Impacts of Argentine ants on invertebrate communities with below-ground consequences
The Argentine ant, Linepithema humile (Mayr), is an invasive species that has been associated with various negative impacts in native communities around the world. These impacts, as for other invasive ants, are principally towards native ant species, and impacts on below-ground processes such as decomposition remain largely unexplored. We investigated the relationship between Argentine ants and invertebrate fauna, litter decomposition and soil microbial activity between paired invaded and uninvaded sites at two locations in Auckland, New Zealand, where there has been no research to date on their impacts. We examined the diversity and composition of invertebrate and microorganisms communities, and differences in soil and litter components. The composition of invertebrates (Order-level, ant and beetle species) was different between invaded and uninvaded sites, with fewer ants, isopods, amphipods, and fungus-feeding beetles at the invaded sites, whereas Collembola were more abundant at the invaded sites. There were significant differences in soil chemistry, including higher carbon and nitrogen microbial biomass at uninvaded sites. Several litter components were significantly different for Macropiper excelsum. The fibre content of litter was higher, and key nutrients (e.g. nitrogen) were lower, at invaded sites, indicating less breakdown of litter at invaded sites. A greater knowledge of the history of invasion at a site would clarify variation in the impacts of Argentine ants, but their persistence in the ground litter layer may have long-term implications for soil and plant health in native ecosystems.
KeywordsLinepithema humile Litter decomposition Microbial biomass Collembola Beetle communities
Biological invasions by exotic social insects pose serious threats to native species and ecosystems worldwide (Holway et al. 2002; Moller 1996; Williams 1994). Most studies on invasive ant impacts focus on their direct effects, such as their ability to reduce native ant species richness and abundance, modify competitive dominance hierarchies, and subsequent co-occurrence patterns (Hoffmann et al. 1999; Holway and Suarez 2006; Gotelli and Arnett 2000; Sanders et al. 2003). Ants are also known to modify invertebrate species composition in communities they invade, with some species and functional groups reducing in density to the extent that they are undetectable, while some species show substantial increases in abundance in the presence of invasive ants (Hoffmann et al. 1999; Human and Gordon 1997; Morrison 2002).
Several studies have examined indirect effects (sensu Wootton 1994) of invasive ants on pollination and seed dispersal mutualisms (Blancafort and Gomez 2005; Christian 2001; Traveset and Richardson 2006). However indirect effects (positive or negative) of ant-mediated impacts on ecosystem function are poorly understood. In one of the few examples, yellow crazy ants, Anoplolepisgracilipes, have been shown to indirectly increase seedling recruitment, change plant composition and slow litter breakdown on Christmas Island through removal of the island’s dominant omnivore, the red land crab (Gecarcoidea natalis) (Green et al. 1999, 2008; O’Dowd et al. 2003). The nest building and foraging activities of the red imported fire ant (Solenopsis invicta) affect soil properties and enhance plant growth though the increase of NH4+ (Lafleur et al. 2005). More recently, Dunham and Mikheyev (2010) have shown that the little fire ant, Wasmannia auropunctata, affects litter decomposition and soil nutrient cycling via a top-down trophic cascade. Some native ants, particularly mound-building ants, are also known to have impacts on soil parameters, such as bulk density, microbial biomass, nutrient concentrations and vertical distribution of nutrients in the nests (Andersen and Sparling 1997; Dóstal et al. 2005 Lavelle et al. 1995; Wardle et al. 2011; Zelikova et al. 2011), but the link between the invasion of alien ants above ground to processes below ground has been largely unexplored (Cammeraat and Risch 2008; Kenis et al. 2009).
The Argentine ant is native to South America (Suarez et al. 2001) and is now invasive in a number of habitats and countries; it typically invades areas of high light availability, such as shrubland and disturbed/edge forest ecosystems (Cole et al. 1992; de Kock and Giliomee 1989; Holway 1998; Suarez et al. 1998; Thomas and Holway 2005; Walters 2006; Ward and Harris 2005). Although Argentine ants have been the subject of much recent study around the globe (Hartley et al. 2006; Roura-Pascual et al. 2004, 2011; Sanders et al. 2003; Thomas and Holway 2005; Walters 2006), much of this work has focused on the interaction with native ant species in continental systems (Holway et al. 2002; Sanders and Suarez 2011). Reduced diversity and biomass of invertebrates and vertebrates as a result of Argentine ant invasion are also evident (Human and Gordon 1997; Krushelnycky and Gillespie 2010; Suarez and Case 2002; Suarez et al. 2005), and these effects can cascade across ecosystems and disrupt processes, such as pollination and seed dispersal (Christian 2001; Blancafort and Gomez 2005; Sanders and Suarez 2011). There is no research describing changes in soil chemistry and litter decomposition where Argentine ants have invaded (Sanders and Suarez 2011).
