Acacia tortilis subsp. heteracantha productivity in the Tugela Dry Valley Bushveld: preliminary results

Acacia tortilis Hayne subsp. heteracantha (Burch.) Brenan dominates secondary succession in the Tugela Dry Valley Bushveld of the Natal midlands. The parts of KwaZulu in this veld type are impoverished, overpopulated and over-grazed. Preliminary results indicate that at a density of 2 700 ± 600 trees/ha there is a standing crop of c.2,87 t/ha (DM) of acacia twigs suitable for hand pruning and milling into fodder, but that this is a costly process. Herbage biomass peaked at 0,73 t/ha (DM). Veld condition assessments suggested a stocking rate of|0,l AU/ha (grazers), but actual grazer stocking rates may be many times this density. It is recommended that the browser/grazer ratio be altered to make use of the c. 1,05 t/ha (DM) of shoot growth produced annually by A. tortilis subsp. heteracantha.


INTRODUCTION
This paper presents preliminary findings of a study of the productivity of Acacia tortilis Hayne subsp.heteracantha (Burch.)Brenan relative to that of the herbage at contrasting sites in secondary acacia veld in Bioclimatic Zone 10 of Natal (Phillips, 1967).The aim of the study is to assess the carrying capacity of the veld for browsers and grazers in order to maximize production and minimize environmental degradation.

Study species
Acacia tortilis (Forsk.)Hayne, a flat-topped thorn tree with indehiscent, coiled pods, is widely distributed in the drier parts of Africa -from the northern Cape and Natal to Arabia (Ross, 1975).Subspecies heteracantha (Burch.)Brenan, the subject of this study, occurs throughout the interior of Natal and indeed throughout most of southern Africa.Its basionym is A. heteracantha Burch., and synonyms are A. litakunensis Burch., A. spirocarpoides Engl, and A. maras Engl.

Study area
All data presented in this study were collected on a farm on the Weenen-Msinga border (Fig. 1), 26 km upstream from Tugela Ferry, but the study will be extended to other areas of A. tortilis subsp.heteracantha semi-deciduous bush, where this occurs as a secondary community in Tugela Dry Valley Bushveld (Edwards, 1967).This type of vegetation falls within Bioclimatic Group 10 (Phillips, 1973), which Thorrington-Smith et al., (1978) describe as having low agricultural potential, and as being overstocked, denuded and eroded, with spectacular bush encroachment.The topography is broken with many stoney hills.Temperatures are extreme with frost in winter and a mean daily summer maximum of 30°C.Rainfall is low (637 mm), unreliable, and generally accompa nied by electric storms (Edwards, 1967;West, 1950).
The population density of Bioclimatic Zone 10 of KwaZulu is given as 50 people/km2 (Thorrington-Smith et al., 1978), but hut counts suggest that it may be higher (Schulze, 1969).Topographic factors and resettlement schemes have resulted in areas of locally high population density, and the environ mental degradation in the vicinity of these is severe and is characterized by vegetation destruction, dongas and drying up of springs.As Thorrington-Smith et al. (1978) point out, cultivation without irrigation is hazardous, so much of the money and food required is brought in from the cities or irrigated white farms, thereby artificially boosting the carrying capacity.Livestock, particularly cattle, are cherished as insurance to be sold in times of dire need or to be sacrificed on special occasions.Tribal land is common land exploited by all to provide fuel, grazing and building materials, but where carrying capacity has been exceeded the "system of the commons is absolutely intolerable" and the environ ment is degraded, further reducing carrying capacity (Hardin, 1977).
Secondary acacia communities are the product of ploughing, over grazing or other forms of disturban ce (Edwards, 1967;West, 1950).Under heavy grazing such bush lands deteriorate rapidly and become more woody or eroded (Kelly & Walker, 1976;Whyte, 1947).Removal of thorn trees would be not only costly and difficult to implement, but would deprive local people of 'free' fuel and fencing.
As browse appears to be under-utilized and herbage over-utilized in semi-arid rangelands (Whyte, 1947;Trollope, 1975;Kelly, 1977) it may be cheaper, more acceptable and more productive to alter the grazer-browser ratio.

