Research Article
Floristic classification of the vegetation in small
wetlands of Kenya and Tanzania
Miguel Alvarez, Mathias Becker, Beate Böhme, Collins Handa, Matthias Josko, Hellen
W. Kamiri, Matthias Langensiepen, Gunter Menz, Salome Misana, Neema G. Mogha,
Bodo Maria Möseler, Emiliana J. Mwita, Helida A. Oyieke & Nomé Sakané
Abstract: Small wetlands in East Africa are increasingly converted into sites for agricultural production. The resulting changes in land
use and cropping systems will impact on the wetlands’ vegetation. We characterized the plant communities in four wetlands of Kenya
and Tanzania, each comprising four types of land use differentiated by the degree of anthropogenic disturbance (cropland, fallow,
grazing land and unused). Since no syntaxonomical scheme was available as a reference, a first classification of vegetation units and
the identification of diagnostic species is proposed. We collected 207 relevés in the representative wetlands in relation to the current
land uses. The plant communities were determined using a modified TWINSPAN classification. For each vegetation unit, diagnostic
species were selected according to their fidelity index (phi coefficient). Floristic relationships between vegetation units were surveyed
by nMDS ordination analyses. We identified 15 plant communities and selected 147 diagnostic species. The communities were differentiated into (1) semi-natural wetland vegetation (associated with less disturbed environments), (2) grassland and fallow vegetation,
and (3) weed communities (associated with eu-hemerobic, drained and cultivated cropland). While the semi-natural vegetation was
distinctly matched with unused fields, the differential matching of the other plant communities with land use types was less clear. According to the floristic similarity, the weed communities associated with cropland tended to be aggregated in the nMDS configuration
while the semi-natural vegetation was dispersed. The results of the ordination did not differ when involving all species or only the selected diagnostic ones. As the plant communities described are rankless syntaxa, the establishment of a comprehensive syntaxonomic
classification for African wetlands will require further vegetation surveys as well as their comparison with published data.
Keywords: flood plain; inland valley; land use; modified TWINSPAN; papyrus swamp.
Nomenclature: African Plant Database (CJB & SANBI 2010).
Abbreviations: nMDS = non-metric multidimensional scaling; SWEA = Small Wetlands in East Africa; TWINSPAN = Two-way indicator species analysis.
Received: 14 November 2010 – Accepted: 29 September 2011 – Co-ordinating Editor: Jürgen Dengler.
Introduction
Wetlands are important landscape elements as they fulfil a diverse set of ecological and social and economic functions. In East Africa, reported wetland
services include preservation of biodiversity, pollutant buffering, water supply for
consumption and irrigation, sites for hunting and gathering, collection of thatching
material and medicinal plants and sites for
agricultural production, including cattle
grazing and various cropping strategies
(Chapman et al. 2001, Denny & de Ruyter
van Steveninck 2001, Junk 2002). Prolonged periods of water (or soil moisture)
availability and higher soil fertility in
comparison to the surrounding uplands
provide a large potential for the expansion
and intensification of agricultural production (van der Heyden & New 2003).
However, the conversion of wetlands into
sites of agricultural production is threatening their ecological functions, particularly, when conversion is associated with
large-scale drainage. Other factors associated with anthropogenic disturbances and
driving changes in the plant ecosystems
involve eutrophication, pollution and
salinization through agrochemicals and
waste from settlements, and hydrological
changes due to excessive water extraction
for industrial or household purposes
(Kalinga & Shayo 1998, Kiai & Mailu
1998, Junk 2002). These disturbance factors and the associated vegetation forms
and species compositions are seen to be
used to assess ecosystem health and the
wetlands' ability to fulfil ecological functions.
While large wetlands like those of Lake
Victoria (Gaudet 1977, Harper et al.
1995) or Lake Naivasha (Kassenga 1997,
Kairu 2001) have been the focus of many
reviews on the impacts of the land use on
their ecological functions, little attention
has been paid to the study of vegetation
and ecology of East African wetlands
with sizes smaller than 10 km². Such
small wetlands show a continuous and
seasonally variable transition between
aquatic and terrestrial ecosystems and
may be permanently or temporary flooded
(Aselmann & Crutzen 1989). Conversion
In: Dengler, J., Oldeland, J., Jansen, F., Chytrý, M., Ewald, J., Finckh, M., Glöckler, F., Lopez-Gonzalez, G., Peet, R.K., Schaminée, J.H.J. (2012)
[Eds.]: Vegetation databases for the 21st century. – Biodiversity & Ecology 4: 63–76. DOI: 10.7809/b-e.00060.
63
of semi-natural wetlands into sites of production will alter the hydrological regime
and the trophic level of soils and water,
which in turn is likely to affect the vegetation. For a better understanding of the effects of land uses in East African wetlands, it is necessary to characterize the
current state of the vegetation in terms of
species composition of the phytocoenoses
and its relation with different forms of
land uses.
While responses of single species to allogenic changes can be used as indicators
of disturbance (Gleasonian approach; van
der Valk 1981), the first step must involve
the characterization of vegetation units
(Clemensian approach), which can potentially provide vegetation groups associated with the level of anthropogenic disturbances as well as their interactions with
geographical and environmental factors.
Assessing the current state of wetland
vegetation thus requires the characterization of both the structure and the species
composition. One of the most comprehensive methodologies for vegetation classification is offered by phytosociology
(Dengler et al. 2008). Resembling traditional subjective procedures, new statistical approaches have been developed to
make phytosociological classifications
more objective. These may involve the
implementation of statistical fidelity
measures (Bruelheide 2000, Chytrý et al.
2002, Tichý & Chytrý 2006, Willner et al.
2009), the improvement of the TWINSPAN algorithm (Roleček et al. 2009),
and the development of a specialised
software like Juice (Tichý 2002), which
executes all mentioned statistical analyses. Additionally Willner (2006) and
Dengler et al. (2008) offer conceptual reviews on the association concept and phytosociology, respectively.
In contrast to the situation in Europe
where both an established syntaxonomical
scheme (Rodwell et al. 2002) and a large
number of relevés are available (Schaminée et al. 2009, Dengler et al. 2011),
vegetation surveys in Kenya and Tanzania
have been mainly carried out in dry ecosystems such as savannas, bushlands and
grasslands (Barkham & Rainy 1976,
Bronner 1990, Cornelius & Schultka
1997). More detailed floristic information
is only available for papyrus swamps
(Gaudet 1977, Kassenga 1997, Kairu
2001, Owino & Ryan 2007). However,
these studies are based only on a simple
listing of plant species or general descriptions of plant formations. The first proposal for a general classification of East
African vegetation was made by Phillips
(1930). This classification, however, did
not follow the criteria of phytosociology,
and it is currently outdated. This is particularly true for the small wetlands that
are subject to most intense land use dynamics. The integrated research project
“Agricultural Use and Vulnerability of
Small Wetlands in East Africa” (SWEA)
provided the opportunity to improve our
knowledge about species commonly encountered in wetlands with different attributes and under various types of land
use in Kenya and Tanzania. The main aim
of this work was the development of a
floristic classification of small wetland
vegetation under different land uses and
use intensities based on relevés with the
aims (1) to derive a vegetation classification according to species composition, (2)
to establish the relationships between
vegetation units and wetland uses, and (3)
to determine the distribution of vegetation
communities in different wetlands.
Materials and methods
Study site
According to a first survey on geophysical, socio-economical and vegetation attributes of small wetlands in East Africa
(Sakané et al. 2011), we selected four
contrasting wetlands to cover the prevailing diversity of attributes such as wetland
type (floodplain vs. inland valley swamp),
altitude (lowland vs. highland), demography (high vs. low population density in
the surroundings), and market accessibility (Table 1). Rumuruti is located in the
Ewaso Narok floodplain on the Laikipia
plateau at 1,811 m a.s.l. and is characterized by a semi-arid climate. Tegu represents an inland valley wetland of the humid highlands and is located close to
Mount Kenya at 1,722 m a.s.l. The two
Tanzanian wetlands are Malinda located
in the sub-humid Mkomazi floodplain at
357 m a.s.l., and Lukozi, an inland valley
in the humid Usambara highlands at
1,765 m a.s.l. (Fig. 1).
As no long-term meteorological registers are available for the study sites, we
described the respective climate on the
basis of data from the interpolation model
offered by the International Water Management Institute at http://www.iwmi.
cgiar.org/WAtlas (New et al. 2002), combined with own data collected during the
field works and registers of meteorological stations located in physical proximity
to the study sites. Rainfall in Malinda is
concentrated in one rainy season between
March and June (Kamiri 2010). Annual
rainfall reaches 1,000 mm, while mean
monthly temperatures oscillate between
23 °C in July and 31 °C in February
(mean temperature 25.2 °C). The other
localities are characterized by a bimodal
rainfall pattern, with a long rainy season
between March and June and a short rainy
season between September and November. In Rumuruti the mean monthly temperatures range between 16 °C and 20 °C
and the annual precipitation is about 500
mm (Thenya 2001). Monthly temperatures in Tegu are very similar to those in
Rumuruti, but the annual precipitation is
higher (between 1,000 and 1,500 mm). In
Lukozi the monthly temperatures oscillate
between 15 °C and 21 °C (mean 18.6 °C)
with an annual precipitation of approximately 1,000 mm (Kamiri 2010).
Table 1: Geographic and physical attributes of the studied wetlands in Kenya and Tanzania. Climate type according to the classification of Köppen.
