Exotic woody plant invaders of the Transvaal

The frequency and abundance o f exotic, woody plant invaders were recorded in 60% of the quarter degree squares in the study area. Sixty-one invaders were encountered o f which the most important and aggressive were Acacia dealbaia, Populus spp., Melia azedarach, Opuntia ficus-indica, Salix babylonica and Acacia mearnsii. Invasion patterns are discussed and an attempt is made to correlate distribution with environmental factors. Attention is drawn to the areas of greatest invasion and the areas that are liable to show the greatest expansion in the future-


Survey history and objectives
The objectives of this survey are: to produce a checklist of the major exotic, woody invaders of streambank, roadside and veld habitats in the Transvaal (see Appendix); to determine the pattern of exotic woody invasion as a whole and for individual species; to attempt to correlate distribu tion with environmental factors; to determine which are the most important and aggressive invaders.
The first part of the Transvaal that was surveyed was the south central region, 10 000 km2 (Wells et al., 1980).The survey was then extended to include the section east of Pretoria to the Kruger National Park, and stretching from Volksrust in the south to Louis Trichardt in the north (Duggan & Henderson, 1982).The survey has now been completed with the inclusion of 113 samples in the western sector and a further 58 samples in the eastern sector (Fig. 1).

The study area
The Transvaal covers an area o f approximately 262 449 km2.The greater part of the province consists of table lands (highveld) ranging from 600 -1 800 m in altitude, the highest point being in the east where the Drakensberg overlooks the lowveld.The lowveld (altitude range of 200 -600 m) and the low-lying portions of the northern and central sections o f the province have a tropical to sub-tropicai climate.The southern highveld and the upper slopes of the escarpment are more temperate and experience frost in winter.
The highest rainfall in the Transvaal occurs along the upper slopes of the Drakensberg escarpment and the average rainfall can exceed 2 000 mm per annum.
The driest parts of the Transvaal occur along the Limpopo River Valley in the north and surrounding country, the eastern lowveld and the south-western Transvaal.These parts experience a mean annual rainfall of 250 -500 mm.
The rest of the Transvaal receives a mean annual rainfall of 500 -900 mm, with the rainfall generally increasing from west to east.

Sampling method and intensity
The sampling method was specifically designed to make use of otherwise unproductive travelling time whilst officers are engaged on other projects.
The method consisted of making continuous recordings of roadside and veld invaders from a moving vehicle and of streambank invaders at water-course crossings (preferably while the vehicle was stationary).Abundance ratings were based on frequency of encounter within each sample unit (the quarter degree square).Abundance ratings for streambank habitats were based on estimates at specific sites.

Abundance ratings
Since the last publication (Duggan & Henderson, 1982) changes have been made to the abundance ratings for roadside and veld habitats.Two ratings have been inserted between the original 5 and 6 ratings and 2 have been inserted below the old rating 1, i.e. there are now 11 abundance ratings.This was done because the gap between the 5 and 6 ratings was too large and the rating 1 was not low enough to accommodate the most infrequently occurring species.
The abundance ratings for roadside and veld habitats and streambank habitats are given in Table 1.

Sampling level envisaged and achieved
The minimum sampling level that was envisaged for the Transvaal was 40% of the total quarter degree squares at an average of 64 km per square.This was based on an analysis of the effectivity of sampling at various levels (Henderson & Duggan, unpublished report).
The sampling level that has been achieved is: 60% (248 of the total 410 quarter degree squares) at an average of 33 km per square.That is, a greater number of sampling units was sampled but at a less intensive level.Twenty-two out of the total of 24 veld types recognized by Acocks (1975) as being present in the Transvaal were sampled.The smallest veld types, numbers 53 and 55 were unsampled.Invader species were recorded separately in each veld type passed through in each sample unit.
The sampling coverage for each Acocks veld type in terms of quarter degree squares sampled and total kilometres travelled is shown in Table 2.The number of water-course recordings, quarter degree squares sampled and kilometres travelled in each of the four broad veld type categories, tropical veld types, tem perate veld types and the whole survey area are shown in Table 3. Sixty-one invaders were encountered in the Transvaal.These species are listed in the checklist together with a further 17 which were obtained from various literature and other sources (see Appendix).

