4 KwaZulu-Natal Nature Conservation Service, PO Box 13053, Cascades, Pietermaritzburg 3202, South Africa

* Corresponding author, e-mail: [email protected]

Received 23 April 2001, accepted 14 August 2001

The soil types and woody vegetation of Weenen Nature Reserve (WNR), KwaZulu-Natal, were described. Mispah, overlying shale, and Shortlands, overlying dolerite were the most common of 18 the soil forms encountered. Six vegetation types were identified by TWINSPAN which differed in slope, aspect, land-use history, species richness and tree density. Vegetation types ranged from species-rich, dense ‘Coddia rudis-Dombeya cymosa Closed Woodland’ overlying dolerite and occurring on steep north-east facing slopes that had never been cultivated, to previously cultivated, species-poor ‘Acacia sieberiana Open Woodland’ on flat topography, overlying shale. Ordination analyses further described the main associated floristic-environmental gradient from steep slopes supporting a high diversity of broadleaved species through to species-poor, flatter areas with an abundance of microphyllous plants. A second main gradient described was from uncultivated sites on steep slopes of usually dolerite-derived soils through to flat, previously cultivated land on shallow soils overlying shale. Topo-edaphic variables and previous cultivation, which were confounded, were therefore the main determinants of floristic variation in WNR, whereas fire had no influence. Half of the dominant woody species exhibited a reverse-J size structure, indicating relatively constant population change. A number of species had a preponderance of individuals in the smallest (0.5—1.5m height) size class, owing possibly to recruitment during the exceptional wet season of 1995/6, and/or the inability of individuals to escape the ‘fire-trap’. Several, mostly Acacia, species appear to have recruited intermittently. The vegetation changes of this system will have an affect on browse availability. The woody vegetation of WNR appears to be changing in response to successional processes and management influences.

Keywords: carrying capacity, classification, environmental gradient, ordination, savanna

African Journal of Range & Forage Science 18 (2+3) 2001, 93—101
© 2001 NISC Pty Ltd, www.nisc.co.za

Sampling efficiency for species composition assessments using the wheel-point method in a semi-arid savanna

e-mail: [email protected]

Received 16 March 2000, accepted 8 April 2001

A pilot survey to determine the sampling efficiency of the wheel-point method, using the nearest plant method, to assess species composition (using replicate similarity related to sampling intensity, and total sampling time) was conducted on three plot sizes (20 x 20m, 30 x 30m, 40 x 40m) at two sites in a semi-arid savanna. Total sampling time included the time to demarcate a plot. A 40 x 40m plot had the highest precision and repeatability (number of point observations required to achieve a stated asymptotic value), and was followed by a 20 x 20m plot. These relationships were consistent across sites. A 40 x 40m plot recorded the highest sampling times due to the time required to demarcate a plot. A 20 x 20m plot was the most efficient. A model to assist in making decisions regarding a suitable plot and sample size, and for planning and budgeting of vegetation monitoring programs was developed. Future surveys should include a Monte-Carlo data simulation as an adjunct to data collection.

Keywords: grass layer, plot size, point methods, replicate similarity, sample size

African Journal of Range & Forage Science 18 (2+3) 2001, 103—109
© 2001 NISC Pty Ltd, www.nisc.co.za

The influence of tree density on the grass layer of three semi-arid savanna types of southern Africa

CGF Richter¹, HA Snyman² and GN Smit²*

1 Free State Department of Agriculture, Private Bag X01, Glen 9360, South Africa

* Corresponding author, e-mail: [email protected]