Wetland
country
coordinates
Altitude (m a.s.l.)
landform
area (ha)
sampled plots
climate type
population
market access
64
Rumuruti
KE
0° 17' 18" N
36° 33' 57" E
1,811
flood plain
917.27
53
Cwb
dense
high
Tegu
KE
0° 28' 26" S
37° 06' 08" E
1,722
inland valley
19.71
42
Cwb
sparse
low
Malinda
TZ
5° 06' 20" S
38° 20' 29" E
357
flood plain
792.20
49
Aw
sparse
low
Lukozi
TZ
4° 39' 29" S
38° 15' 11" E
1,765
inland valley
103.08
63
Cwb
dense
high
Biodiversity & Ecology 4
2012
Fig. 1: Location of the studied wetlands (yellow circles) in Kenya and Tanzania. Digital elevation model obtained from the
CGIAR-CSI SRTM 90 m Database (Jarvis et al. 2008). Shape files of lakes, streams, and country borders downloaded from the
Digital Chart of the World (URL: http://www.maproom.psu.edu/dcw).
In the Malinda and Rumuruti floodplains, the dominant soil types are Fluvisols and Vertisols, while the inland valley
swamps of Tegu and Lukozi are characterized by Gleysols and Histosols, and, in
the case of Lukozi, with some colluvial
over-lay of gneiss and granite material
originating from adjacent steep valley
slopes (ARI Mlingano 2006, Kamiri
2010).
While all land use types are encountered at each wetland site, agricultural
land use differs between sites with a predominance of intense horticultural production in the peri-urban setting of Lukozi, small-scale subsistence food cropping in the rural setting of Tegu, and cattle grazing in the sub-humid and semi-arid
savanna environments of Rumuruti and
Malinda. Also the type of flood-tolerant
crops differs by altitude with lowland rice
(Oryza sativa) in Malinda and Taro
(Colocasia esculenta) in Tegu (Sakané et
al. 2011).
Biodiversity & Ecology 4
2012
Data collection
To cover the prevailing diversity of vegetation and land uses in each wetland, the
vegetation was sampled in 207 rectangular plots, each measuring 10 m2. The plots
were located in areas of homogeneous
vegetation, and in the case of cropped
fields, were managed by one farmer
(preferential sampling; Dengler et al.
2008). The land use of each plot was assessed according to field observations and
interviews with the owners and was classified into four main types: (1) crops (cultivated plots), (2) grazing areas, (3) temporarily abandoned fallows, and (4) unused plots dominated by semi-natural
vegetation. The species composition of
each plot was assessed through relevés
according to the Braun-Blanquet method
(Wikum & Shanholtzer 1978, Dengler et
al. 2008). All vascular plant species present in each plot were listed and their
abundance was recorded following visual
estimates of the respective percentage
cover. Cover ratings of 6 to 100% were
scaled in 10% steps, while cover ratings
of 1 to 5% were scaled in 1% steps. Cover
values below 1% were registered either as
sparse (+: more than one individual) or
rare (r: only one individual). In farmland
plots, the crop species itself was not considered in the estimation of cover. Unknown specimens were collected and
identified at the East African Herbarium
of the National Museums of Kenya in
Nairobi (EA) and at the Herbarium of the
University of Daar es Salam in Tanzania
(DSM), and using the determination keys
of the Flora of Tropical East Africa (Polhill 1949–2010). For sedges and rushes
the key of Haines & Lye (1983) was used.
Bryophytes, algae and lichens were not
included in the sampling. The relevés
were stored in SWEA-Dataveg, a vegetation database in TURBOVEG format (Alvarez et al. 2012). This database is registered with ID AF-00-06 in the Global Index of Vegetation-Plot Databases (GIVD;
Dengler et al. 2011).
65
Statistical analysis
For a floristic classification of the relevés
stored in SWEA-Dataveg, a vegetation
matrix including the original cover values
was exported into the Juice software
(Tichý 2002) to use the modified TWINSPAN algorithm (Roleček et al. 2009).
The minimum group size was set at 3 and
the threshold levels of cover were set at
0%, 5% and 50%. Whittaker’s beta diversity index (Whittaker 1972) was used for
the analysis of heterogeneity of groups, as
it provides balanced classifications, respecting group size and heterogeneity, but
also because of its robustness (Roleček et
al. 2009). The phi coefficient of association was calculated after standardizing the
size of the groups according to Tichý &
Chytrý (2006) to compile the lists of diagnostic species for each relevé group
derived from the TWINSPAN analysis.
To select diagnostic species, we set an
arbitrary threshold value of 0.32, which
provided a list of diagnostic species for
more than one group that allows the detection of species that link related relevé
groups. Only those species were considered as diagnostic that showed a significant concentration in the respective vegetation unit at α = 0.05 according to
Fisher’s exact test (Chytrý et al. 2002).
To analyse floristic similarities between
the vegetation units, relevés were pooled
within groups and the mean value of the
percentage cover (absences not considered) and the percentage constancy of
each species in each vegetation unit was
calculated (importance value; Wikum &
Shanholtzer 1978). The resulting matrix
was analyzed using a non-metric multidimensional scaling (nMDS) restricted to
two dimensions, to allow an easy visual
interpretation of the scatter plot (Lepš &
Šmilauer 2003, Leyer & Wesche 2007),
and measuring the distance between
groups according to the Bray-Curtis index
(Bray & Curtis 1957). This index was selected because of its usual robustness
when ecological factors are measured
(Faith et al. 1987). The same procedure
was applied by using only the diagnostic
species. Both ordinations (considering all
species and only diagnostic ones) were
compared using procrustes rotation
(Peres-Neto & Jackson 2001, Olden &
Jackson 2001). The software Juice (Tichý
2002) was applied for the classification
analysis, the calculation of the phi coefficient of association, the testing of significance levels, and the edition of the summary tables. Non-metric multidimensional
scaling (nMDS) and the procrustes rota-
66
tion were performed in the computing
environment R (R Development Core
Team 2009), using the package “vegan”
(Oksanen et al. 2009).
Results and discussion
Vegetation units and diagnostic
species
We recorded 400 plant species in the 207
relevés. Of those species, 266 occur in
Kenya and 286 in Tanzania. Considering
the single localities, 206 species were recorded in Rumuruti (highland floodplain
in a semi-arid environment), 188 in Malinda (lowland floodplain in a sub-humid
environment), 180 in Tegu, and 152 in
Lukozi (both highland inland valley sites).
Some 315 species were present in cropped
fields, 301 in fallow land, 116 in grazing
areas, and 111 in the unused wetland areas. According to the TWINSPAN classification, we obtained 15 relevé groups.
Ten of those groups include plots of a
single locality with its specific land use,
while the other groups cut across several
localities or land uses (Table 2). Considering a threshold value of 0.32 for the phi
coefficient, 147 diagnostic species were
selected for the different vegetation units
(Table 3). The number of diagnostic species per vegetation unit ranged from 2
(groups 1 and 3) to 31 (group 5). The
relevé groups are described as follows:
Group 1: Typha capensis community
This group occurred in seven plots and
contained two diagnostic species: Epilobium hirsutum and Typha capensis. This
is the only cluster containing exclusively
unused plots, distributed in both Tanzanian wetland types, the floodplain in Malinda and the inland valleys in Lukozi
(Plate D). This vegetation unit has very
low species diversity, dominated by T.
capensis as the common pioneer for eutrophic wetland vegetation.
Group 2: Cyperus papyrus-Cyperus exaltatus community
The group was encountered in ten plots
and contained nine diagnostic species,
namely Cyperus exaltatus, C. papyrus,
Grangea maderaspatana, Heliotropium
indicum, H. steudneri, Ipomoea aquatica,
Mimosa pigra, Phragmites australis, and
Trianthema portulacastrum. This group
included both fallow plots and unused
areas and was mainly associated with
oligotrophic permanently flooded sections
of floodplains (Plate E). The most impor-
tant diagnostic species C. papyrus reached
highest cover percentages in undisturbed
areas. Despite a low fidelity, Typha capensis was also occasionally encountered
in this community, mainly in areas experiencing eutrophication. Other hydrophytes in this group include Ph. australis,
M. pigra, C. exaltatus and I. aquatica.
From the list of common species associated with papyrus swamps in Lake Naivasha (Gaudet 1977), sole C. papyrus
was encountered in the present survey.
The occurrence of group 2 in fallow areas
of the lowland floodplain in Malinda was
associated with areas formerly used for
paddy rice production.
Group 3: Cyperus exaltatus-Callitriche
oreophila community
This group was encountered in five plots
and contained only two diagnostic species: Callitriche oreophila and Cyperus
exaltatus. The group is dominated by
grazing areas, all located on the fringe of
the highland floodplain of Rumuruti. In
general, all species are representing strong
anthropogenic disturbances as indicated
by the dominance of C. exaltatus and the
ubiquitous perennial grass Cynodon dactylon.
Group 4: Pennisetum mezianumSporobolus pyramidalis community
This group occurred in ten plots with 13
diagnostic species, namely Cycnium
adonense, C. tubulosum, Heliotropium
steudneri, Hyphaene compressa, Pennisetum mezianum, Pluchea dioscoridis, Sida
acuta, Sphaeranthus steetzii, Sporobolus
pyramidalis, Tragia furialis, Vernonia
amygdalina, V. colorata, and V. glabra.
