Invasion of Chromolaena odorata in the Lowveld Region of Swaziland and its Effect on Herbaceous Layer Productivity
A study was conducted to investigate
pattern of Chromolaena odorata invasion and its effect on grass
layer in three land use systems (communal, government ranch and game reserve)
and two soil classes (lithosol and raw mineral soil). Six sites (2 ha
each) were selected, one on each of the two common soil classes for each
land use system. The communal land had significantly lowest (p<0.05)
density (465 SE ha-1) of C. odorata compared to the
game reserve and there was no significant difference in total density
in the two soil classes (mean = 691 SE ha-1). The greatest
proportion occurred in height class of >1-1.5 m and the lowest in >0-0.5
m. Shrub (>0.5-3 m) and seedling (>0-0.5 m) densities were significantly
(p<0.05) highest in the game reserve and commercial ranch, respectively.
Total dry matter (DM) grass production within and outside invaded areas
showed significant differences (p<0.05) in many sites. More pronounced
trends of DM production were observed for the most dominant palatable
species which included Urochloa mozambicensis, Panicum deustum
and P. maximum. It is concluded that the invasion of C.
odorata was rated to be moderate or high while past land management
may be the major factor in the difference between land use systems. The
effect of invasion of C. odorata on the grass layer was also severe
to constitute a threat to livestock and game industries. Adaptive control
strategies are therefore recommended while further work will be required
to identify the causes of difference between land uses, to determine soil
characteristics and overall productivity potential as affected by C.
odorata invasion as well as interaction effects between biotic and
Southern Africa savanna rangelands support large, diverse
and dynamic ecosystems which sustain both wildlife and agricultural enterprises.
In several cases, however, alien shrubs have been noted to invade these
rangelands. Such invasion has been known as a recent ecological phenomena
characterized by a plant density of invasive shrubs beyond the level the
ecosystem can carry under a climax equilibrium condition. Chromolaena
odorata is a perennial scrambling shrub of neotropical origin forming
a dense tangled bushes of up to 3 m (Strathie-Korrubel, 2000) and produces
copious quantities of efficiently dispersed seeds capable of establishing
even in undisturbed habitats. It forms a dense, permanent thickets over
large areas, thereby suppressing other vegetation. Solomon et al.
(2006) contended that the invasion of savannas by C. odorata appears
to be episodic in nature and the change from savannas to shrub land can
happen very rapidly. The population of C. odorata in southern Africa
is commonly associated with high rainfall areas and annual rain as little
as 500 mm or below (Foxcroft and Martin, 2002).
Chromolaena odorata has been reported to be a
major threat to livestock, game farming and biodiversity conservation
with a negative impact on the productivity of the grassland ecosystem
in southern Africa. The productivity and long term economic viability
of agriculture as well as the ecological integrity of savannas can be
undermined by the invasion of such plants at the expense of the herbaceous
cover, especially grasses. In Swaziland, savannas represent a valuable
economic resource as a means of sustenance to the people. There are two
main types of land tenure and usage in this ecosystem namely; the Swazi
National Land which is mainly the communal land and the Title Deed Land
which is mainly private ranching (Sweet and Khumalo, 1994). In both tenures,
livestock production is the most valuable activity and in recent time,
an invasion of C. odorata has been reported in the lowveld savannas
of Swaziland (Solomon, 2006; Solomon et al., 2006). However, information
on the extent of infestation and the magnitude of its associated problems
in the different land uses and soil types has not been fully understood.
This study was therefore designed to investigate the pattern of C.
odorata invasion and its effect on the grass layer productivity under
three land use systems (game reserve, commercial cattle ranch and communal
grazing land) and at two soil classes (lithosol and raw mineral soil).