Received 5 June 2001, accepted 14 August 2001

The influence of bush density on the grass component of the herbaceous layer was investigated for three savanna veld types (Molopo Thornveld, Mixed Vaalbos Thornveld, and the Eastern Grass and Bushveld) in the Northern Cape and North-West Province of South Africa. Relationships between bush density and the density, species composition and phytomass production of the grass component were determined over a four-year period. Bush encroachment had an adverse influence on the density and production of the grass component, but not on the species composition. Fluctuations in seasonal rainfall appeared to be the most important factor governing changes in the species composition of the grass component. Bush enchroachment (2 500 tree equivalents ha^-1) decreased the potential grazing capacity by as much as 331%, 149% and 58% in the Molopo Thornveld, the Mixed Vaalbos Thornveld and the Eastern Grass Bushveld, respectively, in comparison to sites with tree densities of less than 400 tree equivalents ha^-1. In the encroached condition grazing capacities of 15, 13 and 10 ha LSU^-1 are recommended for the Molopo Thornveld, the Mixed Vaalbos Thornveld and the Eastern Grass and Bushveld, respectively. The implementation of bush control measures in all problem areas is proposed in order to restore the productivity of these areas in the long term.

Keywords: bush encroachment, grass-bush interactions, above-ground phytomass, grazing capacity

African Journal of Range & Forage Science 18 (2+3) 2001, 111—116
© 2001 NISC Pty Ltd, www.nisc.co.za

Quantified range condition assessment of open Camelthorn savanna along a degradation gradient

I Zimmermann¹*, FP Graz¹, MH Berry^2, NNT Imasiku², ND Muroua², MM Sibalatani² and CS Sikopo²

* Corresponding author, e-mail: [email protected]

Received 18 March 2001, accepted 31 August 2001

Measurements were taken at 20 different sites in a relatively homogenous area of open Camelthorn savanna in eastern Namibia. Ordination of grass species composition did not provide a good degradation gradient, due to domination by different species of annual grasses at many of the sites. However a centred PCA ordination of perennial grass species indicated a degradation gradient. This gradient was used to allocate a range condition score to each site, which was then correlated with other measurements. The range condition score did not correlate well with mulch cover, basal cover, proportion of bare soil, above-ground grass biomass, proportion of moribund grass, proportion of seedlings amongst the perennial grasses, density of woody plants lower than 0.5m and canopy cover of browsable woody plants over 0.5m tall. A weak correlation (r² = 0.68, p<0.0001, n = 20) was obtained through a power trend line with woody canopy cover of only the bush thickening species. A better correlation (r² = 0.90, p<0.0001, n = 20) was obtained with mean distance from sampling point to the nearest perennial grass. This may therefore be an appropriate indicator to include in range condition score for such types of savanna.

Keywords: ordination, distance measurements, Namibian rangeland, western Kalahari

African Journal of Range & Forage Science 18 (2+3) 2001, 117—124
© 2001 NISC Pty Ltd, www.nisc.co.za

Abiotic and herbaceous vegetational characteristics of an arid rangeland in Kenya

WN Ekaya¹*, JI Kinyamario² and CN Karue¹

1 Department of Range Management, University of Nairobi, PO Box 29053, Nairobi, Kenya

2 Botany Department, University of Nairobi, PO Box 29053, Nairobi, Kenya

* Corresponding author, e-mail: [email protected]

Received 15 June 2001, accepted 13 September 2001

A two-year study was conducted with an overall objective of characterising the structure and function of an arid rangeland in Kenya. A plot measuring 100 x 100m was used for this study. Data on rainfall and temperature were recorded at the plot site whereas data on evaporation rates and relative humidity were obtained from the meteorological office near the study site. Herbaceous aboveground material was sampled at monthly intervals using a rectangular 0.25m² quadrat frame. Clipped material was separated by species and classified dead or live by physical examination. A 5cm diameter metal soil corer was used to sample belowground plant material, at monthly intervals. The sampled material was washed with running water over 2mm sieves and classified dead or live using the vital staining technique. All weights and calculations were based on organic weight.

