The distribution patterns and impacts of the approximately 750 tree and 8000 herbaceous plant species that have invaded over 10 million hectares of land in South Africa (Department of Environmental Affairs and Tourism [DEAT] 2006) have been extensively documented (Henderson 2001; Joubert 2009; Macdonald, Kruger & Ferrar 1986; Macdonald et al. 2003; Richardson & Van Wilgen 2004; Wilson et al. 2014). Much less has been published on the introduced fauna, although previous reviews have listed alien animals within some specific habitats or regions or within specific taxonomic groups. These include listings of the introduced fauna of South African aquatic ecosystems (De Moor & Bruton 1988) and South African National Parks (Spear et al. 2011) and several reviews documenting progressively increasing numbers of introductions into the marine environment (Griffiths et al. 1992; Griffiths, Robinson & Mead 2009; Mead et al. 2011a, 2011b; Robinson et al. 2016). Introduction pathways for South African vertebrates, invertebrates and plants have been compared and temporal patterns of invasion via the defined pathways examined (Faulkner et al. 2016). In taxonomic terms, the alien vertebrates have received a fair amount of research interest (Brooke, Lloyd & De Villiers 1986), with various reviews devoted to introduced mammals (Skead 2011), birds (Dean 2000), reptiles (van Wilgen et al. 2010) and fishes (De Moor & Bruton 1988; Ellender & Weyl 2014; Griffiths, Day & Picker 2015). Less attention has been devoted to invertebrate groups, but there have been some attempts to list the introduced species within specific taxa such as terrestrial molluscs (Herbert 2010) and earthworms (Plisko 2010), or functional groups such as biological control agents (Klein 2011). Reviews of the pests of cultivated plants (Annecke & Moran 1982; Prinsloo & Uys 2015; Visser 2009) also incorporate many species that are introduced. However, until recently no attempt had been made to produce an inventory of the entire regional alien fauna. The first such listing appeared in the semi-popular book by Picker and Griffiths (2011). The draft National Environmental Management Biodiversity Act: Alien and Invasive species list (Republic of South Africa 2014, Notice 3: National Lists of Invasive Species, lists 3–10) also provides incomplete listings of various groups of faunal invasives, which in fact also include numerous species that have not yet invaded any part of South Africa. For example, list 5 of the Invasive List (Notice 3) mentions 35 alien reptile species – none which has been recorded as an established alien, but fails to list the single species which is an established invader (Brahminy blind snake, Ramphotyphlops braminus)!

The aims of this paper are to analyse the most comprehensive available listing of alien animal species of South Africa, that in Picker and Griffiths (2011), in terms of invaded habitat, taxonomic composition of the fauna, proportions of intentionally and unintentionally introduced species, invasion status, historical rate of accumulation of species, geographical origins of the fauna and spatial distribution across the region. These analyses provide baselines against which to measure future rates and patterns of faunal invasion in the region and also allow for comparisons of invasion rates and patterns between the various taxa and across the different habitats examined, as well as between South Africa and other regions. The species listing provided in the Appendix also documents which species have (or have not) in fact been reported in the region and can provide a basis for updated legislation, be used to compare listing from other regions, etc.

This analysis is based entirely on the listing of 571 alien species given in Picker and Griffiths (2011). Species documented after 2011 are excluded, as are translocated species and cryptogenic marine species (as listed by Mead et al. 2011a, 2011b). Data from the Appendix of Picker and Griffiths (2011) were used for the analysis of taxonomic composition, geographical origin and species richness of the South Africa alien fauna. All 571 species listed in the Appendix of Picker and Griffiths (2011) were also used for the analysis of modes of introduction and for the species accumulation curves. The original sources of data from which the listing in Picker and Griffiths (2011) were derived included various earlier reviews (Annecke & Moran 1982; Coates 1970; Herbert 2010; Heyns 1971; Klein 2011; Mead et al. 2011b; Millar 1994; Plisko 2010; Visser 2009) and scattered taxonomic papers as cited under the individual species entries in Picker and Griffiths (2011). In cases where only a subset of the data could be used (viz. Figures 2 and 3), the number of species used and source of data are given in the figure legends.

When examining the invasion status of species, we followed the framework proposed by Blackburn et al. (2011) and applied these categories to the 220 species featured in Picker and Griffiths (2011) for which we deemed the distribution sufficiently well known to be evaluated. Species were considered to have remained restricted to the site of entry (Category C3) if their distribution was mapped as a point entry (or two points in the case of multiple known sites of introduction, e.g. to two harbours). Species were considered to have ‘spread significantly’ (Category D) if they had expanded < 100 km from site of entry and to be fully invasive (Category E) if they had spread > 100 km or occupied numerous sites (e.g. several widespread cities). In analysing geographical origin of the alien fauna, we used the seven-continent model. Where the native range of a species spanned two continents, each continent was scored 0.5 for that species, and where the home range spanned three continents, each received a score of 0.33. This accounts for the fractional scores shown in terms of numbers of species originating from the various continents. Assigning a full score to a species that occurs naturally on two continents would effectively bias the scores, as it would assume that the species had been introduced twice, once from each source continent and thus give it a double weighting. For parsimony, the fractional scores assume each species has been introduced from a single continent within its natural range.