The group includes plots that were either
used as grazing lands or left to fallow and
occurred mainly at the two floodplain
sites (Malinda and Rumuruti, see Plate C)
in semi-arid or sub-humid environments
and high densities of grazing cattle. The
presence of S. pyramidalis and some
woody species like S. acuta and H. steudneri are indicative of the heavy grazing
pressure. As in group 3, Cynodon dactylon was frequently associated and sometimes dominating the plots.
Group 5: Pentodon pentandrus-Leersia
hexandra community
Group 5 was encountered in 29 plots.
With 31 diagnostic species, this group
was the most diverse vegetation unit
(mean species richness of 29). Diagnostic
species comprised Abildgaardia buchananii, Acmella uliginosa, Asystasia
gangetica, Azolla pinnata, Basilicum
Biodiversity & Ecology 4
2012
polystachyon, Cyperus distans, C. latifolius, C. laxus, Dryopteris inaequalis,
Eclipta prostrata, Enicostema axillare,
Ethulia conyzoides, Euphorbia hirta, Ficus sur, Fuirena umbellata, Hibiscus
cannabinus, Ipomoea aquatica, I. sinensis, Leersia hexandra, Ludwigia octovalvis, Melanthera scandens, Melochia
corchorifolia, Panicum maximum, Pentodon pentandrus, Pluchea dioscoridis,
Rhus pyroides, Sesbania speciosa, Sorghum arundinaceum, Torulinium odoratum subsp. auriculatum, Vigna vexillata,
and Zehneria scabra. This large species
diversity was also reflected in the diverse
land uses encountered here comprising
fallow plots, croplands, and unused areas.
However, all plots occurred in the lowland floodplain of Malinda (Plate F),
characterized by a strong seasonality of
flooding. Consequently, diagnostic species represented plants typical of an
aquatic habitats in the wet season (e.g. A.
pinnata, I. aquatica, L. octovalvis and
Cyperus spp.) as well as plants of dryer
environments typical during the dry season or located on the floodplain fringe
(e.g. I. sinensis, a savanna species according to Bronner 1990). This group also includes diverse weed species typically associated with lowland rice (e.g. Leersia
hexandra) or occurring in small drainage
canals between field plots.
Group 6: Trifolium semipilosum-Sida
schimperiana community
This group occurred in seven plots and
comprised 13 diagnostic species: Aeschynomene schimperi, Cyathula uncinulata, Crotalaria juncea, Croton megalocarpus, Echinochloa colona, E. haploclada, Eragrostis tenuifolia, Panicum
coloratum, Sida schimperiana, Solanum
incanum, Sphaeranthus suaveolens, Trifolium semipilosum, and Vernonia
poskeana. While this group was associated with all land uses, it dominated in
very wet fallow and grazing areas and
was restricted to the highland floodplain
site of Rumuruti. The vegetation unit is
dominated by grass species of humid environments and Trifolium. semipilosum.
Among the typical indicators for high soil
moisture are Echinochloa haploclada
(mentioned by Phillips 1930 as a typical
pioneer grass on moist alluvium) and
Sphaeranthus suaveolens (Gaudet 1977).
In addition, Croton megalocarpus is a
species of the Croton forest (Muasya et al.
1994), which together with other woody
species like A. schimperi, S. schimperiana, and S. incanum are indicative of
Biodiversity & Ecology 4
2012
Table 2: Composition of the relevé groups according to the localities (a) and main
land uses (b).
Groups
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Total of plots
Groups
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Total of plots
Rumuruti
2
5
2
7
22
2
8
1
4
53
Cropping
lands
12
2
4
22
20
1
13
4
1
24
103
(a) Localities
Tegu
Malinda
3
8
8
29
4
38
1
42
49
(b) Main land uses
Grazing
Fallows
lands
6
1
3
6
4
14
3
1
9
3
4
4
4
1
3
2
12
62
18
overgrazing in wet environments (Bronner 1990).
Group
7:
Sacciolepis
africanaOxygonum sinuatum community
Goup 7 was restricted to only four plots,
however with 22 diagnostic species: Acacia mearnsii, Ageratum conyzoides, Amaranthus blitoides, Bidens biternatus,
Commelina erecta, Crotalaria agatiflora,
C. lebrunii, Eragrostis tenuifolia, Galinsoga quadriradiata, Haplosciadium abyssinicum, Justicia betonica, J. flava, Oxalis
corniculata, O. latifolia, Oxygonum sinuatum, Melinis repens, Phyllanthus amarus, Physalis peruviana, Sacciolepis africana, Stellaria sennii, Tagetes minuta,
and Tephrosia villosa. While group 7
represents a very diverse vegetation unit
(mean species richness of 32) it is only
associated with maize or vegetable fields
in completely drained highland inland
valleys. The majority of the diagnostic
species in this cluster are typical upland
weeds, most of them classified as dry
bushland species (Bronner 1990), and
Lukozi
4
16
4
3
36
63
Unused
fields
7
4
1
3
1
8
24
Plots per
group
7
10
5
10
29
7
4
22
40
9
17
4
4
3
36
207
Plots per
group
7
10
5
10
29
7
4
22
40
9
17
4
4
3
36
207
many originating from North and South
America (e.g. G. quadriradiata, P. peruviana and T. minuta).
Group 8: Sonchus oleraceus-Galinsoga
quadriradiata community
Group 5 was encountered in 22 plots but
with only five diagnostic species, namely
Crotalaria lebrunii, Galinsoga quadriradiata, Schkuhria pinnata, Sonchus oleraceus, and Trifolium lugardii. Similar
to group 7, this cluster included only
croplands, however not from an inland
valley but rather the highland floodplain
of Rumuruti (Plate B). Many of the diagnostic species are also typical upland
weeds and the two naming species of the
cluster (S. oleraceus and G. quadriradiata) are characteristic of weed communities in Europe (Oberdorfer 2001).
Group 9: Hydrocotyle sibthorpioides
community
This is the largest cluster group, encountered in 40 plots and comprising 23 diagnostic species: Acmella caulirhiza, Agera-
67
tum conyzoides, Ammannia baccifera,
Lythrum rotundifolium, Coelachne africana, Crassocephalum vitellinum, Cyperus esculentus, Dichondra repens, Galium
aparinoides, Gymnanthemum auriculiferum, Hydrocotyle sibthorpioides, Kyllinga brevifolia, K. pumila, Lantana
camara, Ludwigia abyssinica, Oxalis corniculata, Panicum trichocladum, Pycreus
nigricans, Rorippa nasturtium-aquaticum,
Sida ovata, Stellaria sennii, Trifolium
rueppellianum, and Triumfetta flavescens.
This group occurs in all land uses, however with a predominance of upland field
crop plots located in inland valleys (Plate
A). Some of the diagnostic species are
typical upland or fallow weeds (e.g. A.
conyzoides), while the majority are repre-
sentative in wet grasslands (e.g. D. repens, H. sibthorpioides, Kyllinga spp. and
L. rotundifolium) and Panicum trichocladum is typical for pioneer grass communities in moist, oligotrophic alluvial soils
(Phillips 1930). In contrast, Panicum
maximum has a broader ecological spectrum ranging from the savanna to lower
forest regions (Bronner 1990).
Plate: Pictures of some of the sampled fields in the small wetlands from Kenya and Tanzania.
A: Landscape of the Tegu wetland showing in foreground a taro field (Colocasia esculenta), commonly colonized by the
Hydrocotyle sibthorpioides community (Photo M. Alvarez).
B: Field of maize (Zea mays) in Rumuruti with the weeds of the Sonchus oleraceus-Galinsoga quadriradiata community
(Photo M. Alvarez).
C: Grassland dominated by Sporobolus pyramidalis in Malinda. Characteristic of this wetland is the doum palm (Hyphaene
compressa) with branched trunks. In the background the Usambara Mountains are visible (Photo M. Alvarez).
D: Terrace on the fringe of a Typha capensis stand in Lukozi. In the foreground (left corner) Polygonum usambarensis with
its characteristic dark pink flowers can be seen (Photo M. Alvarez).
E: Cyperus papyrus stand in Rumuruti with signs of clearcut and trampling by grazing animals (here mainly cattle) (Photo
M. Alvarez).
F: Rice paddies in Malinda. In the background there are also individuals of bananas (Musa paradisiaca), coconut palms
(Cocos nucifera), and mangoes (Mangifera indica) (Photo M. Alvarez).
68
Biodiversity & Ecology 4
2012
Table 3: Summary table of the relevé groups showing the constancy (left) and the fidelity (right) of the plant species. Only positive fidelity values (phi coefficient) are shown. Also non-significant fidelity values (according to a Fisher’s exact test at α = 0.05)
are skipped. The table is shortened, excluding non-diagnostic species (companions). Diagnostic species (values grey-shaded)
were selected applying a threshold of phi ≥ 0.32.
2
10
12
9
3
5
8
2
(a) percentage
4
5
6
7
8
10 29 7
4 22
18 29 27 32 27
13 31 13 22 5
Group 1: Typha capensis community
Epilobium hirsutum
43 .
Typha capensis
100 60
.
.
.
.
10 38
Relevé group
Nr. of relevés
Mean nr. of species
Nr. of diagnostic species
1
7
3
2
frequency
9 10 11 12 13 14 15
40 9 17 4
4
3 36
32 13 16 27 17 21 24
23 3
7 14 8
5
9
2
10
12
9
3
5
8
2
(b) phi coefficient x 100
4
5
6
7
8
9 10 11 12 13 14 15
10 29 7
4 22 40 9 17 4
4
3 36
18 29 27 32 27 32 13 16 27 17 21 24
13 31 13 22 5 23 3
7 14 8
5
9
. 64 -14 60 32
---
---
-16
---
---
---
---
---
---
---
---
---
---
1
7
3
2
Diagnostic species of single groups
.