MATERIALS AND METHODS
This study was conducted in Simunye area in the lowveld of the Kingdom
of Swaziland which represents about 33% of the total land area of the
country. The landscape is undulating with the co-ordinates of approximate
center of the study area are 26° 05 S, 31° 58 E and
the altitude ranges between 250-500 m above sea level. The climate is
generally semi-arid and mean annual rainfall ranges between 400-600 mm
with most of the rainfall occurring between October and March. The mean
monthly temperature in January and July are 26 and 18° C, respectively
(Monadjem and David, 2005). The vegetation has been classified as lowveld
savanna (Acocks, 1988) with broad leaved woodland predominating in the
West, microphyllous (acacia) savanna in the East and the riverine forest
along the rivers and major drainage lines (Hess et al., 1990; Sweet
and Khumalo, 1994). Basal and sandstone-claystone are the geological features
of this area, while raw mineral soil and lithosol are the most common
soil classes. The raw mineral soil is shallower and sandier with a higher
content of gravel and sand than the lithosol (Murdroch, 1970).
Site selection, sampling procedure and data collection
Three land use systems, namely; communal grazing land, commercial
government ranch and game reserve located adjacent to each other were
identified. Shewula land represents the communal grazing land; Nkalashane
ranch and Mlawula represent the commercial government ranch and game reserve,
respectively (Fig. 1). Six sites (2 ha each) were selected,
one on each of the two common soil classes for each land use system. The
sites were identified with the help of the geographical information system
of the country (GIS) which included patches of land invaded by C. odorata
and the adjacent non-invaded area. Three 100 m transects were laid out
randomly along the invaded patch on each site, giving a total of 18 transects.
Along each transect 3x100 m2 quadrats were evenly spaced to
record the density of C. odorata. All the rooted Chromolaena
species were counted in each 100 m2 quadrat and recorded into
one of the five height classes of >0-0.5; >0.5-1; >1-1.5; >1.5-2
and >2-3 m. An arbitrary decision was made to count a stem as a separate
individual if it is > 50 cm from the nearest stem. The plant data were
then standardized to shrub equivalent (1 SE = 1 shrub, 1.5 m high) for
determining the total density per unit area of land. Shrub equivalent
height was assumed to be the average height for the shrub. This method
was chosen because the distribution of C. odorata was impossible
to detect using remote sensing techniques (Abaye, 2003).
||Map of Swaziland (a) and location of the study sites
in the lowveld of Swaziland (b) 1 and 2 Game reserve, 3 and 4: Communal
grazing land, 5 and 6: Commercial government cattle ranch; L: Lithosol,
r: Raw mineral soil
Dry Matter yield (DM) and botanical composition of the
grass layer were determined from 7x0.25 m2 quadrats laid randomly
within each 100 m2 quadrat invaded patches and the adjacent
non-invaded areas. The quadrat samples were cut to stubble height of 2
cm and dried in the oven at 72 °C for 48 h. All vegetation data were
collected towards the end of the rainy season (March-April, 2006).
Soil Sampling and Analysis
Topsoil samples to a depth of 200 mm were taken from each transect
at 15 random locations per transect. Each set of 15 samples was bulked,
thoroughly mixed, air dried and sieved through a 2 mm mesh screen prior
analysis. Soil texture of the experimental samples was determined by means
of standard Bouyoucos (hydrometer) method (Day, 1965). Soil pH was measured
in a 1:2.5 soil water relation extract method while the Kjeldahl method
was used to determine percentage total Nitrogen (N) (Van Reeuwijk, 1992).
