939).
The females have a light crimson patch on the flanks but this is not easy to
see in the field and therefore it is not possible to tell the sexes apart. The con-
trasting black, grey and white colours, and especially the white and black on
the wings and on the long, broad and somewhat graduated tail are fully
exploited in territorial display.
The Grey-backed Fiscal is found over a large area in East, Northeast
and West-Central Africa (Olivier, 1944, p. 247) being actually more wide-
spread than its coastal East African relative, the Long-tailed Fiscal (Lanius
cabanisi Hartert) which Williams (1963) describes in his popular field guide.
The typical habitat of the Grey-backed Fiscal is acacia or open bush vegeta-
tion with short grass in areas below 5,000 feet in elevation. The bulk of the
following observations were made in and around Kampala, but the birds
were also observed in the Queen Elizabeth and Murchison Falls National
Parks. All the populations belong to the typical race, L. e. excubitorius Pr6v. &
Des Murs (Mackworth-Pread & Grant, 1960). These observations depend
on field work, but could not be backed by colour ringing as facilities for this
were not available. Further investigations involving colour ringing and collect-
ing are to be carried out.
Voice and Territory Defence
Because of its size, its perching habits, mentioned later, and its noisy
behaviour, many naturalists have written on the Grey-backed Fiscal, but
sometimes in a dramatically exaggerated way. In almost all cases, however,
there is no mention of any territorial behaviour. Instead, the marked vocaliza-
tion so often noted has been regarded just as part of the birds' tendency to
form gregarious parties. Mackworth-Praed and Grant (1960, p. 591) say,
for instance, that "these birds are most gregarious, often dancing about all
over a tree in a party, like Babblers, or flying screaming from one tree to
another .... and occasionally assemble into large parties in chase of flying
ants." The 'gregarious' and 'dancing' parties are a confusion of two things.
First, there are the family groups consisting of the breeding pair and its more
Fig. 1 Territories of Lanius excubitorius at Makerere, 1960-63.
GREY-BACKED FISCAL SHRIKES 203
or less mature brood or broods before the latter disperse. These are the indivi-
duals likely to be found amicably chasing flying ants together. Second, there
are the rival family units found engaged in territorial display (' dancing about')
and where several territories meet and a common intruder appears, as many
as twenty birds may be seen altogether, but each family group defends a
corner of its own territory. Only van Someren (1962) has hinted before that
the so-called 'dancing' behaviour may be territorial in function. His passing
reference ("Fiscal Shrikes parcel out their territories and keep up a rasping
chatter") was published in a commercial brochure and although he had evident-
ly assembled supporting observations at about the same time as the present
author (van Someren, 1962, in litt.), unfortunately he did not live to publish
Observations on Territorial Behaviour and Territories
Territorial display and defence. Territorial defence is carried on by charact-
eristic behaviour consisting of self-advertisement, threat, display and fighting.
Self-advertisement often takes the form of perching on high, exposed or other-
wise conspicuous positions within the territory and, as in the case of the Ameri-
can Loggerhead Shrike (Miller, 1931, p. 157) it forms part of the feeding
behaviour. It was noticed, however, that Grey-backed Fiscals tended to spend
more time, especially in the morning and evening, feeding near those bounda-
ries where neighboring shrikes were active. This is perhaps another explanation
for the 'large parties' which are said to assemble 'in the chase of flying ants'.
Another form of self-advertisement is boundary patrolling. The birds fly
along the boundary calling, whilst a neighboring territorial group is dis-
playing against an intruder.
From a repertoire of half a dozen or so calls, only two are concerned with
territorial display. The first is the alert or threat call, uttered at the beginning
of a territorial threat or display as one of the birds alerts its mate and then
they both fly towards the intruder. The call is a metallic monosyllable which
may be rendered as kyoir-l. As the defending birds, one following the other, go
for the intruder, the alert call changes to the display call which sounds like
kyoir-l, kyoi, kyo-ooh. This is a dueting chorus, uttered by the displaying birds
repeatedly and out of step with each other. It has been described variously
by different authors, such as harsh cries (Bates, 1930; Mackworth-Praed &
Grant, 1960), rasping, shrill, jarring or noisy chattering (Jackson, 1938;
van Someren, 1962, Williams, 1963) or screaming (Mackworth-Pread &
The flight at the intruder is direct, with rapid wing beats, but as the
defenders get near to it they change to gliding flight. With lowered heads
and open bills, outstretched wings and fanned tails, they present their con-
trasting black and white areas to the maximum view of the intruder, pro-
bably as an intimidation. If the intruder does not leave, the threat changes
to the display proper which consists of the dueting accompanied by the move-
ments usually termed 'dancing'. The defenders alight near the intruder and
start alternately jumping up and down, to a half or one foot into the air.