.
.
.
.
.
.
.
.
.
.
24
.
.
.
.
.
.
Group 2: Cyperus papyrus -Cyperus exaltatus community
Cyperus papyrus
. 100 40 . 34 43
Trianthema portulacastrum
. 50 .
. 10 .
Phragmites australis
. 50 . 20 3
.
Heliotropium indicum
. 30 . 10 .
.
Mimosa pigra
. 30 . 10 7
.
Grangea maderaspatana
. 20 .
.
.
.
.
.
.
.
.
.
5
.
.
.
.
9
.
8
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
-------
64
59
56
45
41
35
-------
-------
15 21
-- --- --- --- --- --
-------
-------
-------
-------
-------
-------
-------
-------
-------
Group 3: Cyperus exaltatus -Callitriche oreophila community
Callitriche oreophila
.
. 20 .
.
.
.
.
.
.
.
.
.
.
--
--
44
--
--
--
--
--
--
--
--
--
--
--
--
------------
------------
------------
67
63
63
60
56
52
46
44
37
35
34
------------
------------
------------
------------
------------
------------
------------
------------
------------
------------
------------
------------------------
------------------------
------------------------
------------------------
86
82
67
66
64
60
54
51
49
48
48
47
47
46
46
45
44
44
44
43
43
40
39
------------------------
------------------------
------------------------
----------------------21
------------------------
------------------------
------------------------
------------------------
------------------------
------------------------
.
Group 4: Pennisetum mezianum -Sporobolus pyramidalis community
Pennisetum mezianum
.
.
. 50 3
.
.
.
.
.
.
.
Vernonia colorata
. 10 . 50 .
.
.
.
.
.
.
.
Hyphaene compressa
.
.
. 50 10 .
.
.
.
.
.
.
Vernonia glabra
.
.
. 60 .
.
.
.
.
.
.
.
Sporobolus pyramidalis
. 10 . 90 24 .
. 14 13 22 29 .
Cycnium tubulosum
.
.
. 40 . 14 .
.
.
.
.
.
Vernonia amygdalina
. 10 . 40 .
.
.
.
5 11 .
.
Sida acuta
.
.
. 20 .
.
.
.
.
.
.
.
Sphaeranthus steetzii
. 30 . 70 24 43 . 32 18 22 . 25
Tragia furialis
.
.
. 20 .
.
.
.
.
.
6
.
Cycnium adonense
.
.
. 20 .
.
.
.
. 11 .
.
Group 5: Pentodon pentandrus -Leersia hexandra community
Pentodon pentandrus
.
.
.
. 76 .
.
.
.
.
.
.
Melanthera scandens
.
.
.
. 69 .
.
.
.
.
.
.
Ethulia conyzoides
.
.
. 30 69 .
.
.
.
.
.
.
Abildgaardia buchananii
.
.
.
. 45 .
.
.
.
.
.
.
Basilicum polystachyon
.
.
. 10 52 .
.
.
.
.
.
.
Ipomoea sinensis
.
.
. 10 48 .
.
.
.
.
.
.
Ficus sur
.
.
.
. 31 .
.
.
.
.
.
.
Acmella uliginosa
.
.
.
. 28 .
.
.
.
.
.
.
Azolla pinnata
. 10 .
. 38 .
.
.
8
.
.
.
Fuirena umbellata
.
.
.
. 24 .
.
.
.
.
.
.
Enicostema axillare
.
.
.
. 24 .
.
.
.
.
.
.
Cyperus distans
. 10 . 20 45 .
.
.
3
.
.
.
Eclipta prostrata
.
.
.
. 28 .
.
.
5
.
.
.
Hibiscus cannabinus
.
.
. 10 31 .
.
.
.
.
.
.
Panicum maximum
.
.
. 20 38 .
.
.
.
.
.
.
Sorghum arundinaceum
.
.
. 20 38 .
.
5
.
.
.
.
Sesbania speciosa
.
.
.
. 21 .
.
.
.
.
.
.
Melochia corchorifolia
.
.
.
. 21 .
.
.
.
.
.
.
Euphorbia hirta
.
.
.
. 21 .
.
.
.
.
.
.
Asystasia gangetica
.
.
. 10 34 .
.
5
5
.
.
.
Ludwigia octovalvis
.
.
.
. 38 14 .
.
5 11 .
.
Zehneria scabra
.
.
.
. 38 29 .
.
5
.
.
.
Cyperus laxus
.
.
. 10 41 .
.
. 25 . 24 .
Torulinium odoratum subsp.
.
.
. 10 24 .
.
.
3
.
.
.
auriculatum
Dryopteris inaequalis
.
.
. 10 24 .
.
.
5
.
.
.
Rhus pyroides
.
.
.
. 14 .
.
.
.
.
.
.
Vigna vexillata
.
.
.
. 41 14 25 5 15 11 .
.
Cyperus latifolius
.
.
. 20 28 .
.
.
.
.
6
.
Leersia hexandra
.
. 20 10 97 86 100 91 90 .
6 50
Biodiversity & Ecology 4
2012
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
33 .
. 22
.
.
.
.
.
.
.
.
.
3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3
.
.
.
.
.
3
.
.
3
.
.
.
.
3
.
6
3
.
.
.
--
--
--
--
38
--
--
--
--
--
--
--
--
--
--
.
.
.
.
.
.
.
.
.
.
.
.
.
3
.
------
------
------
------
36 -36 -35 -34 -33 27
------
-- --- --- --- -30 30
------
------
------
------
------
------
69
Table 3 (continued).
Relevé group
1
2
3
4
(a) percentage frequency
5
6
7
8
9 10 11 12 13 14 15
(b) phi coefficient x 100
6
7
8
9 10 11 12 13 14 15
1
2
3
4
5
.
.
.
.
25 .
.
.
.
.
.
.
.
.
.
3
.
.
.
.
.
.
.
.
. 25 67 17
.
.
.
.
.
.
. 11
.
.
.
.
.
.
.
.
.
.
.
.
-------------
-------------
-------------
---30
---------
-------------
-------------
-14
-18
13
------27
13
17
----24
----14
-------------
-------------
-------------
-------------
-------------
-------------
30 .
.
.
.
.
.
5 11 . 50 .
.
.
3
.
.
.
.
.
.
3
.
.
.
.
.
.
20 .
.
.
.
.
.
65 11 . 25 .
.
.
10 . 12 . 25 33 56
3
.
. 25 .
.
.
53 .
6
. 25 . 19
8 11 .
.
.
.
.
63 . 18 .
. 67 61
.
.
.
.
.
.
.
35 .
. 75 .
. 17
3
.
.
.
.
.
.
---------------
---------------
---------------
---------------
-- -- 71
-- -- 61
-- -- 59
-- -- 58
-- -- 57
22 -- 55
-- -- 55
-- -- 54
-- -- 54
-- -- 51
-- -- 51
-- -- 49
-- 30 47
-- -- 46
-31
12
-----30
---25
--
24
---19
31
--22
-26
-7
--
---------------
---------------
------------32
--
---------------
---------------
------25
---26
----
---
---
---
---
---
---
---
41 -40 17
---
---
---
---
---
---
---------------------
---------------------
---------------------
---------------------
-- --- --- --- --- --- --- --- --- 31
29 --- --- --- --- --- --- --- --- --- --- --
---------------------
---------------------
86
67
53
50
48
45
43
42
42
42
41
41
38
37
35
35
34
34
34
34
---------------------
---------------------
---------------------
---------------------
---------------------
---------------------
----
----
----
----
----
----
----
----
56
41
34
----
----
----
----
----
Diagnostic species of single groups
Group 6: Trifolium semipilosum -Sida schimperiana
Sida schimperiana
.
.
.
Trifolium semipilosum
.
. 20
Solanum incanum
.
.
.
Panicum coloratum
.
.
.
Aeschynomene schimperi
.
.
.
Sphaeranthus suaveolens
. 10 .
Vernonia posk eana
.
. 20
Cyathula uncinulata
.
. 20
Echinochloa haploclada
.
. 40
Crotalaria juncea
.
.
.
Croton megalocarpus
.
.
.
Echinochloa colona
.
.
.
community
.
. 43 .
. 13
20 . 100 . 41 45
10 . 43 .
5
.
40 . 57 . 27 .
10 10 43 . 18 3
.
. 29 .
9
3
20 21 86 25 5 63
.
. 29 .
.
.
10 7 57 . 14 13
.
. 14 .
.
.
.
. 14 .
.
.
. 10 43 . 36 23
Group 7: Sacciolepis africana -Oxygonum sinuatum community
Acacia mearnsii
.
.
.
.
.
.
Sacciolepis africana
.
. 20 .
.
.
Commelina erecta
.
.
.
.
.
.
Tephrosia villosa
.
.
.
.
7
.
Bidens biternatus
.
.
.
.
.
.
Phyllanthus amarus
. 10 . 20 52 .
Physalis peruviana
.
.
.
. 34 14
Justicia flava
.
.
.
.
.
.
Amaranthus blitoides
.
.
.
.
. 29
Crotalaria agatiflora
.
.
.
.
3
.
Oxalis latifolia
.
