Calculation of grazing capacity and browse capacity for game species

G.N. Smit

Department of Animal, Wildlife and Grassland Sciences, University of the Free State,

P.O. Box 339, Bloemfontein 9300

1. Introduction

The basic requirement of management systems for sustainable game production from veld is to balance the stocking rate of the various game species with the grazing and browsing capacity of the veld. The grazing capacity of the grazeable portion of a homogeneous unit of vegetation can be defined as the area of land required to maintain a single animal unit (AU) over an extended number of years without deterioration of the vegetation or soil (ha/AU).  An animal unit (AU), also commonly referred to as a large stock unit (LSU), is defined as an animal with a mass of 450 kg, which gains 0.5 kg/day on forage with a digestible energy percentage of 55 %.  The stocking rate can be defined as the area of land in the system of management, which the manager has allocated to each animal unit in the system, and is expressed per length of the grazeable period of the year (ha/AU). 

The main difference, in practical terms, between the grazing capacity of the veld and the stocking rate is thus that the grazing capacity refers to the true number of animals that the vegetation can sustain and the stocking rate to the number of animals the manager perceived that the vegetation can sustain.   The ability to balance the true grazing capacity of the veld with the applied stocking rate sounds simple, but due to various reasons it can be difficult to achieve.  One of the more important reasons is that the effect of any particular combination of herbivore species and numbers is often difficult to judge and changes take place so gradually that this is not realised until veld degradation has already taken place. 

2. Factors that affect the productivity of the vegetation

The state of plant succession may have a profound effect on the productivity of the vegetation. Plant succession has been defined as a progressive development of vegetation in an area through a series of different plant communities, finally terminating in a climax community (Trollope et al. 1990).  Climax vegetation, in turn, has been defined as a final stable plant community in an ecological succession which is able to reproduce itself indefinitely under existing environmental conditions (Trollope et al. 1990).  Vegetation in a pioneer state, dominated by low yielding, annual grasses, will have a much lower grazing capacity than vegetation in a predominantly climax state. 

Important determinants of successional and retrogressional changes of the herbaceous layer include the spatial and temporal changes of the soil water regime (MacDonald 1978; Yeaton et al. 1986; Peel et al. 1991; Mott et al. 1992), and fire regime (Edroma 1981; Yeaton et al. 1988), as well as soil disturbances (Yeaton et al. 1986; Belsky 1987; Tueller & Platou 1991), soil nutrient status (Grunow et al. 1970; Walker & Knoop 1987), grazing (Walker 1980; Edroma 1981; Furniss 1982;  Archer et al. 1988; Friedel & Blackmore 1988; O'Connor 1991; Peel et al. 1991; Smit & Rethman 1992), and several other determinants like altitude, aspect, slope, stoniness and soil texture (Roberts 1971; Bredenkamp 1977). 

The effect of both intensity, frequency and season of grazing on herbaceous species composition has received much attention in the literature (Gammon 1978; Friedel & Blackmore 1988; O'Connor 1991; Smit & Rethman 1992).   Classification of grass species as Increasers and Decreasers in reaction to certain grazing regimes, is an example of the outcome of these investigations (Vorster 1982; Westfall et al. 1983; Friedel & Blackmore 1988; Janse van Rensburg & Bosch 1990; Smit & Rethman 1992).  The effect of grazing may also interact with other determinants.   Mott et al. (1992) concluded that drought during the dry season led to major mortality of defoliated Themeda triandra plants during the following growing season, while Heteropogon contortus and Panicum maximum did not display the same sensitivity.  Long term total exclusion of grazing also resulted in grass species changes.   Within savanna on a red clay soil, protected from grazing and fire, Enneapogon scoparius and Cenchrus ciliaris replaced climax grasses like Cymbopogon plurinodis, Themeda triandra, Heteropogon contortus, Elyonurus argenteus and Hyparrhenia species (Louw 1973). 

3. Plant characteristics that may influence the grazing and browsing capacity

3.1  Grasses

Trollope et al. (1990) defined sweetveld as veld in which the forage plants retain their acceptability and nutritive value after maturity, as opposed to sourveld, which is defined as veld in which the forage plants become unacceptable and less nutritious on reaching maturity.  In semi-arid areas, generally regarded as sweetveld, quantity is often more limiting than quality, and quality is subsequently regarded less important than quantity. Nevertheless, qualitative aspects cannot be ignored unconditionally, and warrant consideration.  A high fibre content and the presence of chemical substances like volatile oils may limit the acceptability of certain grass species to grazing herbivores.   In some grass species this acceptability may change with plant age as some species may be acceptable when young or after drying, while being avoided at other times.