The streambank habitat
Four hundred and ten recordings were made in which 40 different exotic, woody invader species were recorded, with up to 9 species in one sample.

Analysis according to veld type
The number and abundance of invader species recorded in streambank samples in each of the four broad veld type categories are given in Table 4. 1. Salix babylonica is not the most frequent invader throughout the temperate veld types but in the false grassveld i t 'i s superseded by Acacia dealbata and joined by Populus albalcanescens and Acacia mearnsii (Fig. 5), 2. Melia azedarach is not the most frequent invader throughout the tropical veld types, there being several more frequent invader species in the forest veld types, 3. Acacia dealbata and Acacia mearnsii are more frequent, particularly in the tem perate veld types, than was previously recorded.Although these species (together with Psidium guajava and Caesalpinia decapetala) were not recorded in fhe western Transvaal -a total of 113 quarter degree squaresthis is most likely due to unavailability of source plants and not unsuitability to the environmental conditions, 4. A n additional 18 exotic, woody invader species have been recorded since our last report.These include: Agave spp., Eucalyptus spp., Pinus spp.including P. patula, Opuntia ficus-indica, Toona ciliata, Populus nigra, Pyracantha angustifolia and Bambusa balcooa.
In the temperate veld types Salix babylonica retains top position but the gap between this species and Acacia dealbata is reduced.Populus albalcane scens retains third position.In the tropical veld types Melia azedarach and Populus albalcanescens retain first and second positions respectively.Acacia dealbata supersedes Salix babylonica and Ricinus communis to take third position.

Roadside and veld habitats
Exotic woody invaders were encountered in 212 (86%) of the 248 quarter degree squares sampled.
Sixty invader species were recorded, with an additional 24 species being recorded since the last publication.The new species include in the western Transvaal: Opuntia cultivar with round, blue-green cladodes, Gleditsia triacanthos, Prosopis velutina, Schinus molle and Trichocereus sp. and in the eastern Transvaal: Toona ciliata, Tecoma stans, Bauhinia variegata, Bambusa balcooa, Passiflora edulis and Rosa sp.

Analysis according to veld type
The number and abundance of invader species recorded in roadside and veld samples in each of the four broad veld type categories are shown in Table 5.A comparison of Fig. 13 (forest) and Fig. 14 (bushveld) reveals great differences: 1.In forest, approximately 24 species have a percentage frequency greater than 10% compared with only about 10 species in bushveld.
2. In bushveld only Opuntia ficus-indica and Melia azedarach were recorded in more than 30% of the sample units, whereas in forest at least 11 species were recorded in more than 30% of the sample units.
3. A number of species which are well represented (27% frequency and higher) in forest namely Solatium mauritianum, Rubus spp., Caesalpinia decapetala, Pinus spp.and Acacia mearnsii are poorly represented in bushveld.
These differences indicate that invasion has progressed furthest in forests and that the bushveld has a greater resistance to invasion.The greater degree of invasion in forest can be attributed mainly to the more favourable growing conditions throughout and the greater disturbance that these veld types have been subjected to.Large tracts of bushveld are very dry and inhospitable to the more mesic invader species such as those mentioned under point 3 above.

Importance:
The importance values for roadside and veld invader species were calculated in the same way as for streambank invaders.Again, the species with the highest values and which collectively added up to a total of approximately 160 points were plotted (Figs 20,21 & 22).
A comparison of the percentage frequency histogram Fig. 19 and importance histogram Fig. 22 shows that for the whole survey area Opuntia ficus-indica, Melia azedarach and Eucalyptus spp.retain first, second, and third positions respectively.Solanum mauritianum climbs from about 11th position to 4th position.Acacia dealbata and A. mearnsii shift from 6th to 5th position.Both Agave spp.and Ricinus communis drop considerably.
Solanum mauritianum climbed from obscurity in the percentage frequency histogram to prominence in the importance histogram because of the very large numbers of individuals that were recorded in a few quarter degree squares along the Drakensberg escarpment.In contrast the Agave spp.and Ricinus communis plummeted in position because they generally occurred in small numbers.
In the tem perate veld types Acacia dealbata and A. mearnsii overtake Eucalyptus spp. to take first and second positions respectively, thus indicating that they are very much more abundant than the Eucalyptus spp.In the tropical veld types Opuntia ficus-indica and Melia azedarach retain first and second positions respectively.Solanum mauritianum and Eucalyptus spp.climb to third and fifth positions respectively.Jacaranda mimosifolia drops one position to fourth place and Agave spp.and Ricinus communis drop to tenth and eighth positions respectively.