Total aboveground standing crop ranged from 84.6g m^-2 to 295.4g m^-2, with a mean of 162.3 ± 60.6g m^-2. Mean monthly aboveground standing crop for 1992 and 1993 was 142.8 ± 53.8 and 178.5 ± 63.3g m^-2 respectively. The two values were significantly different (p<0.10). Aboveground biomass yield ranged from 17.7g m^-2 to 242.7g m^-2, with a mean of 104.3 ± 58g m^-2 and a coefficient of variation of 58%. Mean aboveground standing crop was 59 ± 24g m^-2. Monthly values ranged from 28.8g m^-2 to 120g m^-2, with a 38% coefficient of variation. The range for total belowground standing crop was from 83.3g m^-2 to 232.7g m^-2, and a mean of 155.2 ± 46g m^-2. The values had a coefficient of variation of 30%. Mean total monthly belowground plant material yield for 1992 and 1993 was 137.6 ± 41g m^-2 and 169.9 ± 46g m^-2. The coefficients of variation were 59% and 28% respectively. The mean monthly belowground biomass yield was 51.6 ± 33g m^-2 with a coefficient of variation of 64%. Mean monthly yield for belowground dead material was 103.7 ± 32g m^-2, with a coefficient of variation of 31%. There was no significant difference (p>0.01) in the mean belowground dead material yield between 1992 and 1993. In 1992, annual NPP was 439.2g m^-2, giving a net primary productivity of 1.22g m^-2day^-1. Monthly NPP ranged from 17.2g m^-2 to 90.1g m^-2. In 1993, annual NPP was 944.5g m^-2, equivalent to a net primary productivity of 2.62g m^-2 day^-1. Monthly NPP was between 27.4g m^-2 and 548.6g m^-2. Over the 1992—1993 period, NPP was 1 383.7g m^-2, equivalent to a productivity of 1.92g m^-2day^-1. Trends in monthly NPP closely followed the trend in rainfall. On the whole, herbaceous vegetation production and productivity were episodic in nature and closely linked to rainfall. The high primary productivity puts arid and semi-arid rangelands under sharp focus as CO₂ sinks, whose role in the amelioration of greenhouse effect could be more important than is currently appreciated.

Keywords: climatic elements, aboveground biomass, belowground biomass, primary productivity

African Journal of Range & Forage Science 18 (2+3) 2001, 125—129
© 2001 NISC Pty Ltd, www.nisc.co.za

WN Ekaya¹* and JI Kinyamario²

1 Department of Range Management, University of Nairobi, PO Box 29053, Nairobi, Kenya

* Corresponding author, e-mail: [email protected]

Received 5 June 2001, accepted 9 October 2001

Data on litter production and decomposition in an arid rangeland in Kenya was collected over a two-year period. Litter sampling was carried out at monthly intervals using a rectangular 0.25m^-2 quadrat frame. Litter within the quadrats was handpicked and washed with running water to get rid of soil particles, dried, and weighed. Weights were expressed on organic matter basis. Monthly litter production ranged from 31.4g m^-2 to 130.0g m^-2. Mean monthly yield was 92.5 ± 26g m^-2, with a 28% coefficient of variation. There was no significant difference (p>0.01) in litter yield between 1992 and 1993. Rate of decomposition for aboveground material ranged from 0.005g g^-1 day^—1 to 0.084g g^-1 day^-1. The mean annual rate of decomposition was 0.026g g^-1 day^-1. Belowground plant material rates of decomposition spread from 0.009g g^-1 day^-1 to 0.062g g^-1 day^-1, with a mean annual rate of 0.041g g^-1 day^-1. Belowground material consistently decomposed faster than aboveground material. Peaks in both aboveground and belowground material decomposition rates coincided with rainfall peaks. Overall, in this arid environment, litter production and decomposition is pulsed in nature, and trends are closely related to rainfall occurrence. Moisture is thus a limiting factor both to the production and decomposition of litter. Belowground litter plays a significant role in nutrient cycling.

