Introduction

Considered one of the worst invasive pest ants (Lowe et al. 2000), the little fire ant Wasmannia auropunctata (Roger) (Hymenoptera: Formicidae) has negatively impacted both biodiversity and agriculture (Le Breton et al. 2003; Wetterer and Porter 2003; Walker 2006). Its distribution is nearly pantropical, and greenhouse infestations have been reported as far north as Canada and the United Kingdom (Jourdan et al. 2002; Wetterer and Porter 2003). Despite having been present in Florida for approximately 85 years, W. auropunctata is not considered a major pest there; on invaded Pacific islands however, it has much larger ecological and economic impacts (Wetterer and Porter 2003).

While W. auropunctata workers are rather slow-moving and diminutive in size, the potent venom from their stings is painful, and they have become an important deterrent to farm laborers that harvest infested crops (Smith 1965; Conant 2000). They also affect agriculture by tending homopterans, which directly damage crops and may vector disease (Spencer 1941; de Souza et al. 1998). Ecological impacts include the stressing of vertebrates (Wetterer 1997; Wetterer et al. 1999), reduction of invertebrate populations (Ulloa-Chacón et al. 1991), and displacement of native ants (Le Breton et al. 2003; Walker 2006).

Improved quarantine and prompt eradication of invasive populations are key to controlling the spread of W. auropunctata (Wetterer and Porter 2003). Current W. auropunctata detection methods employ a food item like peanut butter, which is placed on the ground for an unstandardized length of time and later inspected for the presence of ants (Causton et al. 2005; Kirschenbaum and Grace 2007). However, this technique is not specific to W. auropunctata and thus species ID becomes more complicated by the presence of multiple species. For example, peanut butter was used to survey for W. auropunctata on the island of Maui (Starr et al. 2007). Although W. auropunctata was not detected in the Maui study, over 20 different species of ants were caught, demonstrating the complexity of surveying for a single ant species using a food bait with broad appeal to many ant species. In addition to lacking specificity, food baits are often messy to use, are more susceptible to spoilage, and may be eaten by vertebrates in the field. Given these drawbacks to the use of food baits, there is a need for the development of improved detection methods that increase trapping efficiency and ease—of—use.

Trapping methods are often improved with semiochemical attractants, because these compounds tend to be more species—specific than generic food bait. This is particularly true of pheromones, which usually attract only a single species. Recently, as part of an effort to increase bait specificity, two components of the W. auropunctata alarm pheromone were identified as 2,5-dimethyl-3-(2-methylbutyl)pyrazine (2-MeBu-diMePy) and 3-methyl-2-(2-methylbutyl)pyrazine (2-MeBu-MePy) (Showalter et al. 2010). In field and laboratory bioassays, both pyrazines induced attraction and arrestment, and increased locomotion (Troyer et al. 2009; Showalter et al. 2010). The strength and longevity of attractiveness demonstrated by these pheromones is promising for their use in ant detection and control.

Many detection methods improve efficiency and species-specificity by combining the use of an attractant (often a pheromone) with some means of capture (often a trap and/or an adhesive material) (El-Sayed et al. 2006). In a previous study, when 2-MeBu-diMePy was deployed in combination with Tanglefoot, a sticky catch material, groups of W. auropunctata were observed to arrest around the adhesive area for up to eight days (Troyer et al. 2009). Ants probed the catch material with their antennae, but demonstrated a strong aversion to venturing onto the coated substrate. These results suggest that Tanglefoot is a poor choice for W. auropunctata detection, but do not rule out other adhesive materials as potential capture materials.

Another detection method, which involves combining one—way traps with pheromone, was developed through serendipitous observations of perforated weedmat. While conducting field tests, we observed that ants could squeeze through holes in only one direction. This observation led to the development of a one—way trap prototype, which could provide a new means to capture and retain ants. In addition, the combination of a one-way trap and pheromone may increase species specificity by attracting only W. auropunctata. It is expected that other small ants will not be attracted by the pheromone and will have little reason to explore the trap while larger ants will be excluded from the trap by the small hole size in the weedmat. The combination of a one— way trap and an attractant has previously been used to trap tephritid fruit flies in Hawaii, demonstrating the feasibility of this trapping technique (Hiramoto et al. 2006; Jang 2011). However, further testing needs to be conducted to determine the efficacy and specificity of this trapping method for W. auropunctata and whether it could be an improvement on current detection methods.