Argentine ants were found in Auckland, New Zealand in 1990 (Green 1990), and although they remain concentrated in the Auckland and Northland regions, they are increasingly being found in many other urban centres in New Zealand (Ward et al. 2005, 2010). There is no published research on the impacts of Argentine ants in native ecosystems in New Zealand. New Zealand ecosystems have evolved without the influence of an abundant and diverse ant fauna (n = 11 species), most of which are present in low densities and are cryptic litter-dwelling forest species and unlikely to compete with Argentine ants (Andersen 1997; Don 2007; Ward 2009). Furthermore, New Zealand native ants do not have key roles in ecosystem functioning (Don 2007; Ward 2009), such as seed dispersal and soil engineering, as they do in other ecosystems, where much of the previous Argentine ant impact research has occurred (Holway et al. 2002; Rowles and O’Dowd 2009; Sanders and Suarez 2011). Thus, New Zealand offers an opportunity to study the effects of Argentine ants on ecosystem function in a system that lacks a large and diverse native ant fauna (Krushelnycky and Gillespie 2010; Ward 2009).
The goal of our research was to compare the diversity and composition of invertebrate and microorganism communities, chemical components of the soil and litter, and litter decomposition at sites invaded by Argentine ants with uninvaded sites. We predicted that the indirect effects of Argentine ants on ecosystem processes, such as litter decomposition, would be mediated via disassembly of native litter and soil fauna communities rather than disassembly of native ant communities.
We used four sites, each 30 × 30 m in size, within urban reserves of Auckland city, New Zealand. We used a paired site design, where each of two invaded sites was paired with a similar uninvaded site. Argentine ants have been present at both invaded sites since at least 2002. Each site was divided into 5 × 5 m plots (36 plots per site), where all environmental, invertebrate, soil and litter sampling was undertaken.
The first of the paired sites was on the North Shore, with two sites approximately 4 km apart: Rotary Reserve (Lake Rd, Northcote, 36°48.32′S, 174°44.68′E), which has been invaded by Argentine ants, and Sylvan Park Reserve (Lake Pupuke, Milford, 36°46.5′S, 174°45.9′E), where Argentine ants are absent. These paired sites are both on volcanic soils on balsaltic cones, within the same environmental classification (Land Environment of New Zealand A6.1b; Leathwick et al. 2003). The second location was Waikowhai Reserve, one of the largest remnants of native habitat (c. 12 ha) in Auckland City. Within this large reserve we chose two sites (36°56.0′S, 174°44.1′E) approximately 500 m apart: one invaded site and one uninvaded. These paired sites are both on Waitemata strata within the same environmental classification (Land Environment of New Zealand A7.1a; Leathwick et al. 2003). All four sites are within the same Ecological District classification (Tāmaki Ecological District; McEwen 1987). The vegetation at all sites was similar, consisting primarily of broadleaf trees and shrubs, predominantly Melicytus ramiflorus, Macropiper excelsum, Corynocarpus laevigatus, Cordyline australis, and several Coprosma species.
From the centre of each 5 × 5 m plot, we measured soil temperature (oC); soil moisture (% saturation of the top 10 cm of soil); % canopy cover using a densitometer; and % ground vegetation cover of the plot. From ten randomly chosen plots we also recorded the number of seedlings (<15 cm in height) from a 49 cm central radius (0.75 m2) (National Vegetation Survey methodology; Allen 1993), and compared litter biomass (dry mass per cm2) by collecting and drying (at 60 °C for 72 h) all the litter within a 20 × 20 cm quadrat from each of the ten randomly chosen plots at each site.