.
.
.
.
.
Melinis repens
.
.
.
.
.
.
Oxygonum sinuatum
.
.
.
.
. 71
Justicia betonica
.
.
.
.
.
.
Group 8: Sonchus oleraceus -Galinsoga ciliata community
Trifolium lugardii
.
.
.
.
.
Sonchus oleraceus
.
. 20 .
.
75
100
50
50
50
100
100
50
100
50
100
25
100
25
.
. 18 .
43 75 95 58
Group 9: Hydrocotyle sibthorpioides community
Hydrocotyle sibthorpioides
.
.
.
.
.
. 25
Panicum trichocladum
.
.
.
. 24 .
.
Dichondra repens
.
.
.
.
.
. 25
Lythrum rotundifolium
.
.
.
.
7
.
.
Trifolium rueppellianum
.
.
. 10 .
.
.
Kyllinga pumila
.
.
.
.
3
.
.
Crassocephalum vitellinum
. 10 .
.
. 14 50
Rorippa nasturtium-aquaticum
.
.
.
. 10 .
.
Acmella caulirhiza
.
.
.
. 24 57 .
Ludwigia abyssinica
. 20 .
. 41 14 .
Triumfetta flavescens
.
.
.
.
.
.
.
Galium aparinoides
.
.
.
.
.
.
.
Kyllinga brevifolia
.
.
.
.
.
.
.
Ammannia baccifera
.
.
.
.
7
.
.
Pycreus nigricans
.
.
.
.
.
.
.
Coelachne africana
.
. 20 .
.
.
.
Gymnanthemum auriculiferum
.
.
.
.
.
.
.
Sida ovata
.
.
.
.
.
.
.
Cyperus esculentus
.
.
.
.
.
. 25
Lantana camara
.
.
.
.
7
.
.
Group 10: Hygrophila auriculata community
Hygrophila auriculata
.
.
Acacia elatior
.
.
Cirsium vulgare
.
.
70
.
.
.
.
.
.
.
.
.
.
59
14
9
.
5
5
.
64
14
18
.
64
.
.
.
14
.
.
.
.
14
.
.
23
.
5
23
9
.
.
.
.
14
.
.
.
.
.
.
.
.
.
.
44
11
.
11
.
11
.
22
.
.
11
.
.
.
.
.
.
.
.
6
.
.
.
.
.
.
.
.
.
.
50 50 33
.
3
95 .
.
.
.
.
.
73 .
.
.
.
.
.
48 .
.
.
.
.
.
70 .
6
. 50 33 .
65 .
.
. 50 .
6
25 .
.
.
.
.
.
55 .
.
.
.
.
.
70 . 18 . 50 33 14
73 11 . 25 . 33 3
55 11 .
.
.
.
.
18 .
.
.
.
.
.
18 .
.
.
.
.
.
15 .
.
.
.
.
.
28 .
.
.
.
.
.
18 .
.
.
.
.
6
25 .
.
.
.
.
.
13 .
.
.
.
.
.
13 .
.
.
.
.
.
28 .
.
.
.
.
3
18 .
.
.
.
.
.
.
.
.
33
33
22
.
.
.
.
25
.
.
.
.
.
.
.
.
.
.
56
54
49
47
40
39
38
38
38
37
37
34
----
Biodiversity & Ecology 4
2012
Table 3 (continued).
Relevé group
1
2
3
4
(a) percentage frequency
5
6
7
8
9 10 11 12 13 14 15
(b) phi coefficient x 100
6
7
8
9 10 11 12 13 14 15
1
2
3
4
5
.
6
.
3
. 22
67 42
.
.
------
------
------
------
------
------
------
------
------
------
57
47
46
41
33
------
------
------
--22
23
--
25 .
.
.
25 .
.
.
25 .
.
.
25 .
.
.
25 .
.
.
50 .
. 28
25 .
.
.
25 .
.
.
50 .
.
.
100 25 67 94
-----------
-----------
-----------
-----------
-----------
-----------
-----------
-----24
---32
-----------
-----------
---------21
49
49
44
44
44
42
40
39
35
34
-----------
-----------
-----20
---31
36 -6 25
50 -17 -11 --
------
------
------
------
------
------
------
------
------
------
------
78
71
63
53
48
------
23
-26
13
5
--
--
--
--
--
--
--
--
--
--
--
--
--
65 14
--
--
--
--
--
--
--
--
--
--
20
--
--
65
33 .
67 22
67 14
----
----
----
----
----
----
----
----
----
----
-31
--
----
----
56 -55 13
51 --
35 25 25 . 64
6
.
. 33 44
59 75 . 33 81
.
.
.
. 14
-----
-----
-----
-----
-----
-----
-----
-----
-----
-----
22
-24
--
-----
-----
-----
47
45
38
36
-------------------
48 55 -- -- -- -- -- -54 -- 33 -- -- -- -- -48 -- -- 42 -- -- -- --- -- 40 33 -- -- -- --- -- -- -- 36 43 -- --- -- -- -- -- 47 36 27
-- -- -- -- -- 39 35 --- -- -- -- -- 61 -- 38
-- -- -- -- -- 54 -- 34
-- -- -- 32 -- 44 -- 35
-- -- -- -- -- 52 13 --- -- -- -- -- -- 33 --- -- -- -- -- -- -- --- -- -- -- -- -- -- --- -- -- -- -- -- 19 --- -- -- -- -- -- -- --- -- -- -- -- -- -- --- -- -- -- -- 43 -- --
-------------------
------------63
-23
--36
-- --- --- --- --- --- --- --- --- --- -36 -55 --- -33 49
36 --- 54
-- 50
-- --
-------------------
----------16
-35
25
35
43
37
43
Diagnostic species of single groups
Group 11: Chenopodium opulifolium -Haplosciadium
Amaranthus graecizans
.
.
.
Brachiaria deflexa
.
.
.
Commelina africana
.
.
.
Dactyloctenium geminatum
.
.
.
Ambrosia maritima
.
.
.
abyssinicum
.
.
.
. 14 .
.
3
.
.
.
.
.
.
.
Group 12: Portulaca oleracea -Amaranthus hybridus community
Conyza bonariensis
.
.
.
.
.
.
Veronica anagallis-aquatica
.
.
.
.
.
.
Hibiscus diversifolius
.
.
.
.
.
.
Crambe hispanica
.
.
.
.
.
.
Chloris pilosa
.
.
.
.
.
.
Chenopodium ambrosioides
. 10 .
.
.
.
Dalbergia melanoxylon
.
.
. 10 .
.
Acacia seyal
.
.
.
.
.
.
Cyperus dives
. 20 . 20 7 29
Portulaca oleracea
.
. 20 10 17 43
Group 13: Paspalum vaginatum -Chenopodium
Chenopodium murale
.
.
Paspalum vaginatum
43 .
Stellaria media
.
.
Polygonum aviculare
.
.
Rumex usambarensis
.
.
community
.
.
. 11 47
.
.
.
. 35
.
5
.
. 41
.
. 13 11 65
.
.
.
. 12
.
.
.
.
.
.
.
.
.
.
murale community
.
.
. 14 .
.
.
. 14 .
.
.
.
. 25
.
.
. 14 .
.
.
.
.
.
Group 14: Gnaphalium unionis var. rubriflorum community
Alectra sessiliflora
.
.
.
.
.
.
.
5
5
5
32
.
.
9
95
.
.
.
.
.
.
.
.
.
3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
11 .
22 .
22 76
.
9
.
.
.
.
.
13 .
10 11
.
.
.
.
.
.
.
.
.
6
.
6
.
.
.
.
25
.
.
.
.
.
.
.
.
100 .
100 .
100 .
50 .
50 33
.
.
.
.
.
12
Gnaphalium unionis var. rubriflorum
.
10
.
.
.
29
.
.
5
.
41 25
.
Pimpinella anisum
Plantago palmata
Trifolium campestre
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
18
.
.
3
.
.
.
. 41 .
22 . 25
.
.
.
Group 15: Eleusine indica -Haplosciadium
Setaria verticillata
.
Rumex nepalensis
.
Eleusine indica
.
Scabiosa columbaria
.
abyssinicum community
.
.
.
.
.
.
5
.
.
.
.
.
.
.
.
.
.
.
. 10 . 25 27 13
.
.
.
.
.
.
.
.
Common diagnostic species of two or
Cyperus exaltatus
.
Heliotropium steudneri
.
Ipomoea aquatica
.
Pluchea dioscoridis
.
Eragrostis tenuifolia
.
Galinsoga quadriradiata
.
Crotalaria lebrunii
.
Stellaria sennii
.
Oxalis corniculata
.
Ageratum conyzoides
.
Tagetes minuta
.
Schk uhria pinnata
.
Chenopodium opulifolium
.
Datura stramonium
.
Amaranthus hybridus
.
Malva parviflora
.
Chenopodium album
14
Haplosciadium abyssinicum
.
more groups
90 100 30
60 . 40
50 .
.
10 . 40
.
. 10
.
.
.
.
.
.
.
.
.
.
.
.
. 40 .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10 .
.
Group 10: Hygrophila auriculata community
This group comprised nine plots and three
diagnostic species: Acacia elatior, Cirsium vulgare, and Hygrophila auriculata.
This group occurs in fallows and grazing
lands during the dry season as well as in
the dry fringes of floodplains. Besides the
diagnostic species, Cynodon dactylon and
Cyperus rotundus are typically associated
and frequently dominant species in this
community.