The development of a more efficient and species specific detection method would also allow the investigation of W. auropunctata microhabitat preferences. Wasmannia auropunctata are known to colonize trees as well as the ground, and are sometimes found in houses (Fernald 1947). Arboreal colonies are often found around the bases of trunks or in crevices, depressions, and crooks formed by branches (Spencer 1941). However, ants have also been known to almost blanket the ground, making nests in any slightly protected area (Spencer 1941; Wetterer and Porter 2003). Undetected arboreal W. auropunctata are often difficult to control, since they will recolonize the ground when treatment ends (Vanderwoude and Nadeau 2009). Resolving W. auropunctata preferences for arboreal versus non—arboreal microhabitats could help to refine ecological understanding of this ant and aid in optimizing trap placement for detection and control purposes.

This paper presents a series of field experiments comparing W. auropunctata detection methods that use either 2-MeBudiMePy as a pheromone attractant or peanut butter as a food lure. Detection methods using the pheromone included either a sticky catch material or a one—way trap. Trap data from these comparisons were also used to evaluate ant species diversity and spatial distribution in the macadamia orchard where the experiments were conducted. Additionally, a vegetation survey of ant microhabitat preferences was conducted, which provided information that would be helpful in the optimization of trap placement for W. auropunctata detection/control.

Materials and Methods

Results

Experiment 1 compared 10 types of tape for capturing and retaining W. auropunctata. Tape types that retained ants (2–500 ants per piece of tape) were clear string reinforced, outdoor carpet, colored label, duct, masking, and electrical (ranked highest to lowest). Of these, the double—sided outdoor carpet tape was chosen for further testing based on the comparatively high number of ants captured and the ease of field placement for testing. Tapes that did not retain ants were scotch tape, blue painter's tape, double—sided poster tape, double—sided scotch tape, and brown packing tape. In several instances, ants were observed to walk unimpeded across this latter group of tapes.

Experiment 2 assessed the diversity of ant species along a transect by means of a visual survey (Figures 1 and 3). From tree collections, 10 ant species were identified. These were Brachymyrmex obscurior Forel, Cardiocondyla obscurior Wheeler, Leptogenys falcigera Roger, Monomorium floricola (Jerdon), Pheidole fervens Smith, Solenopsis papuana Emery, Tapinoma melanocephalum (Mayr), Technomyrmex sp. (vitiensis (Emery) and/or albipes Mann), Tetramorium bicarinatum (Mayr), and W. auropunctata (Roger). With the exception of plots 30 and 39, W. auropunctata was found exclusively in the western end of the transect (Figure 3). Additionally, when W. auropunctata was found on a tree, it was the only species collected from that tree 94% of the time. Wasmannia auropunctata was found as the lone ant species at a significantly higher frequency than all other ant species except P. fervens (p < 0.05, Q < 0.001) (Figure 4).

Experiment 3 compared three W. auropunctata detection methods (Figures 5 and 6). In plots where W. auropunctata was detected during the visual survey (Experiment 2), the number of ants caught per tree did not vary significantly between detection methods (ANOVA, p = 0.558) (Figure 6A). In plots where W. auropunctata was not detected during the visual survey, the one—way trap and double—sided tape, which both used a pheromone attractant, caught significantly fewer ants of all species than the peanut butter detection method (pair—wise Mann—Whitney U test, p < 0.01) (Figure 6B). In one instance (plot 39), the one—way trap detected previously unobserved W. auropunctata (Figure 5). There was one instance (plot 26) where C. obscurior was found inside the one—way trap, seven instances of species other than W. auropunctata caught with the doublesided tape, and 25 instances of species other than W. auropunctata on the peanut butter coated stick. Eight of the 41 peanut butter coated sticks (originally placed in trees) were recovered from the ground, some of which no longer had peanut butter. Presumably they were knocked down by vertebrate wildlife (i.e., mongooses, pigs, birds) and licked clean. The one—way trap and double—sided tape captures were 98% and 83% species specific (detecting the known presence of W. auropunctata without capturing other species), respectively. While there were no significant differences between the species—specificity of captures with the one—way trap or double—sided tape captures (p = 0.0571), they were both significantly more selective than peanut butter (39%) (p < 0.01 for both).