3.2  Woody plants

The actual intake of available browse may be influenced by chemical defences of woody plants (Van Hoven 1984; Furstenburg 1991; Bryant et al. 1992), as well as nutritional characteristics of leaves in different phenological stages (Hall-Martin & Basson 1975; Cooper 1982; Owen-Smith & Cooper 1987; Cooper et al. 1988; Styles 1993).  Chemical defences of plants may include chemical substances, which may be poisonous (Smith 1992; Taylor & Ralphs 1992) or reduce palatability (Robbins et al. 1987; Bryant et al. 1992).  A diverse array of secondary metabolites deters feeding by mammals on woody plants.  Condensed tannins are especially important as a defence mechanism in woody plants (Haslam 1974; Van Hoven 1984; Martin et al. 1985; Furstenburg 1991; Hagerman et al. 1992).  Tannins are a diverse group of compounds, widespread among dicotyledonous forbs and trees, which precipitate protein (Asquith & Butler 1985; Robbins et al. 1987) and sometimes act as a toxin rather than a digestion inhibitor (Hagerman et al. 1992).

Herbivory by mammals may affect the chemical defences of woody plants.   In some cases browsing may result in increased defence (Baldwin & Schultz 1983; Van Hoven 1984; Furstenburg 1991) and in others decreased defence (Bryant et al. 1992).   Furstenburg (1991) found that the leaves of Acacia nigrescens trees displayed a 70 % increase in tannin concentration 2 minutes after disturbance, followed by a further, delayed, response after 30 to 100 minutes after the disturbance.   Plants known to have chemical defences against vertebrate herbivory are prominent on nutrient-deficient soils, while those with structural defences (e.g. spines) are predominant on fertile soils (Owen-Smith & Cooper 1987).   The effectiveness of these defences may vary between browser and woody species.   The success of chemical defences of trees of an Acacia nigrescens dominant community was demonstrated by Furstenburg (1991) who observed giraffe selecting plants with a low tannin content.  

Regarding structural defences, Cooper (1982) observed that the presence of straight spines or thorns has little effect on the feeding of goats and impala, while hooked thorns are more effective deterrents. 

4. Grazing and browsing capacity for game species

4.1  Substitution values for game

The use and application of the Animal Unit (AU) or Large Stock Unit (LSU) originated from conventional agriculture and is based on the metabolic mass of the animals involved (mainly cattle and sheep in different age and sex classes).  Comparison of different game species with the AU or LSU based on the metabolic mass, presents problems.  The use of AU/LSU-values for herbivore game species does not allow for ecological separation, and thus overlooks the potential for using the specialized and complementary resource-use habits of wild ungulates to maximize veld utilization.   

In an attempt to find a system more suitable to multi-species systems, Dekker et al. (1996) defined a grazer unit (GU) as the metabolic equivalent of a blue wildebeest (100 % grazer) with a mean body mass of 180 kg.  Similarly he defined a browser unit (BU) as the metabolic equivalent of a kudu (100 %) browser with a mean body mass of 140 kg.  By calculating the overlap for spatial distribution (plant community preferences), habitat variables and diet composition (grass : tree ratios) he was able to calculate substitution values for different game species based on potential competition for the same food source.   

The daily DM requirement of a GU will be 4.5 kg (2.5 % of body mass for a Blue wildebeest) (Owen-Smith 1999) and the daily DM requirement of a BU will be 3.5 kg (2.5 % of body mass for a Kudu) (Owen-Smith 1999) .  The substitution values of a few game species in terms of GU and BU are presented in Table 1

Table 1  Approximate substitution values of a few game species in terms of grazer units (GU) and browser units (BU).