They take off with wings flapping rapidly and land with heads down, wings
outstretched and tails fanned out. On landing they gently wave the tails up
and down. This is the Babbler-like behaviour of the earlier observers. Both
sexes are equally active in the display, as so are the mature progeny which
may be present, either feeding with the parents or holding sub-territories.
Territorialfighting. Threat and display behaviour is sufficient to dislodge
204 UGANDA JOURNAL
a single, casual intruder. This usually flies away before the defenders arrive,
especially if it has already been chased from an adjacent territory, or it flies
away very soon after it is reached. Depending on how far the intruder flies,
and how deep into their territory it had been, the defenders then either fly
after it uttering the display call, or remain on the perch craning their necks
in its direction of flight. In no case have I ever seen actual fighting with a
single intruder, although this probably does take place. Display may, however,
go on for some time, the longest I have timed being thirty minutes, before
the intruder flees.
Physical fighting seems to be rare and only to take place when terri-
torial transgression is by a mated pair of other shrikes. In a long series of
observations since 1960, physical fighting has only been witnessed three
times, between pairs F and G (1 January 1962), E and G (7 January 1962)
and E and F (8 March 1962). These observations, of which details are not
essential here, were all during the earlier of the two periods of maximum
breeding activity and pair F was actually breeding in January 1962. There
was much overlap and display along the boundaries of these three territories.
In the first two encounters the intruding pairs were soon chased off, but in
the third, where pair E trespassed on territory F, combat occurred after
considerable activity. The defending pair (F) resorted to striking when over-
flying, swooping (six times) and attempting to dislodge them physically had
failed. In the ensuing conflict one of the combating pairs was seen to fall to
the ground locked in the fight, but otherwise no apparent physical harm was
Duration ofterritories. It was next attempted to find out how long the terri-
tories were held by each breeding pair and what their average size was. The
whole of Makerere campus, an estate of 201 acres, was surveyed for breeding
pairs of shrikes and their territories were mapped by noting their display
perches or the points at which they doubled back overhead when chased.
The campus was an ideal place for these observations because it is largely
isolated by zones of wood plantations or peasant cultivations, unsuitable for
the shrikes, on roughly three sides. On the other hand, it has within it plenty
of the birds' favoured feeding habitat, open patches of grass lawns with many
tall lookout perches. Unsuitable areas of compact buildings or gardens and
woods are not too extensive for the birds to overfly (Fig. 1).
At the end of 1960, eight territories (A-H) were found almost exclusively
within, and two others (J and K) with small portions within, the college
estate (Fig. 1). Since that time roughly the same number of territories has
been maintained, probably occupied throughout the year by the same per-
manently mated pairs, although it is difficult to prove this absolutely in the
absence of colour ringing. The only major change was the elimination of
territory F during 1963. The reason for this was not found out, but building
activities southwest of Livingstone Hall in 1962-63 had disturbed the birds'
foraging area. After a period of intense territorial hostility from neighboring
pairs A, E and G, an unsuccessful attempt to raise a brood and following upon
one of the adult birds getting caught in a mist net (J. Kingdon, pers. comm.),
the territory was abandoned and carved up between E and G.
The only other notable, although minor, alteration was a prolonged
boundary conflict in mid-1965 between E and H birds which resulted in an
adjustment in favour of H up to University Road (Fig. 1). This again followed
building operations in territory H during 1962-63 and 1964-65 but it is note-
worthy that building activities elsewhere on the campus did not cause any
GREY-BACKED FISCAL SHRIKES 205
disturbance of the other territories. That territories are held and defended
permanently has also been established from my observations elsewhere and
can be inferred from the remarks of Stoneham (1928) and Jackson (1938).
Size of territories. The average size of territory at Makerere was calculated
from the area of the campus and the total number of territories exclusively or
predominantly located there (Table 1). The portions of the two territories, J
and K, centred outside the campus were presumed to be balanced by the
areas outside the campus used for foraging by birds in the other territories.
Territory size was also estimated in the Queen Elizabeth and Murchison Falls
National Parks by noting the distances between family groups, which was
about 440 yards, and assuming that when observed they were at the centres
of their territories which were presumed to be circular. The calculated territory
size, although slightly smaller, was probably of the same order as at Makerere,
considering the method of estimation.