Invasion patterns
The pattern o f invasion for the Transvaal as a whole is uneven with respect to both diversity and abundance of species.The high diversity and abundance areas occur in and around m ajor towns and cities, and in the eastern Transvaal along the foothills and slopes of the Drakensberg escarpment.
The pattern of species diversity for the Transvaal is illustrated in Fig. 23, which gives the distribution of quarter degree squares in which 6 or more roadside and veld invader species were recorded.A similar pattern was obtained for streambank invaders.
The distribution maps of the most important species (Figs 23 -32) also illustrate an uneven invasion pattern.The high abundance areas o f all the species overlap to the greatest extent in the degree squares 2330 (Tzaneen District), 2430 (Pilgrims Rest District) 2530 (Lydenburg District) and 2630 (Carolina District).
Overall the degree square with the greatest abundance and diversity of invader species was 2530 -which includes the towns of Lydenburg, Belfast, M achadodorp, Sabie and Nelspruit.This degree square is also one o f the most variable in the Transvaal with respect to topography, climate and indigenous veld types.It has a high rainfall and has been greatly disturbed.
The western parts of the Transvaal have been subject to much less invasion than the eastern parts.Contributing factors include less disturbance, the absence of large and extensive plantations of invader species, a drier climate and a landscape that offers a smaller range of habitats and niches for invading species.

Sampling
As mentioned in the previous paper (Duggan & Henderson, 1982), the sampling method has its limitations such as the undersampling of certain habitats and the less distinctive species.The method, however, has proved successful and economical in obtaining information that otherwise would have been unobtainable.It has also provided an initial broad survey which can serve as a basis for further intensive work on more restricted areas.
O ne recommendation that we have with regard to the sampling method is that a standard roadside transect length of about 10 km be used within each veld type or sample unit in recording abundance values of roadside and veld species.This would enable all the records obtained to be compared directly and simplify the computation of average and total abundance values.If this recommendation is adopted, the abundance ratings <1 and « 1 will fall away.In addition, the use of many short standard transect lengths instead of one or a few long ones, will also provide more recordings which is statistical ly desirable and would also reveal variation that otherwise would not be apparent.