Game species	Aver. mass (kg)	Intake (% of mass)	% grass	% leaves	GU	BU
Oribi	142	3.6	100	0	0.1	0
Grey Rhebok	202	3.4	100	0	0.2	0
Mountain Reedbuck	232	3.0	100	0	0.2	0
Blesbok	612	2.8	100	0	0.4	0
Bontebok	592	2.8	100	0	0.4	0
Southern Reedbuck	702	2.8	100	0	0.4	0
Gemsbok	2102	2.7	100	0	1.3	0
Red hartebeest	1202	2.7	100	0	0.7	0
Tsessebe	1262	2.6	100	0	0.7	0
Black wildebeest	1402	2.5	100	0	0.8	0
Blue wildebeest	1801	2.5	100	0	1.0	0
Burchell’s Zebra	2162	4.1	100	0	1.9	0
Sable antelope	2152	2.8	100	0	1.3	0
Waterbuck	2282	2.8	100	0	1.3	0
Roan antelope	2352	2.8	100	0	1.5	0
Buffalo	7152	2.4	100	0	3.8	0
Hippopotamus	1 410	1.5	100	0	4.7	0
White rhinoceros	1 727	1.4	100	0	5.4	0
Steenbok	122	4.1	50	50	0.05	0.07
Springbok	372	3.0	70	30	0.2	0.1
Impala	522	2.7	70	30	0.2	0.1
Lichtenstein’s Hartebeest	1712	2.6	80	20	0.8	0.3
Eland	4602	2.4	30	70	0.7	2.2
Elephant	3 8002	0.8	50	50	3.4	4.3
Duiker	212	4.0	0	100	0	0.2
Bushbuck	332	2.9	0	100	0	0.3
Nyala	622	2.6	0	100	0	0.5
Kudu	1401	2.5	0	100	0	1.0
Giraffe	8282	2.2	0	100	0	5.2
Black Rhinoceros	8652	1.5	0	100	0	3.7

       ¹ Average mass of herd     ² Average mass of mature female

4.2  Calculation of the grazing capacity

If the amount of herbaceous dry mass per hectare is known the grazing capacity can be calculated using the formula proposed by Moore et al. (1985), and again described by Moore & Odendaal (1987) and Moore (1989):

y = d  [ DM x f ]

                   r

where y = grazing capacity (ha GU-1)

d = number of days in a year (365)

DM = total grass DM yield ha^-1

f =  utilization factor

r = daily grass DM required per GU (2.5 % of body mass = 4.5 kg day-1)

The utilization factor, expressed as a decimal value, represents that part of the available grass material that can be consumed.  Actual consumption is limited by grazing preferences of the animals and losses due to trampling and environmental factors.  The percentage of available dry matter that the animals will actually consume is determined by factors like palatability of t he plant material and the species of animal (bulk feeder or concentrate feeder).   However, even when the animals will be able to consume a high percentage of the available dry matter, their intake should be limited to pre-determined level to avoid overgrazing.  The utilization factor may thus vary from 0.20 (20 %) to 0.50 (50 %), with the average of 0.35 (35 %) that is commonly used. 

With the DM production of individual species known it is now possible to assign a different utilization factor to each species in order to compensate for differences in the palatability and grazing value of different grass species: 

y = d  [ (DM₁ x f₁) + (DM₂ x f₂) + (DM₃ x f₃) ….. ]

                                              r

where DM₁ = grass DM yield ha^-1 of species 1

           DM₂ = grass DM yield ha^-1 of species 2

           DM₃ = grass DM yield ha^-1 of species 3

           …

 F₁ =  utilization factor for species 1

 F₂ =  utilization factor for species 2

 F₃ =  utilization factor for species 3

4.3  Calculation of the browsing capacity

The browsing capacity for browser game species is far more complex than calculating the grazing capacity for grazing game species.  The following may influence the calculation of the browsing capacity for a specific browser species:

(i)      Acceptability of the plant species available to the browsers,

(ii)     Height distribution of the browse material,

(iii)    Phenology of the plant species (whether they are evergreen, early winter deciduous or late winter deciduous species),

(iv)   Seasonal presence of flowers and pods/seeds with a high nutrient content.

Opposed to grasses where most of the dry matter produced during the season remain available during the dry winter as standing hay, leaves that drop from trees are often not as accessible or acceptable to browser species.  In the relatively dry savannas that receive summer rainfall the browsing capacity is often determined by the amount of food available during the dry months just before the onset of the new season (August to October).   During this time areas that can be described as “critical resource areas”, such as river ecosystems may play an important role in the survival of browsers during this critical pre-season dry period.  