Territory size in Lanius excubitorius under various habitats
Locality Area in acres No. of territories Territory size in
Makerere 1960-62 201.1 8 25.2
Makerere 1963-66 201.1 7 28.7
Kabanyolo Total 340 4 85*
Grass leys 110-130 4 42.5*
Queen Elizabeth Park 22*
Murchison Falls Park 22*
For methods of estimation see text.
The effect of farming practices is seen at Kabanyolo University Farm,
twelve miles north of Kampala. Except for 110-130 acres maintained as
grass leys over the rest of the area of the farm the birds are deprived of their
typical foraging habitat and territory size is effectively about twice that in the
national parks. In spite of this, the feeding is restricted to the neighbourhood
of the few perches (houses, trees, fences and power lines) although the whole
area is nevertheless defended.
Estimation of Population
Population census on Makerere Campus. During 1961-62 a census of all the
birds was carried out in each of the territories at three to four monthly intervals
(Table 2). Making allowance for the occasions when no counts were made
(those with dashes) and those when one or both of the pair known to be present
were not seen, the shrike population varied between 33 and 37 in all the terri-
tories. Later observations have indicated that, in spite of continued breeding
activity, this population stability at the same level was maintained to 1966.
206 UGANDA JOURNAL
Census of Lanius excubitorius at Makerere 1961-63
Territory December April July October February
Science Faculty (A) 2 3 6 3 2
West Road (B) 4 1 2 5 2
College School (C) 7 7 7 8
Old Mitchell Hall (D) 0 2 2 6 -
Library (E) 2 2 2 2 2
Quarry House (F) 5 3 2 2 0
Botanic Garden (G) 4 3 3 3 2
University Hall (H) 5 5 7 6
Mosque Hill (J) 4 3 3 5 3
Wandegeya (K) 4 2 2 4
TOTAL 33 33 36 28 29
Population estimation in Kampala City. With the information from Makerere,
I attempted to estimate the number of territories and breeding pairs of shrikes
in Kampala City and to check this estimate by mapping the location of the
pairs actually found. Out of the total area of the City then of approximately
5,289 acres, the industrial area and commercial centre occupy about 571
acres which area is excluded for the purposes of the estimate (Fig. 2). At 3.3
birds per territory of 29 acres, there should have been 538 birds in 163 territories.
It proved too difficult to count all the birds, but between 1960 and 1965 only
82 pairs were seen, together with about eight of the Fiscal (Lanius collaris L.)
with which the Grey-backed Fiscal probably competes (Fig. 2). Much of the
area of the city is therefore unsuitable for Shrikes, including the unbuilt area,
which may or may not be under cultivation, and the dense residential sec-
tions with close-set, wooded gardens (e.g. Nakasero and Kololo). In the latter
areas the Grey-backed Shrikes are only found in association with games
fields, public open spaces and other patches of grass however small (e.g. street
islands and verges).
The vocal and group behaviour of the Grey-backed Fiscal, previously
attributed to gregariousness, is here clarified as being territorial, that is social
in the sense of Tinbergen (1953, p. 21) and Wynne-Edwards (1962, p. 14),
but not gregarious. The distinctive 'dancing' is a threat display utilizing the
birds' colour pattern and could be interpreted as ritualized intention behaviour
(Daanje, 1950), derived from balanced urges to attack and to flee. Similar
'dancing' behaviour, which is probably also territorial, has been reported for
the Long-tailed Fiscal, Lanius cabanisi (Jackson, 1938), a closely related species
(Olivier, 1944, pp. 56-57). The pair bond in the Grey-backed, and probably
in the Long-tailed, Shrike is so intimate that it is not surprising the species
have evolved a joint (both sexes )rather than an individual (one sex) threat
display against territorial competitors. Significantly, too, the geographical
ranges of the two species do not meet.
It is interesting that, rare as physical combat is, single intruders were
never seen to be physically attacked, unlike mated pairs. The boundaries of
1000 0 4000
208 UGANDA JOURNAL
the territories, permanently held and defended by both sexes, are marked
by prominent perching and display stations. All foraging, roosting and breeding
activities take place within them. The birds in territories A, E, G and H which
were more closely watched roosted in the same place (tree or locality) over a
period of two or three years (1960-63) but the nesting place, although some-
times coincident, was generally changed from season to season.