When defining the mode of introduction as either intentional or unintentional, we consider species to be unintentional introductions only if they arrived inadvertently into the region, typically through association with crops, on ships’ hulls and ballast water, etc. Species are considered as intentional introductions both if purposely introduced directly into the wild, for example, as ornamentals, biological control agents or targets for hunting or fishing, or if introduced intentionally into captivity/aquaculture, from where they subsequently escaped and established in the wild. The mode of introduction was assessed for species in the Appendix of Picker and Griffiths (2011) using original sources of data as listed above.

Dates of introduction were bracketed by decade. While dates of introduction of vertebrates are generally quite accurately known, the date of discovery for the more cryptic invertebrates commonly postdates the true date of introduction, sometimes by decades. Species accumulation curves were plotted separately in Excel 2013 for invertebrates, for vertebrates and for all taxa combined. The slopes of the fitted curve for all taxa combined were calculated separately for the period 1750–1880, which was a period of limited colonisation, and for the period 1880–2000, which was a period of greater international trade initiated by the discovery of diamonds in Kimberley (1871) and gold in the Witwatersrand (1884), the Anglo-Boer wars (1880–1881; 1899–1902) and by immigration of Indian labourers in 1860 for the growing sugarcane industry in KwaZulu-Natal (Christopher 1994). The 1880–2000 curve was extrapolated in Excel 2013.

A graduated map showing the species richness of the alien fauna per half-degree square in South Africa was generated from Picker and Griffiths (2011), using those 242 species for which adequate distributional data were presented. Hardcopies of individual distribution maps for each of these species were scanned and superimposed over a visual half-degree grid map of matched scale to generate a visual half-degree grid map showing the number of alien species in each grid cell. Marine species were included in cells bordering the coastline, meaning that cells falling over both land and sea contain both terrestrial and marine species. From this visual grid, a half-degree cover was constructed for South Africa, and the species count data were transferred from the Excel grid to the grid feature class in ArcGIS 10.3.1. The grid was then symbolised using the ‘Graduated Colour’ symbology option with seven classes. This was subsequently converted to greyscale in Corel Photo-Paint X3.

The 571 alien animal species in South African analysed here certainly represent an underestimation of the true number of alien animals present in the region. Newly introduced species are constantly invading the region and several such new invasions are reported each year. In addition, long-established historical invasions are also being revealed as a result of new ecological surveys or taxonomic revisions, for example, Mead et al. (2011a) uncovered several marine invasions that were associated with dry ballast or wooden vessels and had thus probably remained undiscovered for a century or more. Species may of course also be lost from the fauna, if local populations that were recorded as introduced in South Africa in the historical literature become locally extinct. Such losses can be difficult to detect, but see documentation of the recent local extinction of an alien sea urchin by Mabin, Wilson and Robinson (2015).

While the number of alien vertebrates (41; 7% of all alien species) is considered to be relatively well documented (Picker and Griffiths 2011), the tally of alien invertebrates is certainly a considerable underestimation linked to South Africa’s relatively low ‘country development status’ (McGeoch et al. 2010), which results in resource limitation constraining the state of knowledge of biological invasions. This is best illustrated by comparing the inventory presented here with that produced by the DAISIE Project (Delivering Alien Invasive Species Inventories for Europe), one of the most comprehensive and detailed inventories of any alien fauna. The DAISIE Project records 2740 terrestrial invertebrate species in Europe compared to only 441 species for South Africa (DAISIE 2008). A similar situation exists in the marine environment. Although the number of recorded alien animals in the South African marine environment has increased dramatically in recent years, from 22 species reported by Griffiths et al. (2009) to 79 species recorded here, this is still far less than the 986 alien marine species recorded for Europe (Hulme et al. 2009). This is because many marine habitats (e.g. soft sediments of harbours) and taxa (e.g. Nematoda, Copepoda, Ostracoda and many parasitic taxa) remain poorly sampled for introduced species in South Africa and surely contain many more alien species. Additional factors, such as different duration and intensity of introduction histories and differences in the diversity of available habitats, may also contribute towards these regional differences.

The differences in the composition of the introduced faunas among major habitat types reflect (1) natural differences in biodiversity and faunal composition of terrestrial, marine and freshwater faunas and (2) differences between vectors that have transported these species. The dominance of the alien terrestrial fauna by insects is a reflection not only of the enormous natural diversity and species richness of this group in terrestrial systems but also of their importance as crop pests and their frequent use as biological control agents. Similarly, ascidians, echinoderms, bryozoans and cnidarians are all overwhelmingly marine groups and hence feature more prominently in the marine listing. However, it is of interest to note that no marine fishes have been introduced to South Africa, and no introduced marine nematodes have been reported, despite their high diversity in marine habitats. The latter is probably an artefact of poor sampling and lack of taxonomic expertise in marine nematodes, as numerous invasive nematodes have been reported from other regions, such as Europe (DAISIE). The high diversity of fishes among freshwater introductions stems from their frequent intentional introduction as sport-fishing targets, ornamental and forage species, or as biological control agents. This pattern was not paralleled in marine habitats, where sport-fishing targets are common among the indigenous fish fauna and no intentional alien fish introductions have taken place (other than of species that remain restricted within aquaculture facilities).