Percentage Organic Carbon (OC) was analyzed using colorimetric method
(Baker, 1976). Sodium (Na) and potassium (K) were determined by emission
spectroscopy and magnesium (Mg), calcium (Ca), zinc (Zn), copper (Cu)
and iron (Fe) using atomic absorption spectroscopy (Jackson, 1970). Phosphorous
was detected by the ultra violet Spectrophotometer (Olsen and Sommers,
Since quadrat data on density and height distribution of the study
plant were not normally distributed, the non-parametric Whitney U-test
for two independent samples was used (Kent and Coker, 1992). Soil and
herbaceous data were normally distributed and analysed using the GLM of
SAS software (SAS, 1987). Herbaceous data for the invaded and non-invaded
area were compared using a simple t-test.
||Texture composition (%) of topsoil sampled from the
three land use systems and two soil classes
||Soil pH, organic carbon (OC), nitrogen (N), macro (cmolc
kg-1) and micro elements (Mean ± SE1)
of top soil sampled from the three land use systems and two soil classes
|1: Standard error, Means with the different
superscripts in the same row are statistically different, *: p<0.05;
**: p<0.01, ***: p<0.001
Results on soil texture as presented in Fig. 2 revealed
that the sand, silt and clay contents were not remarkably different in
the three land use systems and the two soil classes. There were no significant
(p>0.05) variations in the soil pH among land use systems and soil
classes. Copper and Fe were below detectable level in all the study sites.
Calcium was significantly (p<0.001) highest in the commercial ranch
but lowest in the communal grazing land (Table 1). Potassium
and Na were lower (p<0.05) in the ranch than in the game reserve and
Phosphorus was significantly lowest (p<0.05) on communal
land than in the ranch, where as the game reserve did not differ (p>0.05)
from both land use systems. Nitrogen was significantly lowest (p<0.01)
on communal land. All soil variables did not show significant differences
between lithosol and raw mineral soils in the study area as shown in Table
Density and Height Distribution of Chromolaena
Marked difference (p<0.05) in the total SE density of C.
odorata existed between the communal land and game reserve. The difference
between the two soil classes (lithosol and raw mineral soil) was not apparent
and less pronounced compared to the patterns observed among the land use
systems (Table 2). Most of the study sites had largest
proportion of height class >1-1.5 m and lowest in height range >0-0.5
||Total density of Chromolaena odorata (mean shrub
equivalent per ha as well number of individuals per ha) under the
three land use systems and the two soil classes
|SE: Standard Error, Means with the different superscripts
in the same column are statistically different (p<0.05)
|| Density (mean number of individuals per ha) of the
seedlings and shrubs of Chromolaena odorata
|Means with different superscripts in the same column
are statistically different (p<0.05)
There was a significantly (p<0.05) more shrubs (>0.5-3
m) in the game reserve than in the ranch and communal land (Table
3). There was a significantly (p<0.05) more seedlings (>0-0.5)
in the ranch than on communal land and game reserve.
Grass Layer Productivity
A total of 26 grasses were identified, of which 23 were perennials,
6 classified as highly palatable, 12 moderately palatable and the remaining
ones as less palatable. Six species each were present only under and outside
the invaded areas (Table 4).
Data on grass DM yield indicated that the ranch had significantly
(p<0.001) the highest production. With regard to soil class, DM production
was remarkably higher (p<0.01) on lithosol than raw mineral soil (Fig.
Grass DM yield within and outside invaded areas did not
show marked difference on lithosol soil of the communal sites, but differed
significantly (p = 0.05) on the raw mineral soil (Table
5). On the commercial government ranch, DM yield of grass within and
outside the invaded areas was similar on the raw mineral soil, while there
was significant (p<0.01) difference on the lithosol soil class. In
the game reserve sites DM yield approached significant variations (p =
0.06) within and outside the invaded areas on both soil classes (Table
Dominant and common grass species were also examined
for their DM production within and outside the invaded areas. The following
grass species were found to be dominant/common to the study sites: Aristida
sciurus, Heteropogon contortus, P. deustum, P. maximum,
U. mosabicensis and Melinis repens.
In the communal land under the lithosol soil, DM yield
of U. mosabicensis and P. deustum were significantly lower
(p<0.05) under the invaded area than outside the invaded area. Panicum
maxiumum was significantly lower (p<0.05) within the invaded area
on both lithosol and raw mineral soil. Except for U. mosabicensis which
showed non-significant variation under the ranch-lithosol site, these
three species showed similar patterns of variation within and outside
invaded areas on the raw mineral soil in the ranch and game reserve. Nevertheless,
H. contortus and M. repens, which were dominant on raw mineral
soil in the ranch did not show marked difference (p>0.05) within and
outside the invaded areas. Similarly, A. sciurus was dominant on
lithosol of the game reserve but did not show significant difference within
and outside the infested areas (Table 6).