The biological functions of territory have been vigorously debated ever
since Howard's book (1920). A recent reviewer concluded that the functions
must be diverse and, because of the complexity of the selective forces governing
behaviour, structure and function, "unqualified answers about the functions
of territory cannot be expected" (Hinde, 1956, p. 363). Territory is evidently
a dispersional convention which helps to regulate the breeding population in
relation to the resources (Wynne-Edwards, 1962) and this seems to be supported
here by the stability of the observed shrike territories and numbers. Further
observations including ringing are, however, necessary to elucidate the biology
and behaviour of this interesting bird.
This paper was written during the tenure of a Rockefeller Fellowship.
I wish to acknowledge the help of my wife and my brother, Aston Kateeba,
in finding shrikes, that of the Uganda National Parks, and of Mr. J. C. Ssebu-
nya of the Department of Geography who re-drew the maps from my original
rough drafts. Mrs. E. Kingdon helped with Italian translations.
BANNERMAN, A. D. (1939) The birds of tropical West Africa, Vol. 5. London, Crown Agents for
BATES, G. L. (1930) Handbook of the birds of West Africa. London, Bale & Danielsson.
DAANJE, A. (1950) On locomotory movements in birds and the intention movements derived
from them. Behaviour 3, pp. 48-98.
HOWARD, H. E. (1920) Territory in bird life, London, Murray.
HINDE, R. A. (1956) The biological significance of territories of birds. Ibis 98, pp. 340-69.
JACKSON, F.J. (1938) The Birds ofKenya Colony and the Uganda Protectorate, Vol. 3. London, Gurney
MACKWORTH-PRAED. C. W. & GRANT, C. H. B. (1960) The Birds of eastern and north-eastern
Africa, Vol. 2 (2nd. edn.). London, Longmans.
MILLER, A. H. (1931) Systematic revision and natural history of the American shrikes (Lanius).
Univ. Calif. Publ. Zool. 38, pp. 11-242.
OLIVIER, G. (1944) Monographie des Pies-Grieches du genre Lanius. Rouen, Lecerf.
STONEHAM, H. F. (1928) Field notes on a collection of birds from Uganda-Part II Ibis pp.
TINBERGEN, N. (1953) Social behaviour in animals. London, Methuen,
VAN SOMEREN, V. D. (1962) Wild life at the source of the Nile. In Nile Centenary Festival
WILLIAMS, J. G. (1963) A field guide to the birds of East and Central Africa. London, Collins.
WYNNE-EDWARDS, V. C. (1962) Animal dispersion in relation to social behaviour. Edinburgh, Oliver
Uganda Journal, 33, 2, (1969) pp. 209-212.
BISHOP HANNINGTON: A MISCELLANY
By H. B. THOMAS, O.B.E.
Over the years the Uganda Journal has given wide coverage to the
death of Bishop Hannington and its aftermath. Major articles are, The last
days of Bishop Hannington, 8 (1940), pp. 19-27 and The death of Bishop
Hannington, supplementary evidence, 23 (1959), pp. 27-37, both by H. B.
Thomas; and Correspondence relating to the death of Bishop Hannington,
12 (1949), pp. 1-22 by Sir John Gray. Other relevant items have more
recently come light and are here brought together.
Dr. John Rowe has kindly made available a copy of the Mombasa to
Mengo, 1892 diary of J. P. Nickisson, a lay member of the C.M.S. party
destined for Uganda under the leadership of Bishop Tucker. Early in Dec-
ember they had reached Mumias and there, after a dramatic search, they
exhumed the remains of Bishop Hannington. The story, somewhat glamourized
perhaps for publication, is told by Bishop Tucker's journal, printed in Church
Missionary Intelligencer, 1893, pp. 272-276. The Rev. A. B. Fisher's contemporary
journal no longer exists, but in a summary prepared by him about fifty years
after the event, which is now in the archives of the C.M.S. in London (Book 2,
1892), he claims that he was the digger who actually struck the box containing
the remains. This claim is substantiated by Nickisson, whose diary is of especial
value as giving a spontaneous contemporary account of the proceedings.
His account has an explicitness which Tucker's rather confused narrative
lacks. He speaks of Hannington's guide as a son of Mumia. This is not impos-
sible although his age was probably well under the estimated 24 years. He
"8 December 1892. Caravan arrived in Mumias .... talk with Mumia.
Giinther was interpreter as he could speak Swahili. The Bishop asked
about Hannington's remains, which he said had been there but had been
removed by Martin and Jackson. He said his son was Kilongosi to Bishop
Hannington. The son was a young sturdy looking fellow of about 24 ....
