Henry Fadamiro, Department of Entomology & Plant Pathology, Auburn University, Auburn, AL 36849, U.S.A. E-mail: email@example.com
Abstract 1. The imported fire ants, Solenopsis invicta Buren (red), S. richteri Forel (black), and their hybrids (S. invicta × S. richteri) are sympatric congeners with overlapping but shifting spatial distributions in northern Alabama and Mississippi, United States.
2. The abilities of workers of the three ant forms (or species) to recognise non-nestmate conspecifics and heterospecifics were compared in separate aggression bioassays using three group sizes of competitor numbers (battles): 1-1, 5-1, and 5-5 resident–intruder battles.
3. Workers of all three forms showed significant aggression toward non-nestmate conspecifics and heterospecifics. The results of the intra-specific aggression bioassays showed that S. invicta had the lowest aggression threshold, whereas S. richteri was the least aggressive. Survival rates 1 day after intra-specific encounters were higher for S. richteri than for S. invicta or hybrids, consistent with its lower aggressiveness.
4. In inter-specific interactions, S. invicta workers showed the greatest aggression and were more aggressive towards S. richteri than to hybrids. Furthermore, survival 1 day after inter-specific encounters was highest for S. invicta workers, whether they were the intruders or residents. The hybrid form was intermediate between both parental forms in aggression and post-aggression survival.
5. The live 1-1 and 5-5 battles produced similar and consistent results, but differences were amplified in the 5-5 battles, suggesting an effect of group size on aggression. These results may explain the observed shifts in the spatial distributions of the three forms and range restriction of S. richteri in southern United States.
Ants and other social insects are known to exhibit aggressive behaviour towards conspecifics from different colonies (intra-specific aggression) and heterospecifics (inter-specific aggression) (Wilson, 1971; Hölldobler & Michener, 1980). Aggressive behaviour in social insects is regarded as an adaptation for nestmate recognition and protection of the colony against threats and intrusion, such as robbery, predation, parasitism, and competition (Wilson, 1971).
Two species of imported fire ants, Solenopsis spp. (Hymenoptera: Formicidae) were accidentally introduced from South America into southern United States (U.S.) in the past century. The black imported fire ant, S. richteri Forel was introduced in the 1910s followed by the introduction of the red imported fire ant, S. invicta Buren in the early 1930s. Both species are believed to have entered the U.S. through the port of Mobile, Alabama, spreading northward in Alabama and into neighbouring states (Wilson, 1958). While S. invicta is now widely distributed throughout southeastern U.S. (Callcott & Collins, 1996; Porter, 2000), S. richteri is restricted in distribution to a relatively small area around the northern border of Alabama, Mississippi, and southern Tennessee (Obin & Vander Meer, 1989b; Streett et al., 2006). Hybridisation between S. invicta and S. richteri has been documented in Alabama, Mississippi, Georgia, and Tennessee (Vander Meer et al., 1985; Ross et al., 1987a, b; Diffie et al., 1988; Vander Meer & Lofgren, 1988; Shoemaker et al., 1996). However, hybrid imported fire ants are known to occur parapatrically with both parent types only in the Mississippi hybrid zone, which consists of areas in northern Mississippi and Alabama (Vander Meer & Lofgren, 1988; Streett et al., 2006). The restricted range of S. richteri has been attributed to displacement by S. invicta, as well as the temporal stability and continued range expansion of the hybrid form (Tschinkel, 1993). In northern Alabama and Mississippi where the three types of imported fire ants (referred to as ‘forms’ in this paper) occur sympatrically, population decrease and range restriction of S. richteri is being observed, whereas population increases and range expansion are recorded for S. invicta and hybrids (Streett et al., 2006).