In it’s simplest form the browsing capacity can be calculated using a similar formula than the one used to calculate the grazing capacity, with the addition of phenology (p) as an additional variable:

y = d  [ DM x f x p ]

                     r

where y = browsing capacity (ha BU-1)

d = number of days in a year (365)

DM = total leaf DM yield ha^-1

f =  utilization factor

p = phenology

r = daily leaf DM required per BU (2.5 % of body mass = 3.5 kg day-1)

The above formula will, at best, render an average browsing capacity value for the year.  In vegetation types dominated by a heterogeneous deciduous woody species a more accurate approach will be to apply the formula above without the incorporation of the p-value (phenology).  The calculated browse capacity value (at peak biomass) is then adjusted or corrected for the specific p-value of each month of the year, thus correcting the browse capacity value in relation to the seasonal fluctuation in the leaf presence (Table 2)

From this example it is clear that the month of September is the most critical and that this bottleneck period is the most important determinant of the actual browsing capacity.

An even more sophisticated approach may be followed.  Like before this approach assumes that a leaf quantification technique (like the BECVOL method – Smit 1996) is used to determine the leaf yield of woody plants on a species basis.  It also assumes that the ranch/reserve was divided into a number of vegetation units (U₁…U_x) and that the leaf biomass was determined in each of these vegetation units.  The different woody species are subsequently classified as (i) evergreens, (ii) late winter deciduous, (iii) intermediate winter deciduous, and (iv) early winter deciduous.  The total leaf dry mass of the various plant groups per hectare up to a specific height (H^(x)) is then calculated as DM₁^(x), DM₂^(x), DM₃^(x) and DM₄^(x).  Since the amount of available browse can vary considerably from month to month, it is best to do a calculation for each month or at least each season (spring, summer, autumn and winter) and for different browsing heights.

Table 2  Example of adjusting the browse capacity at peak biomass for a reduction in browse availability during each month of the year according to specific p-values (leaf phenology).  (Equation:  adjusted browsing capacity = browsing capacity x (1/p).    

Month	P-value (leaf phenology)	Calculated browse capacity at peak biomass (ha/GU)	Adjusted browse capacity (ha/GU)
January	1.0	8.20	8.20
February	1.0	8.20	8.20
March	1.0	8.20	8.20
April	0.9	8.20	9.11
May	0.8	8.20	10.25
June	0.7	8.20	11.71
July	0.6	8.20	13.67
August	0.3	8.20	27.33
September	0.2	8.20	41.0
October	0.6	8.20	13.67
November	0.9	8.20	9.11
December	1.0	8.20	8.20

Calculation of the total amount of browse material available for actual consumption up to a height of 2.0 m (DM^(2.0)) as a total of three vegetation units (U₁ .. U₃):

 

DM^(2.0)U₁ = [(DM₁^(2.0) x f₁ x p₁) + (DM₂^(2.0) x f₂ x p₂) + (DM₃^(2.0) x f₃ x p₃) + (DM₄^(2.0) x f₄x

                   p₄)] x A₁

 

DM^(2.0)U₂ = [(DM₁^(2.0) x f₁ x p₁) + (DM₂^(2.0) x f₂ x p₂) + (DM₃^(2.0) x f₃ x p₃) + (DM₄^(2.0) x f₄x

                   p₄)] x A₂

 

DM^(2.0)U₃ = [(DM₁^(2.0) x f₁ x p₁) + (DM₂^(2.0) x f₂ x p₂) + (DM₃^(2.0) x f₃ x p₃) + (DM₄^(2.0) x f₄x

                   p₄)] x A₃

 

where, DM_1..4^(2.0) = total DM/ha to a height of 2.0 m for each of the plant groups,  

            f_1..4   = utilization factor for each of the four plant groups expressed as

                       a decimal,

            p_1..4  = leaf availability expressed as a decimal, and

            A_1..3 = area cover by each vegetation unit (ha) 

DM^(2.0) = DM^(2.0)U₁ + DM^(2.0)U₂ + DM^(2.0)U₃

 

Total BU (that can utilize browse to a height of 2.0 m) that can be kept during

that month (or season):

 

                                         DM^(2.0)

                                     =      r

 

where, r = daily browse (DM) required per BU (2.5 % of body mass = 3.5 kg)

Limited scientific information currently exists on which to base the utilization factor (f), but indications are that it is very low.  In the case of Black Rhinoceros it can be as low as 8 % (f = 0.08), and up to about 20 % or more (f = 0.20) for other browsers.  The estimated percentage leaf presence (p = phenology) for the various plant groups can theoretically vary from 100 % (p = 1.0) in the case of evergreens to 0 % (p = 0.0) during winter for the early deciduous group.  However, there are indications that browsers may utilize the tips of shoots and twigs, even if no leaves are present.  This implies that the value of p will always be above 0.