Important species
This survey has underlined the invasive importance of Acacia dealbata, Melia azedarach, Opuntia ftcus-indica, Acacia mearnsii, Salix babylonica and Populus alba/canescens.All these species, with the exception of Opuntia ficus-indica, are invaders of the streambank habitat and two of them (Salix babylonica and Populus alba/canescens) are virtually restricted to the streambank habitat.Acacia dealbata, A .mearnsii, Melia azedarach and Opuntia ficus-indica are all important invaders of roadside and veld habitats.
Further sampling since the last publication has confirmed the observation that Acacia dealbata is an aggressive and dangerous invader.Acacia mearnsii has proved to be of greater importance than was previously believed, particularly in the moist forested and false grassveld areas.Both these species are either absent or very rarely encountered in the western Transvaal.If seed was made available, however, both species could extend their range into these parts mainly along water-courses and into the moister bushveld areas, e.g. the Waterberg.
Melia azedarach and Opuntia ficus-indica have proved to be the most widespread invader species in the Transvaal as a whole.Both species are highly penetrative and are found far from any plantings.
Several potentially important invaders have been recorded since the last publication.These include Pinus patula, Toona ciliata, Prosopis velutina, Schinus molle and Trichocereus sp.
Pinus patula is an extremely aggressive invader of the moist and cool, forested regions of the Drakensberg escarpment.This species seeds prolifically, colonizing roadsides, road cuttings and other disturbed places, plantations of other tree species and indigenous forest, scrub and grassland.At least one other species of pine is also invasive in these parts.(According to the Sabie regional forestry office, P. taeda and P. elliotii are also naturalized in places.) Toona ciliata is a tree similar in appearance to Melia azedarach and belonging to the same family, Meliaceae.T. ciliata is naturalized in the Barberton, Nelspruit and Duiwelskloof areas and possibly elsewhere.Like Melia azedarach, it has become a roadside weed and has invaded watercourses, bushveld and forest.This species, however, is less drought and frost resistant than Melia azedarach and therefore has a potentially smaller distribution.
Passifiora edulis is another species that has become naturalized in the moist bushveld and forests of the eastern Transvaal.The climbing growth habit of the species, however, made it difficult to record and therefore we have little information on its percentage frequency and abundance.M r M. J. Wells of the Botanical Research Institute (pers.comm.)reports that it is very frequent, particularly in remnant forests that survive between plantations.
Prosopis velutina, Schinus molle and Trichocereus sp. have been recorded in the dry south western Transvaal.A t present these species are very localized but have the potential to spread.
Pyracantha angustifolia, which was recorded in the pilot study area, has been recorded since the last publication.At present this species occurs in small numbers at scattered localities, mainly in the southern grassland areas of the Transvaal.This species, together with the Cotoneaster spp.(C.pannosa recorded in pilot study area) are potentially important invaders in the colder parts of the Transvaal where they have been planted because of their resistance to cold and for their attractive berries.
The berries of these species are well-liked by birds particularly since they are produced in winter when there is very little other food available on the cold highveld.The dispersal of the fruits by birds enables these species, and other suitable fruit-bearing plants such as Ligustrum spp., Lantana camara, Melia azedarach, Morus alba, Psidium guajava and Solanum mauritianum, to spread far afield and to reach inaccessible places from which reinfestation can take place.
Species which have assumed greater importance, as a result of better recording, since the last publication include Acacia melanoxylon, Solanum mauritianum and Eucalyptus spp.All these species are the most abundant in the cool, moist forested regions of the eastern Transvaal.
Solanum mauritianum is a species which we suspect has been undersampled.It was seldom seen except in the eastern Transvaal where it occurred in greater numbers than any other exotic woody invader in the Transvaal.It is particularly bad as an understorey weed in pine plantations, being able to tolerate considerable shading.
A t present Acacia melanoxylon is fairly localized but has a potentially wide distribution coinciding with the cool, moist parts of the Transvaal.Small plants were often seen far from any planting, particularly along roadsides and in clearings in plantations and forests.
The Eucalyptus spp.are, on the whole, relatively unaggressive species.In the forestry areas of the eastern Transvaal, however, they assume their greatest importance.Several species seed themsel ves, of which the most prolific is E. grandis.

Correlation o f invasion with environmental factors
Overall the warmer tropical forest, bush and savanna veld types were generally more subject to invasion by exotic, woody species than were the cooler, temperate grassveld and false grassveld types.
In the streambank habitat, 36 of the 40 species recorded occurred in the tropical veld types compared with 20 species in the tem perate veld types.
In roadside and veld habitats, 53 of the 60 species occurred in the tropical veld types compared with 42 species in the tem perate veld types.
Further sampling since the 1982 publication has allowed for a more in-depth study of the correlation between invasion and veld type.The results have shown that considerable differences exist between the inland tropical forest types, tropical bush and savanna types, false grassveld and pure grassveld types.
If the degree of invasion is gauged by the abundance and species diversity of invaders per sampling unit and water course recording then the inland tropical forest types show the greatest invasion in streambanks, roadside and veld habitats.The false grassveld types take second place in roadside and veld habitats followed by the tropical bush and savanna and pure grassveld types.
Although the tropical bush and savanna types have the most recorded invader species, they occupy the largest area and account for 63% of the samples.Invasion of the bushveld is also very patchy -with large tracts of dry veld relatively undisturbed while other parts again have been severely disturbed, for example the moist bushveld, Acocks 10, in the eastern Transvaal.
In the streambank habitat there is little difference in the abundance and species diversity values for the bushveld and false grassveld types.The pure grassveld types have the least number of exotic, woody invader species, and the lowest species abundance and diversity values.