Our analysis of degree of establishment of species within the alien fauna suggests that 13% of analysed species have failed to spread from their sites of introduction, 16% have spread significantly (< 100 km) and 71% are fully invasive. However, as this analysis is based on species featured as full entries in the review by Picker and Griffiths (2011), it is (apart from the vertebrates) biased in favour of ‘important’ invasions, which also tend to be those that have spread widely. In addition, species with wider distributions are more likely to have been discovered, and their distribution is more likely to have been mapped when compared to those that remain localised. Thus, we suspect that that species in the early stages of invasion in fact make up a larger component of the fauna than this analysis suggests.

The initial rate of accumulation of the South African alien fauna was slow. During this period of early colonialism, the first Dutch settlers gradually expanded settlements into the interior and eastern parts of South Africa, with the last frontiers of the Northern and Eastern Cape and northern KwaZulu-Natal being breached by colonialists around 1900. During the last decades of the 19th century (1880 onwards), the numbers of alien species increased steeply, with the slope of the accumulation curve increasing from 0.16 to 0.65. This coincided with the discovery of diamonds in Kimberley (ca. 1871) and then gold on the Witwatersrand (1884) – both events galvanising urbanisation, international travel and trade links, which led to increased immigration rates (Christopher 1994; Deacon 1986). The dramatic increase in trade because of the mining industry was also associated with the industrialisation of ports. All these factors contributed to the steep increase in the number of alien animals since 1880. It should be noted that while these figures were derived from species whose invasion history is well known, many (invertebrate) species are only being discovered long after their first date of introduction, and some species may have multiple dates of introduction. The pattern of invasion by alien species in Europe differs in many respects from that in South Africa. In Europe, there was no noticeable increase in the number of new species between the period 1951–1970 and the period 1971–1990 in marine and freshwater habitats, apart for marine invertebrates, which had a steep increase during the period 1951–2007 (Hulme et al. 2009). No new alien birds or mammals colonised Europe after 1951, comparable with the situation in South Africa, where few new alien vertebrates colonised after 1930. However, the number of escapee species appears to be increasing with time (Faulkner et al. 2016). In Europe there was a fairly steep increase in the number of new alien insects after 1951 and a flatter curve for new alien non-insect arthropods. In South Africa the curve was linear and steep after 1890. Roques et al. (2009) estimated that 60% of the alien insect fauna of Europe only established in the last 50 years. A similar trend is evident for South Africa (Figure 2), where approximately 50% of invertebrate introductions occurred prior to 1960. The major pathways for the introduction of South African alien invertebrates (besides deliberate release as biocontrol agents) were as contaminants and stowaways with the number of released biocontrol agents increasing sharply after 1970 (Faulkner et al. 2016). In contrast, the introduction of invertebrates as stowaways and contaminants was gradual in the 1900s and accelerated in the 2000s (Faulkner et al. 2016). The curve may thus not have been influenced to a large degree by biocontrol agents, as few had been introduced at the beginning of the 20th century and comprised only 21% of the 530 alien invertebrates considered here. See also Faulkner et al. (2017) for a discussion of introductions from other African countries.

The highest concentration of marine alien species in South Africa occurs in the vicinity of major urban areas, especially ports, which were the entry point for many groups of alien invaders, including plants (Deacon 1986). For animals, the highest density of terrestrial alien species (130–162 species/half-degree grid square) occurs in the metropolitan area of Cape Town. This area is the oldest port in South Africa, and its Mediterranean-type climate also provides a suitable eco-climatic match for alien fauna from the temperate regions of Europe and North America, with which the developing colony conducted most of its trade (Tribe & Richardson 1994). Interestingly, some alien vertebrates, such as the grey squirrel and European chaffinch, both introduced to Cape Town in the 1890s, have not managed to substantially extend their ranges further into the fynbos biome in over 100 years.

The least invaded region, with less than 50 introduced species per half-degree cell, is restricted to the arid, low population density areas of the Northern Cape in the interior of the country. Many of the alien species here are widely distributed species associated with human habitation, such as mice, house sparrows, cockroaches, booklice and pests of domestic stock and stored products (see individual maps in Picker & Griffiths 2011). The low densities of alien animals in the semi-arid interior may also be related to low habitat diversity and poor ecoclimatic matching of that area. Most of the country has 50–100 introduced animal species per half-degree grid square. Areas with more than 100 species comprise the urban centres surrounding Johannesburg, Pretoria and Durban, plus a broad coastal swath running from Port Elizabeth to Cape Town. Cape Town and its close surrounds stand out as the only region with over 150 species (largely invertebrates) and represent by far the most heavily invaded area of the country. This is similar to the situation in Europe, where both alien plants and insects were found to be associated most closely with urban habitats (parks and gardens), followed by cultivated lands (Pyšek et al. 2010). Sampling intensity may also play some role here, as urban habitats also tend to be adjacent to research institutions and this may result in their being more intensively surveyed than more remote regions of the country.