We conducted this invertebrate sampling between 23 January and 10 February 2006, the austral summer period. We examined invertebrates under the microscope and sorted specimens to Order, and the beetles and ants were then further identified to the lowest taxonomic level (species where possible). All specimens are held at Landcare Research in Auckland.
Within each site, 20 pitfall traps were set in a 4 × 5 grid to sample ground invertebrates (one pitfall per plot; centre 20 plots within the 36-plot site). Each trap consisted of a 100-mm-deep plastic container with a diameter of 105 mm, sunk vertically in the ground, containing a 100 ml mix of ethanol and mono-propylene glycol (70:30). A lid was secured a few centimetres above the trap to minimise debris entering the trap. Traps were left open for 8 days. Due to a limited number of Berlese funnels available, we randomly chose ten plots within each site for litter invertebrate sampling. In each plot, all litter within a 1 × 1 m quadrat was collected and placed in a Berlese funnel for 48 h in the laboratory.
We collected eight soil samples from randomly chosen plots per site using a soil corer (10 cm depth, 25 mm diameter) in summer (30 January) and autumn (24 April). We pooled the eight samples into two stratified samples per site and then analysed these samples for the following chemical variables: water content (% dry mass), total carbon (C; %), total nitrogen (N; %), NO3-N (mg−1 kg−1), NH4-N (mg−1 kg−1), mineralisable N (mg−1 kg−1), and Olsen P (mg−1 kg−1). We also measured microbial biomass C (mg−1 kg−1), microbial biomass N (mg−1 kg−1), and basal respiration (μgC−1 g−1 h−1) to estimate microbial biomass and activity in the soil (Solaiman 2007). Laboratory analyses were done within Landcare Research laboratories, and because each of these components is analysed a different way, with different laboratory steps, detailed methods can be found at Landcare Research (2008).
Litter bag technique
To examine decomposition of plant material we measured dry mass loss of litter using the litter bag technique (Coleman et al. 2004). We compared two native plant species, C. australis and M. excelsum. C. australis leaves are recalcitrant and were expected to decompose slower than M. excelsum. Fresh leaf material was collected and oven dried at 30 °C for 6 days before placement into nylon mesh (5 mm) litter bags (10 × 20 cm). Oven drying at 30 °C was conducted in lieu of air-drying to obtain an initial mass as laboratory conditions for air-drying were too humid. Ten additional litter bags of each plant species were oven dried (60 °C for 48 h) as standards and weighed to obtain an average initial mass to dry mass correction factor. For each litter bag, the mass of leaf litter placed into the bag was recorded (to 2 decimal places), but ranged from 5–8 g of C. australis and 3–8 g of M. excelsum.
At each site, we randomly chose three plots, at least 20 m apart from each other. Thirty litter bags were placed in each plot on 1 February 2006. We pegged down the litter bags with a small wire peg after placing the 30 bags in a 6 × 5 arrangement, with five bags of each species in a row, tied together with fishing line to ease collection. From each plot, we collected a subset of ten bags after 1, 3 and 12 months. Litter bags were collected randomly from a plot. Thus, for each collection period there were 120 litter bags (2 species × 5 bags × 3 plots × 4 sites). Unfortunately the 12-month litter bags had been vandalised at both North Shore sites (Rotary Reserve and Sylvan Park) and therefore were not included in the final analysis.
Samples were oven-dried at 60 °C for 48 h then reweighed to obtain their dry-mass. Dry masses of the initial leaf material placed into the field were obtained using the correction factor (as described above) in the formulae: dry mass = mass initial × 0.32 (for C. australis) or 0.57 (for M. excelsum). Prior to obtaining dry mass, we took approximately 1 g of leaf material from the 1-month samples for microbe analysis (see details below), and consequently these litter bags were not included in the analyses of dry mass loss. After obtaining dry mass from the 3-month samples, we sent six litter bags for chemical analysis (see details below).
Litter samples were dried to prevent spoilage by microbiological activity during storage and then ground so that representative sub samples could be taken when a small sample mass were taken for analysis. As dried samples absorb water readily, they were briefly re-dried for 1 h at 105 °C (Landcare Research 2008).
Analysis of the litter was conducted after 3 months in the field by examining six litter bags per site (for each plant species) and percent content determined for the following litter components: nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), acid-detergent fibre, cellulose, and ADF-lignin. Ratios of carbon:nitrogen and acid-detergent fibre:N were also derived. Because each of these components is analysed a different way, with different laboratory steps, detailed methods can be found at Landcare Research (2008).