Experiment 4 investigated W. auropunctata microhabitat preferences with one-way trap placement (Figure 7). Wasmannia auropunctata were detected in all microhabitats. When analyzed individually, the mean number of W. auropunctata caught differed significantly among microhabitats (Kruskal—Wallis test, p < 0.05). When trap placements were grouped as arboreal or non—arboreal, the tree-associated traps caught significantly more W. auropunctata than those located further from the tree (Mann—Whitney U test, p = 0.01).

Discussion

The assessment of different sticky catch materials for use in W. auropunctata detection (Experiment 1) suggested that tested tapes fall into two groups; materials which W. auropunctata will traverse and sometimes become stuck to, and those which they will traverse without becoming stuck. A third group of materials are those that deter W. auropunctata from traversing, presumably because they are highly adhesive and are perceived as such by the ants. An example of this third group of sticky materials is Tanglefoot, which acts as a barrier that W. auropunctata will not cross or become entangled in when 2-MeBu-diMePy is present as an attractant (Troyer et al. 2009). These observations of W. auropunctata behavior and locomotion with regard to sticky materials suggest that adhesives should be of intermediate strength so as not to deter ants from trying to cross while being sticky enough to entangle ants during crossing.

Survey data (Experiment 2) supports previous research describing the ability of W. auropunctata to displace other ant species (Le Breton et al. 2002; Armbrect and Ulloa-Chacon 2003; Le Breton et al. 2003; Kirschenbaum and Grace 2007). Of all ants surveyed, the data show W. auropunctata to be the least likely to co—inhabit with another ant species in a given tree (Figure 4). The majority of the other ant species were consistently found with 2-3 other species on the same tree.

Field studies of W. auropunctata detection methods (Experiment 3) showed increased species specificity when the pheromone 2-MeBu-diMePy was used in combination with a one—way trap or double—sided tape (Figure 5 and 6). While all methods succeeded in detecting W. auropunctata, only the methods using 2-MeBu-diMePy were significantly more species—specific; detecting the known presence of W. auropunctata without capturing other species. Multiple ant species were observed feeding on peanut butter—coated sticks, and interference from larger animals (presumably mongoose, pigs, or birds) made field testing problematic. Similar problems were encountered in W. auropunctata surveying attempts using peanut butter in the Galapagos (Abedrabbo 1994). Additionally, comparisons between visual survey and trapping data confirmed the detection reliability of one—way traps and double—sided tape used in combination with 2-MeBu-diMePy; W. auropunctata was caught wherever it had been previously observed, with one additional occurrence detected with the one—way trap (Figure 3). These results demonstrate the potential of pheromone detection methods to selectively capture target ant species even when other ants are present and abundant.

Of particular interest to regulatory agencies, pheromone aided detection may decrease the need for taxonomic identification of field samples if only target ant species are attracted to the detection device. There are two commonly used ant monitoring methods: a broad survey of taxa diversity and the targeted detection of single species (often invasive pests) to aid in control. Surveys for ant taxa diversity commonly rely on pitfall traps, baited traps, and leaf—litter extractions (or some combination thereof) for collections (Botes et al. 2006; Neville et al. 2008; Ouellette et al. 2010). This requires trained taxonomists to sort through large sample quantities of entire invertebrate assemblages. Thus, the aforementioned survey methods are utilized with great inefficiency towards the purpose of monitoring a single target species, be it for pest control, tracking the spread of invasive species, or emerging applications like biodiversity monitoring. For the latter, Andersen (2004) describes the dearth of species—level tools for non—specialists as a primary hindrance to implementing biodiversity monitoring programs. However, the high species specificity of both the one— way trap and the sticky tape, in combination with an attractive pheromone, has the potential to be such a tool. Pushing non—specialist detection to an extreme, species—specific trapping might even be applicable to citizen/scientist collaboration monitoring specific ant species (Braschler 2009). Using pheromone traps would enable members of the public to simply report the presence or absence of ants without creating a large number of samples to be examined.