Several factors may account for the observed population dynamics and shifts in spatial distributions of the three forms of imported fire ants in southern U.S., including genetic, ecological, and other factors. For instance, it is possible that hybrids simply preferentially backcross with S. richteri (rather than with S. invicta), thus leading to an erosion of pure S. richteri populations. However, inter-specific competition through exploitative competition and/or inteference mechanisms, such as inter-specific aggression, is the most probable factor (Wilson, 1971; Holway & Suarez, 1999; Calcaterra et al., 2008). Several authors have attributed the ecological dominance of S. invicta in southern U.S. to its superior competitive ability (Jones & Phillips, 1987; Tschinkel, 2006). However, the exact mechanisms which mediate its competitive advantage remain unclear. In southern U.S., S. invicta colonies occur in two social forms, monogyne (single queened) and polygyne (multiple queened) colonies (Ross et al., 1996; Tschinkel, 1998). Polygyne colonies of S. invicta typically show reduced nestmate discrimination compared with monogyne colonies, possibly as a result of increased queen number and population density (Vander Meer et al., 1991; Holway & Suarez, 1999). Thus, the imported fire ant complex in southern U.S. presents a good model system for investigating the mechanisms of inter-specific competition, in particular the connection between aggression (nestmate discrimination) and competitive superiority.
In perhaps the only published study of nestmate discrimination among the three imported fire ant forms, Obin and Vander Meer (1989b) reported no significant differences in the capabilities of S. invicta, S. richteri, and their hybrids to display aggression towards heterospecific competitors. However, our observations of nestmate discrimination in laboratory populations and recent shifts in the distributions of the three forms in southern U.S. suggest the possibility of differences in their aggressive behaviours. Thus, the present study was conducted to re-investigate nestmate discrimination and compare aggression of the three imported fire ant forms towards conspecifics and heterospecifics. The main aim was to investigate whether or not the observed shifts in the populations and spatial distributions of the three forms can be explained by aggression. Based on the range expansion and wider distribution of S. invicta and hybrids relative to S. richteri, we hypothesised that S. invicta and hybrids will demonstrate a lower aggression threshold than S. richteri.
Sources and maintenance of ant colonies
Three colonies of S. invicta, and six colonies each of S. richteri and hybrid (S. invicta × S. richteri) ants were tested in this study. Colonies of S. richteri and hybrid fire ants were collected in northern Alabama and western Tennessee in spring 2008. Colonies of S. invicta were collected in Auburn, Alabama, approximately 200–300 miles (322–483 km) from S. richteri and hybrid collection sites. All colonies were considered monogyne based on colony size, worker size, and presence of only one queen during collection (Morel et al., 1990). Ant workers were collected by transferring approximately 1 litre of soil (containing workers, broods and the queen) from each mound into 1-gallon (3.785-litre) Rubbermaid plastic jars coated with Fluon® (ICI, Wilmington, DE) to prevent escape. Workers collected from each colony (mound) ranged from 1,000 to 5,000 and were maintained in their nest soil in the laboratory at 25 ± 1 °C, 50 ± 10% relative humidity, on a LD 14:10 h cycle and fed 10% sugar water and crickets. The colonies were maintained for 1–2 months after collection before the tests. Major workers from each colony were used for aggression tests, as they are the workers likely to be involved in foraging and colony defence (Moffett, 1985). The average weight of the major workers of S. invicta, hybrid, and S. richteri used for the tests were 212.9 ± 4.7, 212.6 ± 5.7, and 210.3 ± 6.1 mg per 50 workers (five replicates per form), respectively.
Species (form) identification by cuticular hydrocarbon profiling
Colonies were separated and identified as S. invicta, S. richteri, and hybrid by analysis of worker venom alkaloids and cuticular hydrocarbons profiles using gas chromatography (GC). This is the standard method for identification of Solenopsis fire ants (Vander Meer et al., 1985; Ross et al., 1987a, b). Briefly, approximately 50 workers from each colony were soaked in hexane for 24 h to extract their chemical compounds. GC analyses were performed on a Shimadzu GC17A equipped with a Rtx®-1MS column (30 m × 0.25 mm i.d., 0.25 μm, Restek, Bellefonte, PA). The oven was programmed at 15 °C/min from 90 °C to 270 °C with a 2-min initial time and a 16-min final holding time.