Biodiversity & Ecology 4
2012
7
10
45
34
.
.
.
.
.
79
.
.
.
.
28
.
.
.
29
.
.
.
43
57
.
.
14
43
29
29
.
29
14
.
.
.
.
.
.
.
50
100
50
75
100
100
100
.
.
.
50
.
.
100
.
.
.
.
5
82
45
14
27
23
41
50
.
23
68
5
9
5
.
.
.
.
.
67 19
100 8
5 11 12 .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5
.
.
.
.
.
68 .
. 50 .
.
10 .
. 25 .
.
50 .
.
.
.
.
70 .
. 25 25 33
85 .
. 25 .
.
23 .
. 75 .
.
. 11 . 75 .
.
.
. 82 25 .
.
3
. 18 75 100 33
25 . 76 100 . 67
3
. 35 . 100 67
3
.
6 25 75 .
30 . 88 . 25 67
3
.
.
.
.
11
.
.
.
8
44
.
50
61
97
83
58
100
Group 11: Chenopodium opulifoliumHaplosciadium abyssinicum community
This vegetation unit was encountered in
17 plots with seven diagnostic species:
Amaranthus graecizans, Ambrosia maritima, Brachiaria deflexa, Chenopodium
opulifolium, Commelina africana, Dactyloctenium geminatum, and Haplosciadium
abyssinicum. All plots were located in dry
or completely drained fields that were
either intensively cropped or had been
abandoned to fallow after extended periods of cropping. Accordingly, diagnostic
--
species were typical upland weeds associated with intense vegetable cultivation.
Group
12:
Portulaca
oleraceaAmaranthus hybridus community
Group 12 was encountered in four plots
with 14 diagnostic species: Acacia seyal,
Amaranthus hybridus, Chenopodium ambrosioides, Chloris pilosa, Conyza bonariensis, Crambe hispanica, Cyperus
dives, Dalbergia melanoxylon, Datura
stramonium, Hibiscus diversifolius, Portulaca oleracea, Schkuhria pinnata,
71
Tagetes minuta, and Veronica anagallisaquatica. The group includes exclusively
irrigated croplands from the Rumuruti
floodplain fringe, mostly cultivated with
maize. Accordingly, many of the diagnostic species are annual upland weeds,
originating from North and South America (A. hybridus, C. bonariensis, D. stramonium, S. pinnata, and T. minuta). Also
some weed species of Eurasian origin
were associated with this unit (C. abrosioides, C. hispanica, P. oleracea, and V.
anagallis-aquatica). Like Portulaca oleracea most species are indicating dry
conditions or the savanna ecological zone.
Group 13: Paspalum vaginatumChenopodium murale community
This group was encountered in four plots
with eight diagnostic species: Chenopodium album, C. murale, Datura stramonium, Malva parviflora, Paspalum
vaginatum, Polygonum aviculare, Rumex
usambarensis, and Stellaria media. Group
13 is representing a dry grassland vegetation and is only encountered in dry grazing lands or fallow plots of the completely
drained inland valleys in Lukozi. Despite
the prevalence of dicotyledonous weed
species, the dominance of Paspalum
vaginatum gives to this community the
physiognomy of grassland vegetation.
Group 14: Gnaphalium unionis var.
rubriflorum community
Group 14 occurred in only three plots and
contained five diagnostic species, namely
Alectra sessiliflora, Gnaphalium unionis
var. rubriflorum, Pimpinella anisum,
Plantago
palmata,
and
Trifolium
campestre. This small group (2 fallows
and 1 cropland) is restricted to the Lukozi
wetland. It represents weed communities
in less intensive cultivation, especially in
fallows. The naming taxon G. unionis var.
rubriflorum is also mentioned for montane forests in the north of Kenya (Bytebier & Bussmann 2000).
Group
15:
Eleusine
indicaHaplosciadium abyssinicum community
Group 15 is the second largest cluster
with 36 plots and nine diagnostic species:
Amaranthus hybridus, Chenopodium album, C. opulifolium, Eleusine indica,
Haplosciadium abyssinicum, Malva parviflora, Rumex nepalensis, Scabiosa
columbaria, and Setaria verticillata. This
group is dominated by croplands but includes also some fallows, all located in
the inland valley Lukozi. All diagnostic
species are typical weeds, many of them
originating from Eurasia (C. album, C.
opulifolium, S. columbaria, and S. verticillata). Also some species common from
grasslands are selected as diagnostic species of this unit, for example E. indica and
S. verticillata (Phillips 1930). The dominant species here are A. hybridus and Galinsoga parviflora, but the second species
is not selected as a diagnostic one.
Ordination analysis and geographical distribution of plant
communities
According to the ordination analysis using
all the species, only the weed communities tend to be aggregated, most of them
appearing in the upper right side of the
nMDS configuration, with the exception
of the relevé groups 5, 8 and 9 (Fig. 2a).
Group 1 (Typha capensis community)
appears isolated from the rest of the
communities due to its strong floristic
dissimilarity and to its low species richness. The distribution pattern in this ordination follows only partially the sorting of
the groups in the summary table. If we
consider only diagnostic species (Fig. 2b),
we cannot detect big changes in the
nMDS configuration, although in this case
we are only considering 37% of all detected species. This fact suggest that the
companion species (non diagnostic) are
randomly distributed in the all the communities and do not have a big impact in
the floristic similarities. The stress value
of the nMDS is less than 0.13 in both ordinations, which is considered as relatively low (Backhaus et al. 2006).
The classification of the relevés as well
as the nMDS configuration was strongly
influenced by the geographic location,
which is probably the most important factor determining the species composition
of the communities. We therefore calculated the centroids of the wetlands using
the nMDS coordinates of the relevé
groups and the frequency of each wetland
(Fig. 2b). The distribution of the communities along the wetlands was irregular,
many communities were resticted to single wetlands, and none of the communities occurred in all wetlands (Table 4).
The variability of plant communities was
highest in the Rumuruti floodplain and
lowest in the inland valleys of Tegu. The
richness on plant communities seems to
be influenced by both the wetland size
and the diversity in land uses and probably land use intensities.
Table 4: Plant communities described in this work and their presence in the sampled wetlands (Ru: Rumuruti, Te: Tegu, Ma: Malinda, Lu: Lukozi).
Group
Community of...
Ru
Te
Ma
Lu
.
+
+
.
.
.
+
+
.
+
.
.
+
+
+
.
.
.
.
.
+
.
+
.
.
.
.
+
.
.
+
+
+
+
.
.
.
+
.
+
.
.
.
.
+
.
.
.
.
.
.
.
.
.
.
.
+
.
+
+
Wetland vegetation
1
2
3
4
6
10
13
5
7
8
9
11
12
14
15
72
Typha capensis
Cyperus papyrus-Cyperus exaltatus
Cyperus exaltatus-Callitriche oreophila
Grassland vegetation
Pennisetum mezianum-Sporobolus pyramidalis
Trifolium semipilosum-Sida schimperiana
Hygrophila auriculata
Paspalum vaginatum-Chenopodium murale
Weed communities
Pentodon pentandrus-Leersia hexandra
Sacciolepis africana-Oxygonum sinuatum
Sonchus oleraceus-Galinsoga quadriradiata
Hydrocotyle sibthorpioides
Chenopodium opulifolium-Haplosciadium abyssinicum
Portulaca oleracea-Amaranthus hybridus
Gnaphalium unionis var. rubriflorum
Eleusine indica-Haplosciadium abyssinicum
Biodiversity & Ecology 4
2012
Plant communities and land
uses
Considering the physiognomy and the
dominant land uses, we recognize three
sets of plant communities: wetland seminatural vegetation, grassland vegetation,
and weed communities (Table 4). The
three species giving the names to the
communities of semi-natural vegetation
(Typha capensis, Cyperus papyrus and C.
exaltatus) are the most important diagnostic species and potential characteristic
species of a syntaxa of high rank (alliance, order or class). Although those
communities are more or less linked
through the mentioned diagnostic species,
they appear much dispersed in the nMDS
configuration, because those communities
are sharing only few diagnostic species.
In the case of the Cyperus papyrusCyperus exaltatus community, we can
recognize visually in the vegetation matrix (data not shown) two sub-groups,
which were not discriminated by TWINSPAN, namely plots dominated by C. papyrus and missing all other diagnostic
species with the exception of C. exaltatus,
and plots where C. papyrus appears with
very low abundance, accompanied by Typha capensis In a further intent to define
associations, we should split this group
into the two mentioned sub-groups.
According to our field observations, the
stands dominated by papyrus alone represent a more natural situation, while stands
of C. papyrus with T. capensis are more
disturbed and show signs of eutrophication. Since all studied wetlands are temporarily flooded, semi-natural communities of floating or submersed plants, like
those studied by Harper et al. (1995) are
not detected in this survey.