While the one—way trap and the double—sided tape performed similarly in the field, there are differences between the two methods that should be considered in assessing which is better suited for W. auropunctata monitoring. From a design standpoint, the simplicity of the double—sided tape is appealing. However, over longer test periods the sticky surface is likely to become covered with detritus from the environment, which would decrease the capacity to retain attracted W. auropunctata. The one—way trap has the advantage of excluding insects ∼1 mm and much of the environmental detritus, which might affect the tape. Additionally, even though species specificity was not significantly different between these two detection methods, it is our opinion that the one—way trap has the potential to be more selective long—term. The one—way trap also has the potential to capture many more ants given the larger volume inside the trap. Beyond use with W. auropunctata, this trap has the potential to be used with other ant pheromones, providing a model for species specific monitoring of small (1–2 mm) ants.

Results from the assessment of several microhabitats within the macadamia orchard (Experiment 4) suggested that the optimal location for one—way trap placement is in trees. Although all traps caught W. auropunctata, one—way traps in trees or near the foot of macadamia trees caught more ants than those on the ground or in grass or shrub thickets. This speaks to the near ubiquity of W. auropunctata in infested areas, and the preference of W. auropunctata populations for tree-based microhabitats in the macadamia orchard.

Figure 1.

Aerial view of the experimental site, a macadamia orchard outside Papaikou, HI. Red dots mark the transect used in Experiments 2 and 3. High quality figures are available online.

Figure 2.

Assembly diagram for one-way traps. Pushpins, septas, and weedmat pieces were attached with hotmelt glue. The Petri dish pieces were fitted together without adhesive. High quality figures are available online.

Figure 3.

Ant species collected in transect plots. (A) Experiment 2, visual survey/collection. (B) Experiment 2, combined ant captures (peanut butter, one—way trap, and double—sided tape). Red labels indicate that the same ant species was collected (Exp. 2) and trapped (Exp. 3) on the same tree during both experiments. High quality figures are available online.

Figure 4.

Ant species diversity (cohabitation) data. Wasmannia auropunctata was found as the lone ant species in a plot at a significantly higher frequency (p > < 0.05, Q < 0.001) than all other ant species except Pheidole fervens (Fisher's exact test with FDR analysis). High quality figures are available online.

Figure 5.

Experiment 3, number of ants caught by the three detection methods: (A) Peanut butter, (B) one–way traps, (C) double—sided tape. Red circles indicate plots where Wasmannia auropunctata were detected, black circles indicate plots where other ant species were detected. High quality figures are available online.

Figure 6.

Experiment 3, numbers (mean ± SE) of ants captured per tree. (A) Traps capturing Wasmannia auropunctata, primarily at the west end of the transect. Letters represent significant differences (p < 0.05) between arrestments and crossings of different treatments (ANOVA, followed by Tukey's HSD). (B) Traps capturing other ant species, primarily in the middle and east end of the transect. * Median significantly different than other medians (Kruskal—Wallis, followed by Mann—Whitney U test). ** Median significantly different than one—way trap (Kruskal—Wallis, followed by Mann—Whitney U test). High quality figures are available online.

Figure 7.

Experiment 4, numbers (mean ± SE) of Wasmannia auropunctata captured in one—way traps in different microhabitats. Letters represent significant differences (p < 0.05) between arboreal and non—arboreal placement groupings (Kruskal—Wallis, followed by Mann—Whitney U test). High quality figures are available online.