Density, biomass and production o f A. tortilis subsp. heteracantha
Density was estimated from stem and tree (stems joined on or below ground level) counts on 80 10 x 10 m plots distributed over five sites.Stem diameter, referred to as basal diameter in this paper, was measured approximately 20 cm above soil level.To estimate the standing crop of biomass available for milling into fodder, trees were pruned with long handled shears, and the shorn shoots (approximately 1 cm diam.) of individual trees at sites 2, 4, 5, 6 and whole plots at sites 1 & 3 were weighed in the field using a spring balance.Two methods were used to estimate annual production.Firstly, all the new season's twigs were measured, counted and clipped from 9 trees of known diameter, then weighed.Secondly, 10 new season's twigs from each of 10 trees were measured.The number of twigs per branch approximately 2,2 cm diam.was recorded, as was the number of such branches per tree of known diameter.Total twig mass was estimated from total length.This method was also used when comparing the growth of pruned and unpruned trees, and of trees of different ages.
The pod crops of 64 measured trees were weighed in the field, and podbearing trees counted in two quarter hectare quadrats.To obtain diameter-mass regressions, 68 trees were felled at site 6.Diameters were measured and prunable twigs and wood weighed separately in the field.The leaf dry mass/pruned branch fresh mass ratio from site 1 was used to provide a rough estimate of dry leaf yield per unit area.Leaf production and retention data are based on counts of leaves on the terminal 10 cm of new seasons shoots.

Herbage composition, utilization, cover and biomass
Species composition at two sites was assessed using the step-point method (Mentis, in press) and data obtained were used to estimate veld condition (Tainton et al., 1980) and utilization ratios for each species: utilization ratio = % use / % occurrence where use is evidence of grazing.Cover at these sites was measured at two monthly intervals using a 100 point grid half a metre square.Ten 1/4 m2 plots per 1/4 ha quadrat were then clipped, oven dried and weighed to obtain a grid number/mass relationship.The grid was also used when comparing herbage cover beneath acacia canopies with that in the open.At site 6 where shrubby forbs prevented the use of the grid, a perforated horizontal bar was used to assess cover.

The effects o f pruning, size-class and species on shoot growth
Pruned trees tended to produce longer and more branched shoots with greater diameters than unpruned trees, and this trend was more marked under heavy pruning than under light pruning (Table 2).The most dramatic differences were in mean shoot length at Site 1 (F[ .46-8, 6; P = 0,01), and in stem diameter at Site 3 (F[ 23 = 9,2; p = 0,01).As fewer shoots were produced by pruned trees, the total shoot length per branch c. 2,2 cm diam.was only marginally greater in pruned than in unpruned trees (FU2 = 2,9; P= 0,2).

Leaf production and loss
During the summer the density of leaves on new seasons twigs is 1,84 (0,1) leaves/cm.Production of leaves is not significantly affected by pruning, but leaf retention during winter is influenced by site
( F 3 .34 7 = 11,7; P = 0,001), pruning (F1> 238 = 6, 8; P = 0,01), and among pruned trees, by tree diameter class (F2tll7 = 4,4: P = 0,05).Soil moisture during the dry season appears to be the major site factor affecting leaf retention.In July, leaf retention of trees growing on the banks of the Tugela was c. 71% of the summer density at this site, 20 m above the Tugela 57% of the leaves were retained, 200 m above the Tugela on old fields, 30% and on very shallow soils at this altitude, 10% A. tortilis subsp.heteracantha trees are often leafless on higher, drier ground during the winter, and it is possible that frost also affects leaf attention.
The ratio of pruned branch fresh mass to air-dry leaf mass was 1: 0,052.Sites 1-5 would therefore yield 237-355 kg/ha air dry leaves.

Basal diameter/mass relationships and size-class distribution
An old A. tortilis subsp.heteracantha stand with a closed canopy had a wood crop of 23,2 (5,9) t/ha (FM) as compared with the 58-76 t/ha (FM) wood crop of 12 year old A. tortilis subsp.heteracantha trees of a north African subspecies growing in plantations in India (Muthana & Arora, 1980).
Weed, pruneable branch and total tree fresh mass increase logarithmically on basal diameter (Fig. 3).The wood yield of a self-established acacia stand should be estimated from the sum of the products of mass and frequency for each size class, and not from mean tree diameter as neither frequency nor mass are normally distributed in the size classes (Fig. 4).This also applies to branch and total biomass, as it is not uncommon for c. 25% of the trees to constitute c. 75% of the biomass.
During the period August 1980-August 1981 there was no significant change in the circumferen ces of a sample of 92 old stems marked at Site 6.The young trees at Site 1, however, showed a mean diameter increment of 0,58 (0,17) cm, a change significant at the 1% level.
A. tortilis subsp.heteracantha at Site 3 starts to bear pods when in the 6-8 cm diameter class.Pod yield then increases exponentially on diameter (Fig. 5), but possibly falls off steeply in very old trees, c. 30 cm diam.Stunted trees on very shallow shaley soils started to bear in the 2-4 cm diam.class, and yield increased linearly on basal diameter.At site 3 TREE BASAL DIAMETER (cm) -Relationship between pod yield (g) and basal diameter (cm) at Site 3. Note: the largest tree is excluded from the regression calculation.Dairy meal where fruiting trees had a mean basal diameter of 11,2 (0,8) cm, the mean yield was 0,73 (0,25) kg/tree, and the maximum yield was 7,2 kg/tree.Pod yields recorded in the study area range from 1 -139 kg/ha: very low compared with the 10-12 kg/tree in A. tortilis plantations in India (Muthana & Arora, 1980) or the 250-500 kg/ha yield cf pods in natural stands in Zimbabwe (West, 1950).