 References

Archer S., Scifes C. & Bassham C.R. 1988.   Autogenic succession in a sub-tropical savanna: conversion of grassland to thorn woodland.  Ecological Monographs 58: 111-127.

Asquith T.N. & Butler L.G. 1985.   Use of dye-labled protein as spectro-photometric assay for protein precipitants such as tannin.   J. Chem. Ecol. 11: 1 535-1 544.

Baldwin I.T. & Schultz J.C. 1983.   Rapid changes in tree leaf chemistry induced by damage: evidence of communication between plants.   Science 221: 277-279.

Belsky A.J. 1987.   Revegetation of natural and human-caused disturbances in the Serengeti National Park, Tanzania.  Vegetatio 70: 51-60.

Bredenkamp G.J. 1977.   The grasses of the Suikerbosrand Nature Reserve:  their habitat preferences and synecological significance.   Proc. Grassld. Soc. sth. Afr. 12: 135-139.

Bryant J.P., Reichardt P.B. & Clausen T.P. 1992.  Chemically mediated interactions between woody plants and browsing mammals.   J. Range Manage. 45: 18-24.

Cooper S.M. 1982.   The comparative feeding behaviour of goats and impalas.  Proc. Grassld. Soc. sth. Afr. 17: 117-121.

Cooper S.M., Owen-Smith N. & Bryant J.P. 1988.   Foliage acceptability to browsing ruminants in relation to seasonal changes in leaf chemistry of woody plants in a South African savanna.   Oecologia (Berlin) 75: 336-342.

Dekker B., Van Rooyen N. & Bothma J. du P. 1996.  Habitat partitioning by ungulates on a game ranch in the Mopani veld. South African Journal of Wildlife Research, 26: 117-122.

Edroma E.L. 1981.   The role of grazing in maintaining high species composition in Imperata grassland in Rwenzori National Park, Uganda.   Afr. J. Ecol. 19: 215-223.

Friedel M.H. & Blackmore A.C. 1988.   The development of veld assessment in the northern Transvaal savanna I. Red turfveld.   J. Grassl. Soc. South. Afr. 5: 20-37.

Furniss P.R. 1982.  A model of resource allocation for savanna grasses.   S. Afr. J. Bot. 1: 1-6.

Furstenburg D. 1991.   Die invloed van tanniene in plante op die voedingsekologie van Kameelperde (Giraffa camelopardalis).   M.Sc.-thesis, University of Pretoria, Pretoria.

Gammon D.M. 1978.   A review of experiments comparing systems of grazing management on natural  pastures.   Proc. Grassld. Soc. sth. Afr. 13: 75-82.

Grunow J.O., Pienaar A.J. & Breytenbach C. 1970.   Long term nitrogen application to veld in South Africa.   Proc. Grassld. Soc. sth. Afr. 5: 75-90.

Hagerman A.E, Robbins C.T., Weerasuriya Y., Wilson T.C. & McArthur C. 1992.   Tannin chemistry in relation to digestion.   J. Range Manage. 45: 57-62.

Hall-Martin A.J. & Basson W.D. 1975.  Seasonal chemical composition of the diet of Transvaal Lowveld giraffe.   J. S. Afr. Wildl. Mgmt. Ass. 5: 19-21.

Haslam E. 1974.  Polyphenol-protein interactions.  Biochem. J. 139: 285-288.

Janse van Rensburg F.P. & Bosch O.J.H. 1990.   Influence of habitat differences on the ecological grouping of grass species on a grazing gradient.   J. Grassl. Soc. South. Afr. 7: 11-15.

Louw A.J. 1973.  Botaniese samestelling van beskermde persele in Mopanie-veld.  Agroplantae 5: 23-24.

MacDonald I.A.W. 1978.    Pattern and process in a semi-arid grassveld in Rhodesia.   Proc. Grassld. Soc. sth. Afr. 13: 103-109.

Martin M.M., Rockholm D.C. & Martin J.S. 1985.   Effects of surfactants, pH and certain cations on precipitation of proteins by tannins.   J. Chem. Ecol. 11: 485-494.

Moore A. 1989.  Die ekologie en ekofisiologie van Rhigozum trichotomum (Drie-doring).  Ph.D-thesis, University of Port Elizabeth, Port Elizabeth.

Moore A. & Odendaal A. 1987.  Die ekonomiese implikasies van bosverdigting en bosbeheer soos van toepassing op 'n speenkalfproduksiestelsel in die doringbosveld van die Molopo-gebied.   J. Grassl. Soc. South. Afr. 4: 139-142.