In a schematic way Kamiri (2010) proposed sequences of disturbance or use
intensity for each locality mentioning the
respective plant communities. The classification of the plots in this work does not
give a complete matching with those hypothetical communities. In attention to
this, we modify the scheme of disturbance
grades under near-permanent flooding
conditions, starting with low disturbed
communities like those dominated by Typha capensis or Cyperus papyrus (sub
group included in the group 2), the next
level of disturbance in both cases is related to the Cyperus papyrus-Cyperus exaltatus community, characterized by a
high constancy of both naming species
plus T. capensis, the third level is a vegetation dominated by C. exaltatus
(group 3), and advanced levels are indi-
Biodiversity & Ecology 4
2012
Fig. 2: Configuration of the nMDS showing floristic similarities between relevé
groups (numbered as in Table 3). The nMDS analysis was restricted to two dimensions, using Bray-Curtis index as distance metric. The same analysis was applied to
the vegetation matrix including all species (a) and only diagnostic species (b). The
second ordination diagram was rotated by using the Procrustes method. Centroids
of the localities were calculated for the second nMDS configuration, weighting the
coordinates by the constancy of relevés from those wetlands. Stress factors: a =
12.78%, b = 10.86%.
cated by Leersia hexandra (e.g. Pentodon
pentandrus-Leersia hexandra community). No syntaxonomical proposals are
made for the assignment of papyrus
swamps to a syntaxon of high level (class
or order). According to the physiognomy
but also the floristic composition those
communities may be related to the class
Phragmito-Magno-Caricetea Klika in
Klika & V. Novák 1941 (Mucina 1997).
The unused plots of the wetland in
Tegu were not discriminated in the
TWINSPAN classification and included
in the Hydrocotyle sibthorpioides community. Probably the high intensity of
cultivation and the small size of this wetland resulted in a high disturbance of the
few unused plots. There were also frequent introgressions of weed communities
from directly adjacent plots into the wetland plant communities.
Neither fallow plots nor grazing fields
appear to be forming discrete vegetation
units, suggesting a continuous species
73
turnover between both land uses. It was
difficult to differentiate between grazing
lands and fallows, particularly as most
fallow lands were also used for occasional
grazing. We recognized four relevé
groups as typical grassland vegetation
(Table 4) as they were generally close in
the nMDS configuration (Fig. 2), with the
exception of the Paspalum vaginatumChenopodium murale community. No
published floristic studies on weed vegetation in Kenya or Tanzania exist to date.
Thus further studies on the use of weed
species as indicators of soil conditions
and cropping intensity are required. While
the present set of weed communities appears well agglomerated in the nMDS
configuration, many of the diagnostic
species are spread across other communities. Since the Pentodon pentandrusLeersia hexandra community is the only
one growing under longer periods of
flooding in lowland rice paddies, its floristic differences with other weed communities are due to the environmental
conditions contrasting with the otherwise
drained fields. The Pentodon pentandrusLeersia hexandra community shares diagnostic species with groups 2 (Cyperus
papyrus-Cyperus exaltatus community)
and 4 (Pennizetum mezianum-Sporobolus
pyramidalis community), all occurring
within the lowland floodplain. In this last
set, many diagnostic species are mentioned as characteristic for weed vegetation in Europe (Pott 1995, Ellenberg
1996, Oberdorfer 2001), and the most important chorological groups are Eurasia
and South and North America. Those
plants are mixed in the communities with
native species, which also behave as typical weeds. A more detailed survey should
elucidate the relation of these vegetation
units with the European weed communities, especially those belonging to the
class Stellarietea mediae Tx. et al. ex von
Rochow 1951 (Mucina 1997, Rodwell et
al. 2002).
Calculating the mean Bray-Curtis values of pairwise comparisons, we find a
high heterogeneity within relevé groups,
with all values ranging bellow 50% of
similarity (data not shown). This heterogeneity can be justified through the high
variability of disturbance through cropping or other land uses, which is influencing partially the species composition but
especially the relative distribution of
abundance values. Additional factors affecting the heterogeneity of the vegetation
are the seasonality of the studied phytocoenoses (due to flooding dynamics and
cropping) and the heterogeneity of condi-
74
tions within plots, especially in croplands,
where wet furrows are alternated with dry
mounds. Both factors (temporal and spatial variability) explain the co-occurrence
of dryness and humidity indicators in the
relevés. In attention to this and the fact
that a reference syntaxonomical scheme is
missing, we retain the definition of the
relevé groups shown here as plant communities (rankless syntaxa). As typical of
wetland vegetation, semi-natural stands
are relatively poor in plant species
(Chambers et al. 2008). Agricultural use
of the fields together with disturbances in
the soil water balance, especially by
mound cultivation (typical of Tegu), terrace construction (in Lukozi) or drainage
(in all wetlands) result in an increase of
the species richness due to introgression
of plants from the neighboring uplands as
well as through the introduction of weeds.
Conclusions and outlook
According to the results presented here,
we can classify the vegetation of small
wetlands in East Africa floristically into
15 plant communities (syntaxa without
specific rank). While undisturbed areas
are associated with only few species,
more disturbed stands have higher species
richness. A strong floristic relationship is
exhibited by weed communities, while
semi-natural vegetation is dispersed in the
nMDS configuration, suggesting a high
heterogeneity between communities.
However, neither ordination analysis nor
selected diagnostic species show clear
enough linkages to vegetation units to be
used as reference for a hierarchical syntaxonomic classification. More vegetation
surveys as well as the extension of
SWEA-Dataveg as a regional database
will contribute to the establishment of a
baseline of a syntaxonomic classification
of wetlands in Africa.
Acknowledgements
The SWEA-project is supported by the
Volkswagen
Foundation,
Hannover,
Germany.
References
Alvarez, M., Möseler, B.M., Josko, M.,
Becker, M., Langensiepen, M., Menz,
G., Böhme, B., Oyieke, H., Handa, C.,
Kamiri, H., Misana, S., Mwita, E.,
Neema, M., Sakané, N. (2012): SWEADataveg – vegetation of small wetlands
in East Africa. – In: Dengler, J., Oldeland, J., Jansen, F., Chytrý, M., Ewald,
J., Finckh, M., Glöckler, F., LopezGonzalez, G., Peet, R.K., Schaminée,
J.H.J. (2012) [Eds.]: Vegetation databases for the 21st century. – Biodiversity
& Ecology 4: 294–295. Hamburg: Biocentre Klein Flottbek and Botanical Garden. CrossRef
ARI Mlingano (2006): Soil map of Tanzania. – Tanga, Tanzania: Mlingano Agricultural Research Institute (ARI Mlingano).
Aselmann, I., Crutzen, P.J. (1989): Global
distribution of natural freshwater wetlands and rice paddies, their net primary
productivity, seasonality and possible
methane emissions. – Journal of Atmospheric Chemistry 8: 307–358. CrossRef
Backhaus, K., Erichson, B., Plinke, W.,
Weiber, R. (2006): Multivariate Analysemethoden. – Berlin, Germany:
Springer.
Barkham, J.P., Rainy, M.E. (1976): The
vegetation of the Samburu-Isiolo Game
Reserve. – East African Wildlife Journal
14: 297–329.
Bray, J.R., Curtis, J.T. (1957): An ordination of the upland forest communities of
southern Wisconsin. – Ecological Monographs 27: 325–349. CrossRef
Bronner, G. (1990): Vegetation and land
use in the Mathews Range area, Samburu-District, Kenya. – Dissertationes
Botanicae 160: 182 pp. Berlin: Cramer.
Bruelheide, H. (2000): A new measure of
fidelity and its application to defining
species groups. – Journal of Vegetation
Science 11: 167–178. CrossRef
Bytebier, B., Bussmann, R.W. (2000): The
vegetation of Mount Nyiru (Samburu
District, Kenya): a checklist and syntaxonomical survey. – Journal of East African Natural History 89: 45–71. CrossRef
Chambers, P.A., Lacoul, P., Murphy, K.J.
& Thomaz, S.M. (2008): Global diversity
of aquatic macrophytes in freshwater. –
Hydrobiologia 595: 9–26. CrossRef
Chapman, L.J., Balirwa, J., Bugenyi,
F.W.B., Chapman, C., Crisman, T.L.
(2001): Wetlands of East Africa: Biodiversity, exploitation and policy perspectives. – In: Gopal, B. [Ed.]: Biodiversity in
wetlands: Assessment function and conservation: 101–132. Leiden, The Netherlands: Blackhuys Publishers.
Chytrý, M., Tichý, L., Holt, J., Botta-Dukát,
Z. (2002): Determination of diagnostic
species with statistical fidelity measures.
– Journal of Vegetation Science 13: 79–
90.
CJB, SANBI [Conservatoire et Jardin
botaniques de la Ville de Genève, South
African National Biodiversity Institute]
[Eds.] (2010): African Plants Database
(version 3.3.4). – URL: http://www.villege.ch/musinfo/bd/cjb/africa [accessed on
2010-10-15].
Cornelius, R., Schultka, W. (1997): Vegetation structure of a heavily grazed
Biodiversity & Ecology 4
2012
range in northern Kenya: ground vegetation. – Journal of Arid Environments 36:
459–474. CrossRef
Dengler, J., Chytrý, M., Ewald, J. (2008):
Phytosociology. – In: Jørgensen, S.E.,
Fath, B.D. [Eds.] Encyclopedia of ecology: 2767–2779. Oxford, UK: Elsevier.
Dengler, J., Jansen, F., Glöckler, F., Peet,
R.K., De Cáceres, M., Chytrý, M.,
Ewald, J., Oldeland, J., Finckh, M., Lopez-Gonzalez, G., Mucina, L., Rodwell,
J.S., Schaminée, J.H.J., Spencer, N.
(2011): The Global Index of VegetationPlot Databases (GIVD): a new resource
for vegetation science. – Journal of
Vegetation Science 22: 582–597.
CrossRef
Denny, P., de Ruyter van Steveninck, E.