The presence of Argentine ants may result in a decline in invertebrates that consume the microorganisms colonising litter, thereby altering decomposition rates. We restricted our examination of litter microbes to those that dominate the first stage of the decomposition (fragmentation); yeasts and fungi. Hence, our further use of the term ‘microbes’ excludes bacteria. We examined the diversity of yeasts and microfungi from six litter bags of each plant species from each site (n = 48), using the following method. Microorganisms were washed off the leaves in phosphate buffer, serially diluted (10-0 and 10-1), plated on potato dextrose agar (PDA, Merck) and incubated at 20 ± 1 °C for 7 days according to the methods of Waipara et al. (2002). An antibiotic (PDA + Chlorotetracycline: 1 ml per 200 ml media) was applied to the media to restrict bacteria growth, as preliminary results showed bacteria formed dense growth and obscured the counting of other microbes, and yeasts and fungi were of primary interest. Following 7–10 days’ incubation, we identified colony forming units (CFUs) of yeasts and microfungi to recognisable taxonomic units (RTUs) and recorded their abundance (see Waipara et al. 2002).
We used a one-way ANOVA to compare each environmental variable between the invaded and uninvaded paired sites. Differences in invertebrate composition among sites was analysed via multivariate analyses using each pitfall trap as an independent replicate. However, first we tested for spatial autocorrelation by examining the similarity between of each pair of traps versus the distance between these traps. We produced a Bray–Curtis matrix using PRIMER v5.0 software (Clarke and Warwick 2005), for each pairwise combination of pitfall traps for each site (Order, ant and beetles data separately) and used a Pearson correlation to quantity the relationship between similarity and distance. Correlations revealed only weak relationships (range of Pearson r values: Orders = 0.03 to −0.19; Ants = 0.00 to −0.03; Beetles = −0.11 to −0.22). These correlations show that the catch of each pitfall is independent of one another, and consequently each pitfall can be treated as an independent replicate. This is congruent with previous research which also shows that 5 m spacing between pitfall traps produces invertebrate catches which are independent (Ward et al. 2001).
Multivariate analyses of the compositional differences of invertebrate orders, beetle species and ant species were conducted via non-metric multidimensional scaling in PRIMER v5.0 software, using a Bray–Curtis similarity matrix (no transformation) from 50 runs (Clarke and Warwick 2005). Analyses were also done on log transformed and presence/absence data but the results were consistent between different data transformations. We used the Analysis of Similarities (ANOSIM) routine, with 5,000 permutations, to analyse pairwise differences between the four sites. ANOSIM creates an overall test statistic (R) that indicates if differences between sites exist. As R approaches 1, there is more dissimilarity between sites. For ANOSIM, Clarke and Warwick (2005) use the definitions of: well separated R > 0.75, clearly different R > 0.5, and barely separable R < 0.25. In addition we used SIMPER analyses to examine which taxa contributed the most to the differences between invaded and uninvaded sites (Clarke and Warwick 2005). Only taxa which contributed >3 % to the overall difference were included in the results in order to avoid long lists of taxa.
We used a two-way ANOVA both to test for differences in soil chemistry at invaded and uninvaded sites at different sampling periods (summer and autumn) and to compare components of the litter at invaded and uninvaded sites and between locations (North Shore and Waikowhai).
To examine litter bag decomposition, we calculated a decay rate (k), which estimates litter loss, from the percent of dry weight remaining in the litterbags over time (Olson 1963). The usual method to analyse k is by using a negative exponential decay function (Olson 1963; Standish et al. 2004). However, our preliminary statistics indicated an exponential decay model failed to explain the current data, and therefore, we used an unbalanced ANOVA in GenStat v8.0 to examine differences in site invasion, sampling period and location on proportion of litter mass remaining (Wardle et al. 2011; Wider and Lang 1982). For C. australis there was no ‘location’ effect because of vandalism of litter bags at the North Shore sites. For M. excelsum there are only two time points because at the 12-month sampling period no leaf material remained in the bags.