Aggression bioassay protocols
The bioassay used in this study measured aggression by nested (resident) worker ants against non-nestmate intruders of the same (intra-specific aggression) or different (inter-specific aggression) species/forms. The bioassay consisted of a white plastic cup (500 ml) coated inside with Fluon to prevent ant escape. Worker ants were individually placed in the bottom of the cup with a soft forcep. The workers which were first placed in the cup and allowed to acclimate for 30 min were considered resident ants, whereas those placed secondarily in the cup were regarded as intruders. Intruder workers from different colonies of the same form (intra-specific aggression) or a different form (inter-specific aggression) were placed in an empty space in the bottom of the cup.
To differentiate between intruders and resident ants, intruder workers were marked by placing a small dot of non-toxic white acrylic paint on their abdomen (van Wilgenburg et al., 2007). Intruder ants were first chilled in a container at 4 °C for 10 min to reduce activity, marked, and allowed to acclimate to ambient conditions for 30 min before they were introduced with a soft forcep into the cup containing resident ants. No effect of marking or chilling was observed on the behaviour and survival of ant workers in this study (see results of intra-specific aggression tests). The reaction of resident ants to intruders (and vice versa) was observed for 10 min. The ants were then kept together in the same cup for 24 h at the end of which the survival of both intruder and resident ants was recorded.
Aggressive responses of resident ants to intruders were scored on a scale of 0–4 similar to Suarez et al. (1999) and Tsutsui et al. (2000) in which 0 = ignore (contacts between resident and intruder ants in which neither showed any interest, i.e. no antennation or aggression), 1 = touch (contacts with antennation), 2 = avoid (contacts that resulted in intruder and/or resident ants retreating in opposite directions), 3 = aggression (including lunging, and brief bouts of biting or pulling), and 4 = fighting (prolonged aggression, and often consisted of one or both opposing ants locking their mandibles onto a body part of the other, carrying the other with its mandibles, or grappling). Approximately 1 h after the test, a 1-cm3 sugar-agar block (10% sugar water + 1% agar after boiling and chill) was placed in the centre of the bottom of the cup to allow feeding. The survival of the ants was recorded 24 h after the test. The above protocol was used to test for intra-specific and inter-specific aggression in colonies of red, black, and hybrid fire ants.
A pilot test was first conducted which showed that workers of the three ant forms displayed no aggression towards nestmates. The results of this pilot test also confirmed very high survival rates (>98% survival after 24 h) for workers which did not engage in aggression. For non-nestmate intra-specific aggression tests, three colonies of S. invicta were tested in six different pairwise combinations of residents and intruders (10 replicates per pairwise combination × 6 combinations for a total of 60 replicates). For S. richteri and hybrid ants, six colonies of each were tested in six different pairwise combinations of residents and intruders (10 replicates per pairwise combination × 6 combinations for a total of 60 replicates). Based on a recent study which reported an effect of group size on aggression in ants (Tanner, 2006), three group sizes of competitor numbers (battles) were tested in all aggression bioassays: 1-1, 5-1, and 5-5 resident–intruder battles.
For treatments using multiple individuals as residents, the response of all residents (n = 5) within each cup (replicate or set) was averaged and used to score aggression for that set. Similarly, the response of all intruders in the 5-5 battles was averaged. Since preliminary data analysis showed no significant effect of colony on aggression, mean aggression score for each form was calculated by pooling the data for the six pairwise (inter-colony) combinations (n = 60 replicates). Data from this and the second experiment (below) were not normally distributed and transformations did not adequately correct this anomaly. Thus, these data were analysed using non-parametric statistics. Data for each resident–intruder battles were first analysed separately and significant differences in intra-specific aggression among the three forms were established using the Kruskal–Wallis test (P < 0.05; JMP® 7.0.1, SAS Institute, 2007). In this and other data analysed with the Kruskal–Wallis test, designations of significantly different pairs of treatment means were made based on concensus between the Tukey–Kramer HSD (honestly significant difference) parametric test and a simultaneous test using the non-parametric Wilcoxon signed-rank test for each paired comparison, similar to the procedure used by Obin and Vander Meer (1989b). Only the treatment pairs shown to be significant in both tests were designated significant in the results section. For each ant form, aggression scores were also compared among the three resident–intruder battles using the same test. Survival data were generated for resident and intruder ants only for experiments using the 1-1 and 5-5 resident–intruder battles, because both battles have been reported to produce consistent results (Roulston et al., 2003). Percentage survival for the 5-5 battles was calculated by averaging the mortalities of individual ants in the cups. Survival data for residents or intruders were compared among the three forms using the Kruskal-Wallis test (P < 0.05). Significant differences between the survival of resident versus intruder workers in each battle were established using the Matched pairs test (P< 0.05; JMP® 7.0.1, SAS Institute, 2007).