Prospects fo r the future
Invasion is expected to increase in all parts of the Transvaal.The greatest expansion is liable to occur in the moist and warm areas which are favourable for plant growth -namely the forests, moist bushveld and warmer false grassveld areas (south eastern Transvaal).
The stability of the forests has been so weakened by clearing for afforestation that they have been made exceedingly vulnerable to invasion by a wide spectrum of exotic invaders.The moist bushveld has been disturbed to a lesser extent, but the climatic conditions are favourable for the growth of many exotic species.The false grassveld is favourable for fewer exotic species, but offers much less competi tion to invading species because of its relatively few indigenous woody species.
The very dry bushveld areas of the Transvaal are probably in the least danger from invading exotics.The combination of unfavourable climatic condi tions and a well-adapted indigenous woody compo nent act as barriers to exotic species.Rivers and streams, however, are more vulnerable to invasion and permit the entry of moisture-loving species into an otherwise inhospitable environment.Although severe frosts help to limit the spread of many potential invader species in the cold grassveld areas, these veld types are vulnerable because of their lack of a natural woody component.Frosthardy, exotic species are therefore able to invade unoccupied niches with relatively little competition.
The ultimate prospects for the future if action is not taken against the invading species are very bleak indeed: 1.The few remaining relics of natural forest are liable to be almost totally replaced by exotics.
2. In the moist bushveld exotic species will replace many of the indigenous species, altering the basic appearance and character of the veld.
3. In the dry bushveld one can expect fewer exotic invaders but one or a few dominant species which will take over large tracts of land at the expense of the indigenous species.
4. The grassveld and false grassveld areas will probably undergo the greatest change in appearance by acquiring a woody element that was either totally lacking o r poorly represented before.
5. The indigenous stream bank vegetation throughout the Transvaal is in grave danger of being almost totally replaced by exotic, woody invaders.

R E F E R E N C E S
A c o c k s .J . P. H ., 1975, V eld

H
e n d e r s o n , L .& D u g g a n .K. J ., 1981.T h e m inim um effective sam pling level fo r the d eterm ination o f the m ajor exotic, w oody plant invaders, th e frequency o f occurrence and relative im portance and m ajo r invasion trends in th e central T ransvaal.U npublished.H u b b a r d .C. S ., 1926.A review o f th e species o f Populus introduced into South A frica. S. A fr. J. Sci.23: 3 4 0 -365.P a l g r a v e .K. C ., 1977.Trees o f southern Africa.C ape Town: Stm ik, P u r s e g l o v e , J .W ., 1968.Tropical crops, dicotyledons.V ols 1 & 2. L ondon: Longm ans.R o ss J, H ., 1975.T he naturalized and cultivated exotic Acacia species in South A frica.Bothalia 11: 4 6 3 -4 7 0 .

Fig. 2 .
Fig. 2.-The four broad veld type categories in the Transvaal.
Degree squares In which 6 or more species occurred.

TABLE 2 .
-Veld type coverage, excluding pilot study area

TABLE 4 .
-The number and abundance o f invader species recorded in streambank samples

TABLE 5 .-The number and abundance of invader species recorded in roadside and veld samples
type groupings and in the whole survey area are shown in Figs 13 -19.The most frequent and widespread invaders are Opuntia ficus-indica, Melia azedarach and Eucalyptus spp.In the temperate veld types Eucalyptus spp., Opuntia ficus-indica, Acacia dealbata, A .mearnsii, Melia azedarach and A. decurrens are the most widespread species.Melia azedarach and Opuntia ficus-indica are by far the most widespread invaders in the tropical veld types, particularly the bushveld which occupies the greatest area.
types o f South A fric a .Mem. hot.Surv.-S.Afr.40: 1 -1 2 8 .B a i l e y , L. H ., 1963.The standard cyclopedia o f horticulture.a i l e y , L. H. & B a i l e y .E. Z ., 1976.Hortus third.New Y ork, B ape Town: D epartm ent o f N ature and E nvironm ental C onservation o f th e C ape Provincial A dm inistration.W e l l s , M. J ., D u g g a n .K. J .&H e n d e r s o n , L ., 1980.W oody plant invaders o f th e central T ransvaal.Proceedings o f the Lantana camara L.Eucalyptus cinerea F. Muell.exBenth.Argyle appleSolanum mauritianum Scop.