At present, it is unclear whether environmental factors or species-specific characteristics are more important in determining the tendency of a species to become invasive. This problem has led to substantial research efforts over the past 15 years to try to identify those factors that might result in a tree becoming invasive (e.g. see Richardson 1998). This has been part of a wider effort to understand the general basis of invasiveness in plants (Rejmanek 2001). For example, one characteristic that has been frequently used to select trees used in forestry and agroforestry is growth rate, particularly in relation to poor soils. This has often resulted in trees being used that seed heavily and that can survive under a wide range of environmental conditions. Unfortunately, these very qualities can lead to a tree becoming invasive. If more such traits can be identified, it would both help to guide future searches for new forestry species and aid weed risk assessments for new and recently introduced species.
Richardson et al. (1994) were among the first to draw attention to the fact that, of the many pines introduced for forestry, some naturalize but remain largely under their own canopy, while others naturalize and then spread widely. In this review, 68 forestry species were listed in Binggeli (1996) as “possibly naturalized/some degree of invasiveness” but the extent of spread was either limited or unknown. Corroborating evidence that these trees were invasive was obtained for approximately one-third of these species.
There are many cases in which a species known to be highly invasive in at least one country has not shown the same trait when introduced elsewhere (or at least has not been reported as such), e.g. Acacia melanoxylon, A. nilotica and Prosopis juliflora (see Figure 4). It is well known that most invasive tree species have taken something like 50 years or more to attain that state (Hughes 1995). Time is thus a very important variable that needs to be taken into account when recording whether or not a species is invasive.
For those species classed as “naturalized” but not invasive, is it then just a matter of time before they become invasive, or will they remain naturalized? Richardson et al. (1994) cited some examples of naturalized pines that have subsequently become invasive whereas others have remained merely naturalized. If “invasiveness” is independent of time, then invasiveness could be linked to some species characteristic. If, on the other hand, the occurrence of invasiveness is strongly linked to time, then the start of an invasion might indicate some change in dispersal ability or colonization opportunities in the area of introduction. These aspects have received much attention from invasion ecologists and there is now a substantial body of scientific literature on the subject (see below). However, because of the absence of information on the time element in the dataset, identifying those tree species that are truly invasive versus those that are not is fraught with difficulties. This problem is compounded by the incomplete information available for so many of the records and the subjectivity of much of the information.
It would be of more practical significance to examine those species that have been recorded as highly invasive in at least one area, in order to determine whether or not some form of control or management by local communities is contributing to noninvasiveness in other areas.
In the following section, the results of previous studies on the “basis” of invasiveness are summarized in order to provide some background on the subject. Some aspects of the reported invasiveness of the genera Acacia and Prosopis are then evaluated to illustrate the importance of understanding the available data on invasiveness.
Trees have been identified as the most successful group of plant invaders (compared to other plant growth forms) in natural or semi-natural habitats (Cronk and Fuller 1995). Broadly speaking, biological factors that have been suggested as contributing to invasiveness fall into the following categories:
1. Species characteristics. Here, factors such as breeding systems and seed ecology may be important. Most plants are outbreeders with some capacity for self-fertilization and this seems to be the case for invasive plants (Cronk and Fuller 1995). For introduced trees, outbreeding may not be a constraint as many species have been introduced in large numbers, e.g. Pinus radiata. Furthermore, many of the introduced trees produce large quantities of seed (e.g. pines, eucalypts, woody legumes) and Richardson et al. (1994), through their work on pines, have suggested that propagule pressure is an important factor contributing to invasiveness.
2. Seed dispersal mechanisms. Cronk and Fuller (1995) collated information that showed that the dispersal of many invasive plants is facilitated by vertebrates. In New Zealand, over half of the woody invaders have fruits that are adapted for dispersal by birds. Similarly, in South Africa, the most successful Acacia species are those dispersed by birds.
3. Absence of herbivores/competitors/pests. There are many studies that show that trees in their native ranges are attacked by a wide variety of coevolved micro-organisms and herbivores (both insects and vertebrates). However, when trees are introduced they are often relieved of this pressure, as the herbivores etc. in the area of introduction are not coevolved and are usually less diverse. Such observations underpin the concept of “classical” biological control for weeds.