Afterwards the son came and told us privately that Bishop H's head had
been brought to the village and had been secretly buried under the floor
of an empty hut and that the hut had fallen down and been cleared
away and that he could point out the spot where the head was buried
in a tin box ....
Saturday 10 Dec. We turned out at 6 o'clock and at 6.30 the search for
Bp. Hannington's remains began. The men worked with jembies or
native hoes and cleared the ground away with shovels. The earth was
very hard at first. They had to dig through baked clay of which the fallen
walls were made, also the floor. The men told us that there had been a
small-pox epidemic in that quarter of the village, and that the houses had
been knocked down in consequence ....
210 UGANDA JOURNAL
We found the ground so hard that we tried further on after digging for
about an hour and a half. Dr. Baxter, Fisher and Giinther helped also in
clearing out the earth with the shovels. The Bishop thought they would
be an hour or more before getting deep enough and went to his breakfast.
Gunther and I also went to ours. Just after we had gone they found a
very soft place in the gound, and Fisher cleared out the earth with a
shovel, and suddenly struck something hard.
The earth being cleared away a wooden box was found with a tin lining
which was much decayed and broken, and the tin rusted away. Inside
was a skull and a few rib bones and a thigh bone. The skull was recognized
at once as Bp. Hannington's by the gold stopping in one of the teeth.
The Bishop had been sent for at once on finding the box and received the
remains which were placed in a sack together with the fragments of the
box. Another skull was found near the box, which was buried near the
place. Also found in the hole the top of a canteen bucket much corroded
away with rust, part of some old shoes, the soles much worn, and an
india-rubber hot water bottle which no doubt was the property of
The things with the remains were taken to the Bishop's tent and packed
in a tin-lined box to be taken to Uganda to be buried there. The whole
proceedings were carried out as quietly as possible and very few of the
people really knew what we had been searching for."
J. E. M. Hannington (1877-1950), the eldest child of Bishop Hannington,
was a C.M.S. missionary in Uganda from 1903 to 1918. He was stationed at
Jinja in 1913. Busoga was then at the nadir of its misfortunes. In ten years
sleeping sickness and famine had reduced southern Busoga from one of the
most fruitful regions of eastern Africa to an uninhabited jungle. The site of the
Bishop's death, nearly thirty troubled years before, was in danger of being lost.
His son determined to re-locate it, and an account of his search is to be found
in Church Missionary Review October 1913, p. 640.
"In a private letter from Jinja dated 13 August 1913, the Rev. J. E. M.
Hannington reports that he discovered the very spot where his father
Bishop Hannington was killed. He writes:- An old man came to me
some weeks ago and told me he could get someone to show me the place.
Accordingly he got one of the few remaining sons of Luba, a man who at
the time was a boy. On Thursday (7 August) I went across the bay to
Bukaleba in a canoe, and on Friday they took me to the place. It was a
flat rock sacred to some spirit, near the edge of a swamp.
The man said that one morning he was coming from the other side of the
swamp with cows to herd and saw that something had happened. He went
across to see what it was and found father's body lying there by the side of
the stone. That was all I could get out of him but it is without doubt
true. On the Saturday morning early, we got people together to 'cultivate'
the place and bring stones and while they were doing it, I cut on a tree
close by 'Bishop Hannington was killed here October 29, 1885'. On the
Monday I built a cairn of stones to mark the spot and we planted round
it with grass something like pampas grass but much smaller and smelling
of lemon. It was all I could do then.
It is now quite a lonely spot, though then there were a great many people
around but they have all either died of sleeping sickness or gone away,
and the country is very desolate."
Towards the end of 1926 the Rev. A. E. Clarke, who was with the C.M.S.
in Uganda from 1918 to 1951, reported on Luba's Revisited in Uganda Church
Review, No. 5 (January-March, 1927). He found the area in much the same
state of desolation as it was in 1913.
"Leaving road at Sikiro (Luba's present enclosure): bush track 8-9 miles
south, through densely populated country to a Gombolola chief's place.
Another mile and edge of habitation is reached-in this stretch we cross-
ed the old highway cut by Grant's administration, its outline fairly
discernible as a swathe through bush. Another mile through long grass,
guide said we had reached the spot. Huge rock level with ground;
overgrown heap nearby was visible-overgrowth removed revealed cairn
built by Hannington in 1913. The building of a permanent memorial to
mark the spot should not be lost sight of."