Moore A., Van Niekerk J.P., Knight I.W. & Wessels H. 1985.   The effect of  Tebuthiuron on the vegetation of the thorn bushveld of the northern Cape - a preliminary report.    J. Grassl. Soc. South. Afr. 2: 7-10.

Mott J.J., Ludlow M.M., Richards J.H. & Parsons A.D. 1992.    Effects of moisture supply in the dry season and subsequent defoliation on persistance of the savanna grasses Themeda triandra, Heteropogon contortus and Panicum maximum.   Aust. J. Agric Res. 43: 241-260.

O'Connor T.G. 1991.  Influence of rainfall and grazing on the compositional changes of the herbaceous layer of a sandveld savanna.   J. Grassl. Soc. South. Afr. 8: 103-109.

Owen-Smith N. 1999.  The animal factor in veld management.  In: Tainton, N.M.  Veld management in South Africa.  University of Natal Press, Pietermaritzburg.  pp: 117-138.

Owen-Smith N. & Cooper S.M. 1987.  Palatability of woody plants to browsing ruminants in a southern African savanna.  Ecol. 68: 319-331.

Owen-Smith N. & Cooper S.M. 1988.  Plant palatability and its implications for plant-herbivore relations.  J. Grassl. Soc. South. Afr. 5: 72-75.

Peel M.J.S., Grossman D. & Van Rooyen N. 1991.  Determinants of herbaceous plant species composition on a number of ranches in the north-western Transvaal.   J. Grassl. Soc. South. Afr. 8: 99-102.

Robbins C.T., Mole S., Hagerman A.E. & Hanley T.A. 1987.   Role of tannins in defending plants against ruminants: reduction in dry matter digestion.   Ecol. 68: 1 606-1 615.

Roberts B.R. 1971.  Habitat preferences of twenty-seven grasses.   Proc. Grassld. Soc. sth. Afr. 6: 44-49.

Smit G.N. 1996.  BECVOL: Biomass Estimates from Canopy VOLume (version 2) - users guide.  Unpublished manual, University of the Free State, Bloemfontein.  22 pp.

Smit G.N., Rethman N.F.G.,1992.  Inter-related floristic changes associated with different long-term grazing treatments in Sourish Mixed Bushveld.   J. Grassl. Soc. South. Afr. 9: 76-82.

Smith G.S. 1992.  Toxification and detoxification of plant compounds by ruminants: an overview.  J. Range Manage. 45: 25-30.

Styles C.V. 1993.   Relationships between herbivores and Colophospermum mopane of the northern Tuli Game Reserve.  M.Sc-thesis, University of Pretoria, Pretoria.

Taylor C.A. & Ralphs M.H. 1992.   Reducing livestock losses from poisonous plants through grazing management.    J. Range. Manage. 45: 9-12.

Trollope W.S.W., Trollope L.A. & Bosch O.J.H. 1990.    Veld and pasture management terminology in southern Africa.   J. Grassl. Soc. South. Afr. 7: 52-61.

Tueller P.T. & Platou K.A. 1991.   A plant succession gradient in a big sagebrush/grass ecosystem.  Vegetatio 94: 57-68.

Van Hoven W. 1984.  Tannins and digestibility in greater kudu.  Can. J. Anim. Sci. 64 (suppl.): 177-178.

Vorster M. 1982.  The development of the ecological index method for assessing veld condition in the Karoo.   Proc. Grassld. Soc. sth. Afr. 17: 84-89.

Walker B.H. 1980.  Stable production versus resilience:  A grazing management conflict.    Proc. Grassld. Soc. sth. Afr. 15: 79-83.

Walker B.H. & Knoop W.T. 1987.   The response of the herbaceous layer in a dystrophic Burkea africana savanna to increased levels of nitrogen, phosphate and potassium.    J. Grassl. Soc. South. Afr. 4: 31-34.

Westfall R.H., Van Rooyen N. & Theron G.K. 1983.   Veld condition assessments in Sour Bushveld.   Proc. Grassld. Soc. sth. Afr. 18: 73-76.

 

Yeaton R.I., Frost S. & Frost P.G.H. 1986.  Direct gradient analysis of grasses in a savanna.    S. Afr. J. Sci. 82: 482-487.

 

Yeaton R.I., Frost S. & Frost P.G.H. 1988.   The structure of a grass community in Burkea africana savanna during recovery from fire.   S. Afr. J. Bot.  54: 367-371.