Mean (±SE) of environmental variables used to characterise sites
Soil temp (°C)
Soil moisture (%)
Canopy cover (%)
Vegetation cover (%)
Litter density (g/cm2)
#Seedlings (0.75 m2)
SIMPER analysis showing the mean abundance (±SE) of invertebrate orders captured in pitfall traps between the invaded and uninvaded sites on the North Shore, and the cumulative contribution (%) of taxa to the overall difference
SIMPER analysis showing the mean abundance (±SE) of ant species captured in pitfall traps between invaded and uninvaded sites, and the cumulative contribution (%) of taxa to the overall difference
SIMPER analysis showing the mean abundance (±SE) of beetle species captured in pitfall traps between invaded and uninvaded sites, and the cumulative contribution ( %) of taxa to the overall difference
Beetle species (family)
Lithostygnus sp. (Latridiidae)
Stenomalium sp. (Staphylinidae)
Aleocharinae spp. (Staphylinidae)
Syncalus sp. (Zopheridae)
Epuraea sp. (Nitidulidae)
Ptiliidae sp. 1
Anotylus sp. (Staphylinidae)
Zeadolopus maoricus (Leiodidae)
Mandalotus sp. (Curculionidae)
Micrambina sp. (Cryptophagidae)
Ablabus sp. (Zopheridae)
Means (±SE) of soil chemistry variables at both invaded and uninvaded sites
Soil chemistry variable
Total C (%)
Total N (%)
Olsen P (mg/kg)
Water content (% dry mass)
Litter bag technique
Mean (±SE) of litter components (%) for C. australis and M. excelsum at both invaded and uninvaded sites
Carbon: nitrogen ratio
Acid-detergent fibre: nitrogen ratio
Carbon: nitrogen ratio
Acid-detergent fibre: nitrogen ratio
Means (±SE) of microbial colony forming units (CFUs) of yeast and fungi for each plant species at invaded and uninvaded sites, across two dilution factors
Plant species + dilution factor
Although the paired invaded and uninvaded sites in our study were similar in litter and vegetative cover, as well as soil temperature and moisture (key factors in determining ant distribution), there were strong differences in the ant fauna, beetle fauna and other invertebrates associated with the presence of Argentine ants at invaded sites. The invaded sites were characterised by Argentine ants and the presence of the exotic M. abstinens, a cryptic hypogenic ant (Andersen 1997) that is often found in higher densities where Argentine ants have invaded in New Zealand (Ward, unpubl. data). Uninvaded sites were characterised by a richer ant fauna, characterised by several other exotic ant species (e.g. Nylanderia spp., T. grassii). The only native ant sampled in the study, P. advenus, was found only in uninvaded sites. There is abundant evidence to show Argentine ants change ant community composition and negatively affect other ant species (Holway 1998; Human and Gordon 1997; Sanders et al. 2003; Walters 2006).
Although previous studies have shown that Argentine ants do affect the abundances and composition of other invertebrates, the effects on different groups are not consistent (Cole et al. 1992; Holway 1998; Human and Gordon 1997; Walters and Mackay 2003). Our results showed strong differences in the beetle community between invaded and uninvaded sites. Krushelnycky and Gillespie (2010) also found endemic beetles to be highly vulnerable to ant invasion in Hawaii. Of particular interest to decomposition are endemic species in the families Zopheridae, Leiodidae, and Cryptophagidae, which are heavily associated with fungal feeding and the decomposition of leaf litter (R. Leschen, Landcare Research, pers. comm.), and which were almost exclusively found at uninvaded sites. If these species were to be reduced significantly at sites where Argentine ants invade, we would expect a cascading effect on decomposition processes. If fungal feeding beetles alone were reduced by Argentine ants, we would expect fungal populations to increase following enemy release and therefore litter breakdown to occur at a faster rate. However, multiple trophic pathways are affected by Argentine ants, and the overall biomass of fungal feeding beetles is much less in this system than the biomass of other groups, such as Collembola, and we would therefore expect impacts on fungal feeding beetles to have relatively minor effects on overall decomposition rates.