The last journals of Bishop Hannington, edited by E. C. Dawson, was pub-
lished in London by Seeley in 1888. This volume covers, (i) 'At home', 12
June 1883-5 November 1884 (scanty, selected, items only); (ii) 'The Palestine
Journal', 5 November 1884-5 January 1885, when Hannington visited Pales-
tine en route for East Africa. This was, in fact, written up by the Bishop on
his way up-country through Masai-land. After a gap of some seven months
comes (iii) 'The last Journal', 1 August-29 October 1885. Only this last
is directly concerned with events in East Africa.
For this last journey, as is noted in Dawson's introduction, another factual
diary, with anthropological, geographical and similar notes for the use of
future travellers, was also retrieved by Mackay and Ashe in Uganda. All
were sent to England.
Yet after wide enquiry the whereabouts of the bulk of this original material
cannot now be traced. Namirembe Cathedral has the original journal for
1-30 September, and C.M.S. Library, London, that for 1-29 October 1885.
In each case this is a small Letts' Diary for one month only with pages measuring
about 5 by 3 inches. The Bishop's handwriting is minute but legible. An
'exact size' reproduction of a page is illustrated on page 381 of Dawson's
James Hannington: a history of his life and work, London, Seeley, 1887. All else,
including the 'factual' diary is missing. Nor has the diary kept by the Bishop's
chaplain, the Rev. W. H. Jones, who brought the news of his death to the
coast, come to light.
Mackay and Ashe in Uganda had made copies of the diaries before
sending them to England. Thus Ashe had his own transcript when he came
to write Two kings of Uganda in 1889. In chapter XVI he draws attention,
disapprovingly, to various editorial changes made by Dawson. A comparison
of the surviving originals with Dawson's printed text indicates the general
intention of Dawson's changes. He sought to 'refine' Hannington's text to the
possible susceptibilities of a missionary-hearted readership, thereby sacrificing
something of Hannington's rugged spontaneity. Details of difficulties with
carriers, and of game shooting successes are for instance omitted. But it can be
said that no changes of substance have been introduced.
212 UGANDA JOURNAL
Relics. For many years the C.M.S. in London held a collection of Hanning-
ton's personal relics for exhibition purposes. In 1962 all were presented to the
Church of Uganda. The Bishop's compass and wooden chalice are now in
Namirembe Cathedral; the remainder of the material, which includes his
brazier, kettle, saucepan and coffee pot, is lodged with the Uganda Museum.
The Ichabod flag which was carried before the Bishop's caravan on its
return to the coast is in the possession of the C.M.S. in London. This, 'home-
made' by his followers, measures 33 inches wide by 24 deep. It is of dark blue
trade cloth on which white americani letters ICHABOD are sewn.
Episcopal Seal. Early in 1962 a most unlikely visitor called at C.M.S.
headquarters in London. He was an 'East-end' dealer, more at home perhaps
in Petticoat Lane than in Salisbury Square. From his bag he produced and
offered for sale what was unquestionably Bishop Hannington's original seal.
The detective flair which led him to discover the association of the Bishop with
the C.M.S. must be commended. He was clearly marked out for a successful
commercial career. He hoped to sell it as a unique paper-weight. Dr. Pos-
nansky, then Curator of the Uganda Museum, happened to be in England and
was alerted. He was able to negotiate its purchase and it is now in the Uganda
Museum. It is in the original case of its maker, 'Wyon Regent Street',' one of
the outstanding medalists of that day. The steel matrix is of magnificent
craftsmanship; it bears the Diocesan arms and the wording SEAL OF JAMES
HANNINGTON D.D. BISHOP IN EASTERN EQUATORIAL AFRICA
How did this come into the dealer's hands? Not surprisingly he was
uncommunicative, and one is thrown back on conjecture. As a hypothesis I
suggest that the seal may have been among the Bishop's effects brought down
the coast by Chaplain Jones. It was not a clearly personal possession to be
returned to his widow in England. One of the few missionaries at Frere Town
may have put it on one side to be handed to Hannington's successor. Then
at short notice he may have been invalided; with his other possessions the seal
was hurriedly packed up and put on board a home-bound steamer. For years
it may have lain in a battered uniform case in the attic of a country rectory.
Came the inevitable end: removal or death necessitated the disposal of the
debris of a life-time, and some perceptive second-hand dealer seized his oppor-
tunity to acquire the Bishop's seal.
1. The name occurs mis-spelt as Lyon in Lord Twining's article, Uganda medals and decora-
tions, Uganda Journal, 2, 1934-5, p. 212.