|| Dry matter yield of grasses in three land use systems
and two soil classes
|| Life forms and ecological grouping of grass species
identified in the study sites
|1: LP = Less palatable; MP = Moderately palatable;
HP = Highly palatable; A = annual; P = Perennial; 2: Inc
II a = Increaser IIa; Inc IIb = Increaser IIb; Inc IIc = Increaser
IIc (Foran, 1976; Tainton et al., 1980; Vorster, 1982)
||Dry matter yield (kg ha-1 ± SE) of
most common grasses under and outside invaded areas by Chomolaena
|Means with the different superscripts in the same raw
are statistically different at p<0.05(*) and at p<0.01 (**)
Information on soil texture and pH of the raw mineral
soil has been unavailable from the previous report in the lowveld areas
(Murdoch, 1970; Sutcliffe, 1975). Results of the soil texture on lithosol
in the current study indicated that the sand and silt content were 65
and 6% disagreed with the previous report of 53 and 17%, respectively
(Sutcliffe, 1975). The findings also suggested that the soil in the two
classes fall under sandy-clay-loam texture. Soil elements analysis indicated
that most exchangeable cation levels are distinguishable and not consistent
among the three land use systems and the reasons are not clear. With the
exception of Ca, whose level was very low, soil levels of all macro elements
across the study areas were fairly above critical values for plant growth
(Mtimuni, 1982; Katyal and Randhawa, 1983; McDowell, 1985). The level
of phosphorous reported in this study was within the optimum plant requirement
(Lemma et al., 2002) for semi-arid rangelands. However, previous
analysis (Sutcliffe, 1975) indicated that the amount of phosphorous in
the study areas was far lower than the present, while OC and N were reported
to be similar. Among the soil properties total OC is a sensitive soil
quality indicator suggesting that within a narrow range of soil, it may
serve as a suitable indicator of soil fertility (Murage et al.,
2000). Furthermore, soil OC fraction offers further insight into soil
fertility changes and the sustainability of land use and management (Barrios
et al., 1996; Du Preez and Snyman, 2003). Total N analysis indicated
the significantly lowest content on communal land, but overall values
suggest that the area was remarkably deficient in N. There are evidences
that deficiency of N in rangelands can severely limit the productivity
of the grass layer (Barnes and Smart, 1991; Solomon et al., 2007).
As previously discussed, the game reserve had a significantly
greater density of C. odorata than the other land use systems.
This might be due to past management and stocking rate differences of
the area. The game reserve represents an area with a history of more severe
grazing pressure than the other land uses. Prior to establishment (before
1986), the reserve area was divided into four cattle ranches on freehold
land and intensively used with an average stocking rate of 4.4 ha/LSU,
which was well above that of the other land uses and the recommended stocking
rate of 6-8 ha/LSU (FAO, 1994; SNTC, 1996) . The later shift to the present
land use was associated with the depletion of the herbaceous resources
and the progress of bush/shrub encroachment that lowered the grazing capacity
below the economic optimum number of cattle over long term utilization.
Mlawula reserve has been sparsely populated in recent times. The current
finding agreed with Abaye (2003) that, land use system has detrimental
effect on the abundance of Chromolaena in the Kwazul-Natal province
of South Africa and further pointed out that conservation and forest areas
were most affected compared to the communal grazing land. However, Smet
and Ward (2005) had a contrary report in Kimberly thorn bushveld of South
Africa that shrub densities were the highest in the communal grazing lands,
while game reserves had the lowest. Annika (2000) reported that significant
differences in shrub densities were absent between the communal land and
private rangelands in the North-east District of Botswana.