Collembola were substantially more abundant in invaded sites, while Isopods and Amphipods were substantially more abundant in uninvaded sites. That the abundance of Collembola was substantially greater at invaded sites is contrary to the results of Dunham and Mikheyev (2010). Collembola also play an important role in the breakdown of decaying vegetation, but appear to primarily consume bacteria and fungi rather than breaking down decaying vegetation (Rusek 1998). The functional role of Collembola is not determined at the Order level and is often oversimplified by ecologists (Rusek 1998). The trophic level of Collembola in this study, and whether they could mask the effects of Argentine ants on decomposition (assuming they are impacted on by Argentine ants), is unknown at present. However, we suggest that future research on invasive ant impacts on ecosystem function, should examine Collembola at higher taxonomic resolution, given the breadth of their roles in litter and soil. Amphipods also play a key role in these habitats as decomposers, given they principally consume decaying plant material (Friend and Richardson 1986) and are one of the most abundant macro-invertebrates in New Zealand litter (Duncan 1994). Decreases in their numbers could affect decomposition, especially the initial breakdown of leaf material. While in our study, there were significantly more amphipods in uninvaded sites, Walters and Mackay (2003) found increased abundances of isopods and amphipods at invaded sites.
Argentine ant literature reveals variability in impacts on invertebrates in different ecosystems. Strong effects on invertebrates are not universal, nor consistent, and often there are both significant decreases and increases in abundances of different invertebrate groups associated with Argentine ants. For example, in Hawaii, Cole et al. (1992) found Argentine ants negatively affected a range of invertebrate groups, including spiders, earwigs, and some beetles, moths, snails, springtails, bugs, flies and wasps. Similarly, Human and Gordon (1997) showed significant reductions in groups of flies, wasps, and arachnids at invaded sites, but increased numbers of isopods, crickets and ground beetles. Apart from ants, Holway (1998) found there were similar abundances of invertebrates at invaded and uninvaded sites in California. He suggested pitfall traps may have failed to capture certain invertebrate groups, contributing to a lack of differences, or that immigration from adjacent areas may be sufficient to sustain populations of invertebrates. Unlike that of Holway (1998), we found that pitfall trapping was the only method to consistently demonstrate differences between invaded and uninvaded sites. Whether regional differences, such as the depauperate nature of ant communities in New Zealand, and lack of ecosystem engineering by ants relative to the US and Australia (Andersen and Sparling 1997; Dóstal et al. 2005; Lavelle et al. 1995), are responsible for these results is unknown.
Implications for decomposition and soil microbes
Although the paired invaded and uninvaded sites in our study were very similar, there were significant differences in soil microbial concentration and decomposition rates (litter bag technique and litter chemistry) which suggest an indirect association between Argentine ants and decomposition that is worth pursuing in future research on Argentine ants. Although there was variability due to location in the loss of dry mass of litter using the litter bag technique from M. excelsum, the results of chemical analyses of this litter indicated that decomposition was significantly slower at sites with Argentine ants. In particular, the C:N ratio and acid-detergent fibre were significantly higher at invaded sites, indicating a greater fibrous content to leaf litter and less decomposed material. These are extremely important variables in the assessment of litter decomposition, far more indicative of decomposition than litter bag technique (Melillo et al. 1982; Berg and McClaugherty 2008). Conversely, at uninvaded sites, there was a significantly higher concentration of N, Ca, and Mg remaining in the litter which indicates greater litter decomposition at uninvaded sites. Nitrogen concentration increases as a function of litter loss; this a strongly linear relationship often used in lieu of the litterbag technique to compare decomposition rates between sites (Berg and Laskowski 2006; Berg and McClaugherty 2008). In Swedish forests, Wardle et al. (2011) found long-term exclusion of red wood ants (Formica rufa) indirectly resulted in strong increases in decomposition rates as a consequence of increased herbaceous vegetation in the plots and top-down multi-trophic effects on soil fauna.
It is not surprising that C. australis did not show any significant differences in litter chemistry, as this species is very fibrous and recalcitrant, and therefore decomposition is likely to occur over a much longer time span than for M. excelsum. Moreover, the litterbag technique showed that C. australis decomposition was slower at the invaded Waikowhai site than at the uninvaded Waikowhai site (litterbags at the other sites were vandalised) in the first month, but then switched to significantly greater litter loss in the 3–12 month period than the uninvaded site. M. excelsum and C. australis are the two extremes of litter type in these ecosystems, and these results reinforce that decomposition processes are not likely to be occurring in a similar way for all litter types (Berg and McClaugherty 2008). We had considered that invaded sites might have higher fungal and yeast counts on decomposing litter due to trophic cascade as a result of reduced abundance of invertebrates that feed on fungi and yeast. Although key endemic fungus-feeding beetles were reduced at invaded sites, there were no significant differences in the counts of either yeast or fungal colonies on leaf litter between invaded and uninvaded sites. There are likely to be other mechanisms involved, such as competition and regulation among fungal and yeast species on the litter.