(2001): Integrated wetland and water resources management capacity building,
with special reference to Africa. – In:
Bergkamp, G. [Ed.] Integrated wetlands
and water resources management: 29–
36. Gland: IUCN.
Ellenberg, H. (1996): Vegetation Mitteleuropas mit den Alpen. – Stuttgart: Eugen
Ulmer.
Faith, D.P., Minchin, P.R., Belbin, L.
(1987): Compositional dissimilarity as a
robust measure of ecological distance. –
Vegetatio 69: 57–68. CrossRef
Gaudet, J.J. (1977): Natural drawdown on
Lake Naivasha, Kenya, and the formation of papyrus swamps. – Aquatic Botany 3: 1–47. CrossRef
Haines, R.W., Lye, K.A. (1983): The
sedges and rushes of East Africa. – Nairobi, Kenya: East African Natural History
Society.
Harper, D.M., Adams, C., Mavuti, K.M.
(1995): The aquatic plant communities
of the Lake Naivasha wetland, Kenya:
pattern, dynamics and conservation. –
Wetlands Ecology and Management 3:
111–123.
Jarvis, A., Reuter, H.I., Nelson, A.,
Guevara, E. (2008): Hole-filled SRTM for
the globe Version 4. CGIAR-CSI SRTM
90m
Database.
–
URL:
http://srtm.csi.cgiar. org.
Junk, W.J. (2002): Long-term environmental trends and the future of tropical
wetlands. – Environmental Conservation
29: 414–435.
Kairu, J.K. (2001): Wetland use and impact
on Lake Victoria, Kenya region. – Lakes
and Reservoirs 6: 117–125. CrossRef
Kalinga, G.M., Shayo, E. (1998): Agriculture in the Tanzania wetlands. – URL:
http://www.fao.org/docrep/003/x6611e/x
6611e02d.htm [accessed on 2010-0711].
Kamiri, H.W. (2010): Effects of land use
dynamics on attributes of wetland soils
in Africa. –Bonner Agrikulturchemische
Reihe 41. Bonn: INRES, University of
Bonn.
Kassenga, G.R. (1997): A descriptive assessment of the wetlands of the Lake
Victoria basin in Tanzania. – Resources,
Biodiversity & Ecology 4
2012
Conservation and Recycling 20: 127–
141. CrossRef
Kiai, S.P.M., Mailu, G.M. (1998): Wetland
classification for agricultural development in Eastern and Southern Africa. –
URL: http://www.fao.org/DOCREP/003/
x6611E/x6611e02a.htm [accessed on
2010-07-11].
Lepš, J., Šmilauer, P. (2003): Multivariate
analysis of ecological data using
CANOCO. – Cambridge, UK: Cambridge
University Press.
Leyer, I., Wesche, K. (2007): Multivariate
Statistik in der Ökologie. – Berlin: Springer.
Muasya, J.M., Young, T.P., Okebiro, D.N.
(1994): Vegetation map and plant checklist of Al Ari Nyiro ranch and the Mukutan Gorge, Laikipia, Kenya. – Journal of
East African Natural History 83: 143–
197. CrossRef
Mucina, L. (1997): Conspectus of classes
of European vegetation. – Folia
Geobotanica et Phytotaxonomica 32:
117-172. CrossRef
New, M., Lister, D., Hulme, M., Makin, I.
(2002): A high-resolution data set of surface climate over global land areas. –
Climate Research 21: 1–25. CrossRef
Oberdorfer, E. (2001): Pflanzensoziologische Exkursionsflora für Deutschland
und angrenzende Gebiete. – Stuttgart:
Eugen Ulmer.
Oksanen, J., Kindt, R., Legendre, P.,
O'Hara, B., Simpson, G.L., Solymos, P.,
Henry, M., Stevens, H., Wagner, H.
(2009): Vegan: community ecology
package. – URL: http://www.cran.rproject.org/package=vegan.
Olden, J.D., Jackson, D.A. (2001): Spatial
isolation and fish communities in drainage lakes. – Oecologia 127: 572–585.
CrossRef
Owino, A.O., Ryan, P.G. (2007): Recent
papyrus swamp habitat loss and conservation implications in western Kenya. –
Wetlands Ecology and Management 15:
1–12. CrossRef
Peres-Neto P.R., Jackson, D.A. (2001):
How well do multivariate data sets
match? The advantages of a procrustean superimposition approach over the
mantel test. – Oecologia 129: 169–178.
CrossRef
Phillips, J. (1930): Some important vegetation communities in the central province
of Tanganyika territory (formerly German
East Africa) - a preliminary account. –
Journal of Ecology 18: 193–234. CrossRef
Polhill, R.M. (1949–2010) [Ed.]: Flora of
tropical East Africa. – London: Royal Botanic Gardens, Kew.
Pott, R. (1995): Die Pflanzengesellschaften
Deutschlands. – Stuttgart: Eugen Ulmer.
R Development Core Team (2009): R: a
language and environment for statistical
computing. R Foundation for Statistical
Computing, Vienna, Austria. – URL:
http://www.cran.r-project.org.
Rodwell, J.S., Schaminée, J.H.J., Mucina,
L., Pignatti, S., Dring, J., Moss, D.
(2002): The diversity of European vegetation - an overview of phytosociological
alliances and their relationships to
EUNIS habitats. – Wageningen, The
Netherlands: National Reference Centre
for Agriculture, Nature and Fisheries.
Roleček, J., Tichý, L., Zelený, D., Chytrý,
M. (2009): Modified TWINSPAN classification in which the hierarchy respects
cluster heterogeneity. – Journal of Vegetation Science 20: 596–602.
Sakané, N., Alvarez, M., Becker, M.,
Böhme, B., Handa, C., Kamiri, H.W.,
Langensiepen, M., Menz, G., Misana,
S., Mogha, N.G., Möseler, B.M., Mwita,
E.J., Oyieke, H.A., van Wijk, M.T.
(2011): Classification, characterisation,
and use of small wetlands in East Africa.
– Wetlands. 31: 1103–1116.
Schaminée, J.H.J., Hennekens, S.M.,
Chytrý, M., Rodwell, J.S. (2009): Vegetation-plot data and databases in
Europe: an overview. – Preslia 81: 173–
185.
Thenya, T. (2001): Challenges of conservation of dryland shallow waters, Ewaso
Narok swamp, Laikipia District, Kenya. –
Hydrobiologia 458: 107–119. CrossRef
Tichý, L. (2002): JUICE, software for vegetation classification. – Journal of Vegetation Science 13: 451–453.
Tichý, L., Chytrý, M. (2006): Statistical determination of diagnostic species for site
groups of unequal size. – Journal of
Vegetation Science 17: 809–818.
van der Heyden, C.J., New, M.G. (2003):
The role of a dambo in the hydrology of
a catchment and the river network
downstream. – Hydrology and Earth
System Sciences 7: 339–357.
van der Valk, A.G. (1981): Succession in
wetlands: a Gleasonian approach. –
Ecology 62: 688–696. CrossRef
Whittaker, R.H. (1972): Evolution and
measurement of species diversity. –
Taxon 21: 213–251. CrossRef
Wikum, D.A., Shanholtzer, G.F. (1978):
Application of the Braun-Blanquet coverabundance scale for vegetation analysis
in land development studies. – Environmental Management 2: 323–329.
CrossRef
Willner, W. (2006): The association concept revisited. – Phytocoenologia 36:
67–76. CrossRef
Willner, W., Tichý, L., Chytrý, M. (2009):
Effects of different fidelity measures and
contexts on the determination of diagnostic species. – Journal of Vegetation
Science 20: 130–137. CrossRef
75
Miguel Alvarez* (malvarez@unibonn.de), Matthias Josko (josko@unibonn.de) & Bodo M. Möseler
(moeseler@uni-bonn.de)
Vegetation Ecology, INRES, University of
Bonn
Nußallee 9
53115 Bonn, GERMANY
Mathias Becker (mathias.becker@unibonn.de)
Plant Nutrition, INRES, University of
Bonn
Karlrobert-Kreiten-Str. 13
53115 Bonn, GERMANY
Matthias Langensiepen (mlang@unibonn.de)
Crop Science, INRES, University of Bonn
Katzenburgweg 5
53115 Bonn, GERMANY
76
Gunter Menz (g.menz@geographie.unibonn.de)
Geographical Institute, University of
Bonn
Meckenheimer Allee 166
53115 Bonn, GERMANY
Beate Böhme (beate.boehme@tbt.tufreiberg.de)
Soil and Water Conservation Unit, TU
Bergakademie Freiberg
Agricolastr. 22
09599 Freiberg, GERMANY
Salome Misana (smisana@ud.co.tz),
Neema G. Mogha (moghang@yahoo.com)
& Emiliana J. Mwita
(emmyrh@yahoo.com)
Geography, DUCE, University of Dar es
Salaam
P.O. Box 35049
Dar es Salaam,TANZANIA
Collins Handa (handacollins@gmail.com)
& Helida A. Oyieke
(cbd@museums.or.ke)
National Museums of Kenya, Museums
Hill Road
P.O. Box 40658-00100
Nairobi, KENYA
Hellen W. Kamiri
(wangechikamiri@yahoo.com)
Karatina University College
Moi University
P.O. Box 1957-10101
Karatina, KENYA
Nomé Sakané (nsakana@gmail.com)
Plant Production Systems, University of
Wageningen
P.O. Box 430
6700 AK Wageningen, NETHERLANDS
*Corresponding author
Biodiversity & Ecology 4
2012