In the current study the density of C. odorata
was rated moderate (<250-550 ha-1) on communal grazing land,
high (>550-850/ha) in the ranch and very high (>850 ha) in the game
reserve. The invasion of this alien plant was also visually observed at
various levels in a number of grazing sites outside the study areas. The
increase in shrub density of alien plants which can be termed as bush/shrub
invasion has been contentious just as other woody plants in rangeland
ecology and management. Different researchers suggested that the main
causes of shrub invasion are grazing (Solomon, 2003); fire (Van Auken,
2000), rainfall (Kraaij, 2002) and the nature of soil (Wiegand et al.,
2006). These factors influence the structure and function of the arid
and semi-arid savannas of southern Africa. Grazing and fire act partly
by influencing the availability of and competition for water and nutrients.
According to Abaye (2003), the low disturbance of grazing land due to
light stocking rate contributed to the low level of Chromolaena
abundance on communal land in KwaZulu-Natal Province of South Africa.
Joshi (2001) reported that any increase in disturbance of an ecosystem
leads to increased probability of alien plants invasion. The degradation
of natural grasslands in terms of loss of herbaceous cover due to overstocking
provides conducive environment for alien species to establish and become
invasive. Several alien invasive plants identified in Southern Africa
are colonizing at the rate of species that benefit from reduced competition
that follows habitat degradation (Abaye, 2003).
Although all the study sites were dominated by shrubs
(>0.5-3 m), the patterns of height class distribution were different.
The height class represents the age of a plant, while its pattern of distribution
explains the different stages of invasion. Therefore, control of C.
odorata invasion may require a thorough understanding of different
growth stages and life cycle of the plant. Germination of C. odorata
seed occurs from the seedbank and may be favored by episodic heavy rainfall
or fire. Survival of seedlings is influenced by moisture stress, competition
from trees/shrubs and grasses and trampling by animals. The effective
stage at which C. odorata can be controlled is at seedling establishment.
In later reproductive stage the likelihood of control becomes difficult
since the plant has already produced seeds which will remain as seed bank
in the soil for the next generation.
Three states of vegetation (Fig. 4)
leading from savannas to shrub climax (invasion climax) may illustrate
the state and transition model of shrub encroachment in the semi-arid
savannas of Africa. As illustrated in the figure, a change from savanna
state to early invasion and shrub establishment state may be reversed
with minimum control inputs. Nevertheless, further advancement in the
invasion will eventually lead to shrub/bush climax, which at this state
will be unlikely to reverse without huge investment in range rehabilitation
projects. In developing countries, control of C. odorata at the
climax stage can be further constrained by limited financial resources.
|| State-and-transition model showing invasion of Chromolaena
The commercial government ranch had higher total DM grass
production than the other land uses. Some years ago, the ranch was closed
down and grazing was excluded because of the invasion of C. odorata
in some grazing camps. In a study conducted in the high veld of Swaziland,
Tsabedze and Solomon (2006) observed that, game reserve had the highest
DM grass production. The significantly lower DM production of grass on
the raw mineral than lithosol soil was expected because the former is
characterized by outcrops with discontinuous soil cover and occasionally
buried soil sandwiched between stone mantle and consolidated rock which
might not favor abundant grass production.
In this study significant difference in the total and
individual grass DM production were noted under and outside the invaded
areas. Similarly, lower productivity of herbaceous layer under the shrubs/trees
as opposed to the nearby open grasslands was reported elsewhere (Grünow
et al., 1980; Dye and Spear, 1982; Walker and Noy-Meir, 1982).