In our study, microbial biomass C and N were significantly lower in soils at invaded sites, and this reduced soil microbial biomass is likely to be indicative of negative belowground effects, meriting further investigation. This corresponds with Wardle et al. (2011) who found that the absence of red wood ants from plots resulted in strong increases in soil microbial biomass. Soil microbial biomass is an important indicator of soil health (Bending et al. 2004; Burton et al. 2010), and is used extensively as an indicator of soil degradation or improvement (Nannipieri et al. 2003; Wardle 1992).
Our research provides evidence for strong differences in the ant fauna, beetle fauna and other invertebrates associated with the presence of Argentine ants at invaded sites. Although there have been many papers addressing the invasive traits and above-ground impacts of Argentine ants (Kenis et al. 2009; Sanders and Suarez 2011), no research has examined the effects of this invasive species on belowground processes. Our findings of significantly lower soil microbial concentrations and slower decomposition (litter bag technique and litter chemistry) in the invaded sites (cf. similar uninvaded sites) provide the first indication of an association between Argentine ants and decomposition and soil health. One possible mechanism for the association is a trophic cascade via non-ant invertebrates, particularly those with roles in decomposition. This study should act as a driver for experimental research on the association between Argentine ants (and other invasive ants) and belowground processes.
If Argentine ants affect the litter fauna (richness and abundance) and are impacting soil organisms and decomposition, there could well be long-term effects on ecosystems. However, the abundance and persistence of Argentine ants at a site is likely to determine the magnitude of impacts (Parker et al. 1999). Argentine ants are very mobile and often do not persist at sites, especially those at the limits of their abiotic tolerances (Heller and Gordon 2006). The mean abundance of Argentine ants at our invaded sites appears to be lower than in other studies (Holway 1998; Suarez et al. 1998; Walters 2006), and may in part explain why the indirect effects in our study were not strong or consistent. Furthermore, it is possible that ant species such as Argentine ants—which are active mostly in the top of the leaf litter, are transient and mobile (Heller and Gordon 2006; Suarez et al. 1998)—have much less of an effect on decomposition and soil compared with mound-building ants (Andersen and Sparling 1997; Cammeraat and Risch 2008; Dóstal et al. 2005; Lavelle et al. 1995; Wardle et al. 2011). Alternatively, the effect could be more widespread over sites, rather than the irregular, aggregated nature of mound-building ant impacts (Wardle et al. 2011).
We thank staff at Auckland and North Shore city councils for permission to carry out this work; Robyn Simcock, Brian Daly, and Graham Sparling for advice on soil and chemical sampling; Chris Winks, Shane Hona, Josie Galbraith, Sarah Harrison, Sarah Knight for field assistance; Jo Rees for sorting invertebrate samples and Stephen Thorpe for identifying beetles; Nick Waipara and Olimpia Timudo for microbial advice, laboratory work and identification; and Guy Forrester for statistical advice; Richard Toft and Nathan Sanders for helpful comments on earlier versions of this manuscript. This work was supported by the Foundation for Research, Science and Technology (FRST C09X0507).