Keywords: dead herbage yield, dryland ecosystem, ecosystem function, nutrient cycling

African Journal of Range & Forage Science 18 (2+3) 2001, 131—135
© 2001 NISC Pty Ltd, www.nisc.co.za

Seasonal variation in forage quality and mimosine contents of two varieties of Leucaena leucocephala

AA Akingbade¹*, IV Nsahlai¹ and CD Morris²

1 Discipline of Animal and Poultry Science, University of Natal, Private Bag X01, Scottsville, Pietermaritzburg 3209, South Africa

2 Range and Forage Institute, Agricultural Research Council, Pietermaritzburg, South Africa

* Corresponding author, e-mail: [email protected]

Seasonal variation in forage quality and mimosine contents of two varieties (cultivars Cunningham and Spectra) of Leucaena leucocephala were evaluated. Cultivar (cv.) Cunningham was available throughout summer, autumn, winter and spring seasons while cv. Spectra was available during summer, winter and spring seasons. Differences in crude protein (CP) and mimosine contents of young Cunningham and Spectra leaves were not significant. Also, difference in mimosine contents of mature leaves of both cultivars were not significantly different. But mean CP content of mature Spectra leaves was significantly higher (24.9 vs 18.7%, p = 0.023) than that of leaves of mature Cunningham. Mean P content of young Spectra leaves was significantly higher (1.6 vs 2.10gkg^-1, p = 0.046) than that of young Cunningham leaves but mean Cu content of young Cunningham leaves was higher (10.5 vs 9.0mgkg^-1, p = 0.06) than that of young Spectra leaves. Mature leaves of cv. Spectra also had higher P content (2.3 vs 1.6gkg^-1, p = 0.06), but lower Ca (20.8 vs 30.5gkg^-1, p = 0.06) and Mg (5.6 vs 7.6gkg^-1, p = 0.06) contents than those of mature Cunningham leaves. Apart from CP, fat and mimosine, other analysed constituents increased with plant maturity.

Keywords: Nguni goats, forage trees

African Journal of Range & Forage Science 18 (2+3) 2001, 137—142
© 2001 NISC Pty Ltd, www.nisc.co.za

Research Note

A study of animal production under different grazing regimes in the False Upper Karoo at Grootfontein

e-mail: [email protected]

Received 6 May 2001, accepted 14 August 2001

Three systems of animal production on the fragile Karoo veld, i.e. multi-paddock rotational grazing, pauci-paddock rotational grazing and continuous grazing were studied. Of these systems, multi-paddock rotational grazing gave the best results with continuous stocking resulting in the worst. All three systems studied had long histories of both continuous stocking- and of pauci-paddock rotational grazing. The veld, under both these previous systems of management has stabilised, but at different seral stages. The continuously stocked range stabilised in the early developmental seral stage, consisting mainly of annual grasses. The veld under pauci-paddock rotational management stabilised in the early to mid-developmental seral stage, consisting mainly of perennial grasses. Experimentation on veld in a stable condition is a prerequisite if valid conclusions in respect of rotational grazing and continuous stocking are to be drawn. Rotational grazing on overgrazed veld will result in poor animal performance, while it is generally accepted that continuous stocking on pristine veld will result in excellent animal performance for a number of years. Multi-paddock rotational grazing outperformed both pauci-paddock rotational grazing as well as continuous stocking on both types of veld. This is surprising taking into account the mid-seral stage condition of the vegetation. This supports the argument that with improved veld management, an improved production potential is possible on veld even in a degraded or current condition.

Keywords: plains, pediments, grazing capacity

African Journal of Range & Forage Science 18 (2+3) 2001, 143—146
© 2001 NISC Pty Ltd, www.nisc.co.za

Research Note

The relation between canopy spread cover and the aboveground available phytomass of Nama Karoo subshrubs and grasses

e-mail: [email protected]

The canopy spread cover of Karoo bushes and grasses in the Nama Karoo was measured and related to the aboveground phytomass available to the grazing animal. A correlation coefficient of r = 0.7456, calculated for Karoo bushes, indicates that this method can be used to non-destructively estimate available phytomass. This in turn can be used for the prediction of current grazing capacity in areas where Karoo bushes are dominant in the vegetation. In the case of the Karoo grasses, a rather loose definition and interpretation of the canopy spread cover, led to a low correlation coefficient, r = 0.5662. This renders the present method largely unsuited for available phytomass estimations to be used in further predictions of grazing capacity in the mountainous areas where the studied grasses are dominant in the vegetation. In the more arid areas, however, the method can be used to advantage.

Keywords: veld type, grazing capacities