In another instance, grassland productivity was found to be higher under
canopies than in the adjacent open grasslands (Smit and Swart, 1994; Jiménez-Labato
and Vaverde, 2006). The present study indicated that the differences in
biomass of grass species increased dramatically at higher shrub densities
with the corresponding increase in the amount of bare ground litter cover
beneath the shrub. Increased grassland productivity under the invaded
area may be associated with low-shrub density and decreased productivity
-with high shrub density. The decline of grassland productivity under
the invaded areas suggest that the invasion by such shrubs competes for
soil moisture and nutrients which presumably alters the microclimate features
of the herbaceous layer (Hobbs and Mooney, 1985)
Walter (1971) postulated savannas as a two layer soil
water system in which grasses are superior competitors for water and nutrients
on top soil with woody plants having exclusive access to a lower (subsoil)
water supply. According to Walker and Noy-Meir (1982) model, the zero
isocline for woody plants is always below that of grasses. The current
study and the finding of Wiegand et al. (2006) support neither
of these models because relatively even distribution and high densities
of C. odorata roots were observed on the top soil where most of
the herbaceous plant root existed suggesting that some woody/semi-woody
plants can be equally or more competent than grasses for water on the
It is concluded from this study that the invasion of
C. odorata ranges from moderate to high and that history of land
use systems may be the major determining factor for invasion. All the
study sites were dominated by shrubs and at advanced stage, control of
the invasion may require a huge investment in the form of range rehabilitation
projects. The effect of C. odorata on grass composition and production
was severe and could threaten livestock and game industries. In most cases,
dry matter production within the invaded areas was significantly lowest
and dominant palatable species such as U. mosabicensis, P. maximum
and P. deustum were most affected. Adaptive control strategies
are therefore recommended while further work will be required to identify
the causes of difference between land uses, to determine soil characteristics
and the overall productivity potential as affected by C. odorata invasion
as well as its interaction effects.
The authors are grateful to the UNISWA Research Board
of the University of Swaziland for funding the project. The ministry of
Agriculture and cooperation and the National Trust Commission of Swaziland
are fully acknowledged for using their ranches and game reserves for the
study. We also express our appreciation to all staff members of Mlawula
game reserve for their assistance and hospitability.
1: Abaye, K., 2003. Land use and land cover in relation to Chromolaena odorata Distribution, mapping and change detection in ST. Lucia wetland area, Soth Africa. M.Sc. Thesis, International institute For Geo-Information Science and Earth Observation, Enschede, The Netherlands.
2: Acocks, J.P.H., 1988. Veld Types of South Africa. 3rd Edn., Botanical Research Institute, Pretoria, South Africa.
3: Annika, C.D., 2000. Vegetation density and change in relation to land use, soil and rainfall: A case study from North-East District, Botswana. J. Arid Environ., 44: 19-40.
Direct Link |
4: Barnes, D.L. and M. Smart, 1991. Relations between soil factors and herbage yields of natural grassland on sandy soils in the South-Eastern Transvaal. Tydskrif Weidinsveren South Afr., 8: 92-98.
5: Barrios, E., R.J. Buresh and J.I. Sprent, 1996. Organic matter in soil particle size and density fractions from maize and legume cropping systems. Soil Biol. Biochem., 28: 85-193.
6: Day, P.R., 1965. Hydrometer Method of Particle Size Analysis. In: Soil Analysis Agronomy, Black, C.A. (Ed.). Methods of American Society of Agronomy, Madison, Wiconsin, pp: 562-563.
7: Du Preez, C.C. and H.A. Snyman, 1993. Organic matter content of a soil in a semi-arid climate with three long-standing veld conditions. Afr. J. Range Forage Sci., 19: 108-110.
CrossRef | Direct Link |
8: Dye, P. and P.T. Spear, 1982. The effect of bush clearing and rainfall variability on grass yield and composition in South-West Zimbabwe. Zimbabwe J. Agric. Res., 20: 103-118.
9: FAO (Food and Agriculture Organization), 1994. Technical cooperate program, livestock sub-sector review and range survey. Working Papers, Swaziland, Vol. II.