- Allen R (1993) A permanent plot method for measuring changes in indigenous forests: a field manual. Landcare Research Ltd, ChristchurchGoogle Scholar
- Berg B, Laskowski R (2006) Litter decomposition: a guide to carbon and nutrient turnover. Advances In ecological research 38. Academic Press, BurlingtonGoogle Scholar
- Berg B, McClaugherty C (2008) Plant litter: decomposition, humus formation, carbon sequestration, 2nd edn. Springer-Verlag, BerlinGoogle Scholar
- Clarke KR, Warwick RM (2005) Change in marine communities. An approach to statistical analysis and interpretation, 2nd edn. Plymouth Marine Laboratory, PlymouthGoogle Scholar
- Coleman DC, Crossley DA, Hendrix PF (2004) Fundamentals of soil ecology, 2nd edn. Elsevier Academic Press, San DiegoGoogle Scholar
- de Kock AE, Giliomee JH (1989) A survey of the Argentine ant, Iridomymrex humilis (Mayr), (Hymenoptera: Formicidae) in South African fynbos. J Entomol Soc S Afr 52:157–164Google Scholar
- Don W (2007) Ants of New Zealand. Otago University Press, DunedinGoogle Scholar
- Duncan KW (1994) Terrestrial Talitridae (Crustacea: Amphipoda). Fauna of New Zealand series 31. Manaaki Whenua Press, LincolnGoogle Scholar
- Green OL (1990) Entomologist sets new record at Mt Smart for Iridomyrmex humilis established in New Zealand. Weta 13:14–16Google Scholar
- Hoffmann BD, Andersen AN, Hill GJE (1999) Impact of an introduced ant on native rain forest invertebrates: Pheidole megacephala in monsoonal Australia. Oecologia 120:595–604Google Scholar
- Landcare Research (2008) Environmental chemistry laboratory tests. http://www.landcareresearch.co.nz/services/laboratories/eclab/eclabtest_list.asp. Accessed 28 July 2008
- Leathwick J, Morgan F, Wilson G, Rutledge D, McLeod M, Johnston K (2003) Land environments of New Zealand technical guide. David Bateman Ltd, AucklandGoogle Scholar
- McEwen WM (1987) Ecological districts and regions of New Zealand, 3rd ed. Biological Resources Centre Publication No 5. Department of Conservation, WellingtonGoogle Scholar
- Roura-Pascual N, Hui C, Ikeda T, Leday G, Richardson DM, Carpintero S, Espadaler X, Gómez C, Guénard B, Hartley S, Krushelnycky P, Lester PJ, McGeoch MA, Menke S, Pedersen J, Pitt J, Reyes J, Sanders NJ, Suarez AV, Touyama Y, Ward DF, Ward P, Worner SP (2011) The relative roles of climatic suitability and anthropogenic influence in determining the pattern of spread in a global invader. Proc Natl Acad Sci USA 108:220–225PubMedCrossRefGoogle Scholar
- Sanders NJ, Suarez AV (2011) Elton’s insights into the ecology of ant invasions: lessons learned and lessons still to be learned. In: Richardson DM (ed) Fifty years of invasion ecology: the legacy of Charles Elton. Blackwell, New York, pp 239–251Google Scholar
- Solaiman K (2007) Measurement of microbial biomass and activity in soil. In: Varman A, Oelmüller R (eds) Soil biology, vol 11., Advanced techniques in microbiologySpringer-Verlag, Berlin, pp 201–211Google Scholar
- Waipara NW, Obanor F, Walter M (2002) Impact of phylloplane management on microbial populations. NZ Plant Prot 56:125–128Google Scholar
- Walters AC, Mackay DA (2003) The impact of the Argentine ant, Linepithema humile (Mayr), on native ants and other invertebrates in South Australia. Rec S Aust Mus 7:17–24Google Scholar
- Ward DF (2009) The diversity, community composition and seasonality of native ants in northern New Zealand. Myrmecol News 12:195–200Google Scholar
- Ward DF, Harris R (2005) Invasibility of native habitats by Argentine ants, Linepithema humile, in New Zealand. NZ J Ecol 29:215–219Google Scholar
- Ward DF, Harris RJ, Stanley MC (2005) Human-mediated range expansion of Argentine ants Linepithema humile (Hymenoptera: Formicidae) in New Zealand. Sociobiology 45:401–407Google Scholar
- Ward DF, Harris RJ, Hartley S, Lester PJ, Stanley MC, Suckling DM, Toft RJ (2010) Twenty years of Argentine ants in New Zealand: past research and future priorities for applied management. NZ Entomol 33:67–78Google Scholar
- Williams DF (1994) Exotic ants: biology, impact, and control of introduced species. Westview Press, BoulderGoogle Scholar
- Zelikova TJ, Sanders NJ, Dunn RR (2011) The mixed effects of experimental ant removal on seedling distribution, belowground invertebrates, and soil nutrients. Ecosphere 2(5):art63. doi:10.1890/ES11-00073.1