Nutritional value, processing costs, and sundry uses of acacia products
The nutritional value of acacia fodders is acceptable, however, they are not cheap (Table 3).Costs shown include labour, fuel and maintenance of the machinery, but not the capital costs.Dry green leaves were obtained by stacking branches until they were dry, then beating off the leaves.A similar method is used for harvesting Leucaena leaves in Australia (Wildin, 1980) and would have been ideal for poor rural areas had not the thorns been shed with the leaves making the product unacceptable to cattle.Cattle, goats and horses accept milled branches and pods, and cattle have been maintained for up to two years on these products at the C.A.O.co-op farm at Weenen.Goats will eat 300-500 g pod meal per day in addition to browsing, and the meal may also be used as a poultry feed.
In the Msinga District, Acacia tortilis subsp.heteracantha, although not a preferred fuel wood, is much used for cooking and heating; the bark is used to tie bundles and roof laths, and the branches to construct thorn fences around kraals and gardens.These fences, which are renewed every year or two, comprise large branches packed together at a density of about 6,7 branches/m to form a breast high barrier.Three gardens sampled in Nomoya were enclosed by 65 (5) m of thorn fence comprising some 400 branches c. 4 cm diam.or the equivalent of 0,2 ha of bushes-an area ten times as great as that of a garden.

Herbage biomass, condition and response to acacia cover
Biomass on heavily grazed oldlands (Sites 3 & 7) peaked in January at 830 (139) and 630 (76) kg/ha DM.Although if of the grass at site 7 died between mid January and mid March, protein content dropped only slightly from 10,5% in February to 9,9% in April.The April protein content of Bothriochloa-dominaled Site 3 was 5,1%.Fibre content of the herbage at both sites was low (25,9%) probably because the grasses, being continually grazed produced few inflorescences.
A veld condition assessment (Table 4) showed that all three study plots were overgrazed relative to the Bioclimatic Zone 10 benchmark.They also contained shade tolerant and old land species not found in the benchmark.Although the protein content and the species composition score of Site 3 is lower than that of Site 7, the former site is dominated by selected or neutral species, and the latter by avoided species.

CONCLUSION
A standing crop of 5,7 t/ha FM (c.2,9 t/ha DM) of acacia branches was available for hay making in secondary A. tortilis subsp.heteracantha veld, however, milling woody plants is an expensive way of maintaining an artificially high grazer population.The annual production of acacia shoots, 2,1 t/ha FM (c.1,1 t/ha DM) was greater than the herbage biomass peak of 0,7 t/ha.The veld should therefore be stocked with as many or more browsers than grazers (within the limits of carrying capacity).Whereas it may be difficult to convince some people of the merits of increasing the browser/grazer ratio, because of social and cultural attitudes to cattle (ARDRI, 1981), this should be less disruptive and costly than an attempt to remove the trees and establish pasture for beef production -the solution implied by Thorrington-Smith et al., (1980).It may PRELIMINARY RESULTS be feasible to camp and rest common land, but this is also likely to be complicated and costly.
k No specific studies have been undertaken on the growth of key browse species' (Trollope, 1981), and the value of leguminous trees for soil improvement, fuel and fodder has only recently been acknowled ged (Folker, 1981).Although, as Janzen (1973) points out, they are not the solution to land management problems, we cannot afford to get rid of them before we have investigated their potential.

F
i g .1.-Study area in relation to the Weenen and Msinga Districts, and the distribution of Bioclimatic Zone 10 in the Tugela Valley.Inset: South Africa, Natal Province shaded.

F
i g .4 .-Tree d en sity a n d b io m as s d is tr ib u tio n in basal d ia m e te r classes.

TABLE 1 .
-Harvestable Acacia biomass and shoot production

TABLE 2 .
-Pruning, size-class and species effects on shoot growth *Pairs of results which do not differ significantly are italicized.
*t Utilization: see methods.