10: Foran, B.D., 1976. The development and testing of methods for assessing the condition of three grassveld types in Natal. M.Sc. Thesis, University of Natal, Pietermaritzburg. South Africa.
11: Foxcroft, L.C. and B.W. Martin, 2002. The distribution and current status of Chromolaena Odorata in the Kruger National Park. Scientific Services Section. Kruger National Park. South Africa.
12: Grunow, J.O., H.T. Groeneveld and S.H. Du Toit, 1980. Above-ground dry matter dynamics of the grass layer of a south African tree savanna. J. Ecol., 68: 877-889.
CrossRef | Direct Link |
13: Hess, P., H. Forster and D. Gwaitta-Magumba, 1990. National Forest Inventory of Swaziland, Results and interpretation. SGFP, Report No. 5. Ministry of Agriculture and Cooperatives, Mbabane, Swaziland.
14: Hobbs, R.J. and H.A. Mooney, 1985. Community and population dynamics of serpentine grassland annuals in relation to gopher disturbance. Oecol. Berlin, 67: 342-352.
CrossRef | Direct Link |
15: Jackson, ML., 1970. Soil Chemical Analysis. Prentica-Hall, Inc., Englewood Cliffs.
16: Jimenéz-Lobato and T. Vaverde, 2006. Population dynamics of the shrub Acacia bilimekii in a semi-desert region in central Mexico. J. Arid. Environ., 65: 29-45.
Direct Link |
17: Joshi, C., 2001. Invasive Banmara (Chromolaena odorata): Spatial detection and prediction. M.Sc. Thesis, ITC, Enschede, The Netherlands.
18: Katyal, J.C. and N.S. Randhawa, 1983. Micronutrients. FAO Fertilizer and Plant Nutrition Bulletin No. 7, Food and Agriculture Organization (FAO), Rome, Italy.
19: Kent, M. and P. Coker, 1992. Vegetation Description and Analysis: A Practical Approaches. Belhaven Press, London, UK., Pages: 363.
20: Kraaij, T., 2002. Effect of rain, nitrogen, fire and grazing on bush encroachment in Semi-arid Savanna, South Africa. M.Sc. Thesis, University of Stellenbosch, Stellenbosch.
21: Lemma Gizachew, A. Hirpa, F. Jalata and G.N. Smit, 2002. Mineral element status of soils, nutritive value of pastures and cattle blood serum in the mid-altitude of Western Ethiopia. Afr. J. Range Forage Sci., 19: 147-155.
Direct Link |
22: McDowell, L.R., 1985. Nutrition of grazing ruminants in warm climates. Academic Press, Orlando, Florida.
23: Monadjem, A. and K.G. David, 2005. Nesting distribution of vultures in relation to land use in Swaziland. Biodegrad. Conservat., 14: 2079-2093.
CrossRef | Direct Link |
24: Mtimuni, J.P., 1982. Identification of mineral deficiencies in soil, plant and animal tissues as constraint to cattle production in Malawi. CTA Report 6 IFAS University of Florida, Gainesville, Florida, USA.
25: Murage, E.W., N.K. Karanja, P.C. Smithson and P.L. Woomer, 2000. Diagnostic indicators of soil quality in productive and non-productive smallholders’ fields of Kenya's Central Highlands. Agric. Ecosys. Environ., 79: 1-8.
CrossRef | Direct Link |
26: Murdoch, G., 1970. Soils and land capability in Swaziland. Ministry of Agriculture. Mbabane, Swaziland.
27: Olsen, S.R. and L.E. Sommers, 1982. Phosphorous. Methods of Soil Analysis Page. Miller, A.L. and R.H. Keeney (Eds.). pp: 403-430.
28: SAS (Statistical Analysis System), 1987. Applied Statistics and the SAS Programming Language. 2nd Edn., Cary Institute Inc., North Carolina.