4. Ecosystem disturbance. Most invasions by forestry trees occur in semi-natural ecosystems (see Section 4) and “disturbance” has frequently been identified as an essential prerequisite for invasion by an introduced species. Further support for this idea comes from Hughes and Styles (1989) who cited examples from South America of introduced trees that were unable to invade natural ecosystems.
Acacia and Prosopis are two genera in the family Leguminosae, tribe Mimosoidae. Among other characteristics described more fully in CAB International (2000), many species in these genera have bisexual flowers, fix nitrogen through rhizobia, have the ability to regenerate rapidly following the loss of plant parts and can exist in a range of growth forms (e.g. shrub, small tree, large tree etc.), depending on their environment. Within both genera, there are species that can survive extremes of temperature, aridity, shade, frost, weeds, fire, salt, waterlogging and browsing. Their main pollinators are insects and birds. The ability of these trees to tolerate difficult environments and to provide useful products (e.g. fuelwood, timber and nutritious pods that can feed cattle and sheep in savannah-type grazing systems), makes them attractive not only for forestry and agroforestry, but also for soil improvement or stabilization. As a result, they have been introduced to a large number of countries (see Figure 4). However, several species (e.g. Acacia nilotica, A. melanoxylon and P. juliflora) have subsequently become invasive.
Many of the positive attributes listed above could also contribute to invasiveness. For example, vigorous growth, though useful for coppicing systems, can be detrimental if the bushes form unmanageable thickets, particularly if the species concerned are also thorny. However, even the most notoriously invasive Acacia and Prosopis species have only been reported as invasive in a fraction of the areas in which they have been introduced (see Figure 4 and Section 3.4.2).
A correspondence analysis (based on data extracted from CAB International (2000)) was performed on the ability of various Acacia and Prosopis species to tolerate climatic factors (drought, rainfall, maximum temperature etc.) and to regenerate rapidly. The summarized tolerance scores for axes 1 (drought tolerance/rapid regeneration) and axes 2 (frost tolerance) were then used in a principal components analysis (PCA) along with the latitudinal, altitudinal and climatic range of these species. The hypothesis being tested was that “invasive” species would be more tolerant of various extremes, or could be separated from “noninvasives” by some other characteristic. Figure 10 shows the result of this analysis.
According to the PCA, there was considerable overlap in the characteristics and environmental tolerance ranges of both invasive and noninvasive Acacia and Prosopis species. The distribution of the points summarizing the characteristics of invasive and noninvasive species also broadly overlapped. A more comprehensive dataset that included more biological characters (e.g. seed size, seed dispersal, reproductive system, age at first reproduction, etc.) would have allowed a more detailed analysis to be conducted (similar to that undertaken for the genus Pinus by Richardson et al. 1994). The two groups might have appeared more distinct if a broader range of characteristics could have been considered.
The basis on which such an analysis could be performed, however, may be questionable. For example, at what point should a species be defined as “invasive”, considering that it may have been introduced into 20 countries but become invasive in only one? Rather than using average data on a species’ biological characteristics, it may be more appropriate to use local data drawn specifically from sites of invasion versus noninvasion. However, at present, such data appear to be largely unavailable. Furthermore, for such an analysis to be valid, reporting of invasion must be equally reliable for all species. If such data could be provided, then this type of analysis might be a viable tool for better understanding invasiveness, and even for performing risk assessments on the introduction of new species. It is important to remember, however, that management approaches may differ in different areas, and that they, too may affect the status of a species (not invasive, naturalized, invasive, etc.) in different parts of its exotic range.
Acacia species are represented by circles and Prosopis by triangles. Invasive species are represented by open symbols and noninvasives by closed symbols. Eigenvalue axis 1 = 0.297. Eigenvalue axis 2 = 0.250. The first two axes explained 54.7 percent of the variance of the data on species attributes.
How can information from invasiveness studies be used? Results from the international program on the assessment of invasives by the Scientific Committee for Problems of the Environment (SCOPE) showed that many biological attributes (e.g. life history, taxonomic status and genetic constitution) were poor indicators of invasiveness. Nevertheless, some of these characters (in combination with other factors) are being used in risk assessments (see Section 6).