29: Smet, M. and D. Ward, 2005. A comparison of the effects of different rangeland management systems on plant species composition, diversity and vegetation structure in a semi-arid savanna. Afr. J. Range Forage Sci., 22: 59-71.
CrossRef | Direct Link |
30: Smit, G.N. and J.S. Swart, 1994. Influence of leguminous and non‐leguminous woody plants on the herbaceous layer and soil under varying competition regimes in mixed Bushveld. Afr. J. Range Forage Sci., 11: 27-33.
CrossRef | Direct Link |
31: SNTC (Swaziland national Trust Commission), 1996. Malawula nature reserve management plan. Revised Draft, Swaziland.
32: Solomon, T.B., 2003. Rangeland evaluation and perceptions of the pastoralists in the borana zone of southern Ethiopia. Ph.D. Thesis, University of the Free State, Bloemfontein.
33: Solomon, T.B., 2006. Invasion of Chromolaena odorata on the lowveld of Swaziland. A Workshop on Asandanezwe, the Menance: Its effect on Agriculture Land and the Environment. University of Swaziland, Swaziland.
34: Solomon, T.B., B.J. Dlamini and A.M. Dlamini, 2006. Invasion of Chromolaena odorata in the Lowveld of Swaziland and its effect on the herbaceous layer productivity. 41st Annual Congress of Grassland Society of Southern Africa. Bela Bela, South Africa.
35: Solomon, T., H.A. Snyman and G.N. Smit, 2007. Rangeland dynamics in southern Ethiopia: (1) Botanical composition of grasses and soil characteristics in relation to land-use and distance from water in semi-arid Borana rangelands. J. Environ. Manage., 85: 429-442.
Direct Link |
36: Strathie-Korrubel, L., 2000. The South African program on the biological control of Chromolaena odorata (Triffid weed) and its global significance. Plant Protect. News, No. 57.
37: Sutcliffe, J.P., 1975. A field guide to soils of Swaziland. Ministry of Agriculture, Mbabane, Swaziland.
38: Sweet, R.J and S. Khumalo, 1994. Range resources and grazing potentials in Swaziland. Report to the Ministry of Agriculture and Cooperatives and UNDP, Mbabane, Swaziland.
39: Tainton, N.M., P.J. Edwards and M.T. Mentis, 1980. A revised method for assessing veld condition. Proc. Grassland Soc. Southern Afr., 15: 37-42.
CrossRef | Direct Link |
40: Tsabedze, W.N. and T.B. Solomon, 2006. Vegetation patterns and nutrients in relation to grazing management systems in the highveld area of Swaziland. 41st Annual Congress of Grassland Society of Southern Africa. Bela Bela, South Africa.
41: Van Auken, O.W., 2000. Shrub invasion of semiarid grasslands. Ann. Rev. Ecol. Syst., 31: 197-215.
Direct Link |
42: Van Reeuwijk, L.P., 1992. Procedure for Soil Analysis. 3rd Edn., International Soil Reference and Information Center. Wageningen (ISRIC). Netherlands.
43: Vorster, M., 1982. The development of the ecological index method for assessing veld condition in the Karoo. Proc. Grassld Soc. Southern Afr., 17: 84-89.
CrossRef | Direct Link |
44: Walker, B.H. and I. Noy-Meir, 1982. Aspects of the Stability and the Resilience of Savanna Ecosystem. In: Ecology of Tropical Savannas, Huntley, B.J. and B.H. Walker (Eds.). Springer Verlag, Berlin, pp: 556-590.
45: Walter, H., 1971. Natural savannas. Ecology of tropical and subtropical vegetation. Oliver and Boyd, Edinburgh.
46: Wiegand, K., D. Salts and D. Ward, 2006. A patch-dynamics approach to savanna dynamics and woody plant encroachment-insights from an arid savanna. Perspect. Plant Econ. Evol. Syst., 7: 229-242.
Direct Link |