For inter-specific aggression tests, the three ant forms were tested in all pairwise combinations of residents and intruders at the resident–intruder battles of 1-1, 5-1, and 5-5 (six pairwise combinations for each battles) using the procedure described above for intra-specific aggression. Each pairwise combination treatment (set) was replicated 10 times using ant workers from the same colony. The experiment was repeated three times (three sets) each time with a different colony of each ant form (i.e. 10 replicates per colony per pairwise combination × 3 colonies for a total of 30 replicates). Analysis of the data by colony showed no significant colony effect. Thus, data obtained for the different colonies (3 per form) were pooled for further analysis. Mean aggression scores for the different battles were calculated for each form by averaging the scores for the 30 replicates. Differences in inter-specific aggression among the six pairwise combinations were established for each battle using the Kruskal–Wallis test (P < 0.05). As an additional measure of aggressiveness, the percentage of attacks initiated by intruder ants (at the 1-1 battles) was calculated and compared between the two pairwise combinations of the same two ant forms using the Wilcoxon signed-rank test (P < 0.05; JMP® 7.0.1, SAS Institute, 2007). As in the intra-specific aggression test, survival data were generated for resident and intruder ants only for the experiments using 1-1 and 5-5 battles. Significant differences between the survival of residents versus intruders in each pairwise combination of the different forms were established using the Matched pairs test (P< 0.05; JMP® 7.0.1, SAS Institute, 2007).
Species (form) identification by cuticular hydrocarbon profiling
Gas chromatographic traces of whole body extracts of the three ant forms showing their typical alkaloid and cuticular hydrocarbon profiles are shown in Fig. 1. Note that the hybrid profile is intermediate between its parental forms. These profiles are consistent with published profiles for the three forms (Vander Meer, 1986), and have been used by various authors as a taxonomic character for identification of Solenopsis fire ants (Vander Meer et al., 1985; Ross et al., 1987a, b). No notable differences were detected in the cuticular hydrocarbon profiles of conspecific fire ants from different colonies, suggesting that this character is conserved within each form.
Solenopsis invicta, S. richteri and their hybrids showed no aggression towards nestmates, but did display considerable aggression towards non-nestmate conspecifics. Significant differences were recorded among the three forms in the level of intra-specific (non-nestmate) aggression (Fig. 2). Intra-specific aggression ratings for S. invicta were significantly greater than ratings for hybrids and S. richteri at the 1-1 (χ22 = 57.21, P < 0.0001) and 5-5 (χ22 = 80.72, P < 0.0001) battles. At the 5-1 battles, intra-specific aggression was highest for hybrid ants followed by S. invicta (χ22 = 54.48, P < 0.0001). In all cases, the lowest intra-specific aggression level was recorded for S. richteri (Fig. 2). No significant effect of group size (battles) was recorded for S. invicta (χ22 = 0.44, P = 0.80). In contrast, signficantly higher aggression scores were recorded for S. richteri at the 5-1 and 5-5 battles compared with the 1-1 battles (χ22 = 19.16, P < 0.0001), possibly suggesting an effect of group size on aggression. Similarly, higher aggressive responses (χ22 = 40.44, P < 0.0001) were recorded for hybrids at the 5-1 than at the 1-1 and 5-5 battles (Fig. 2).
In order to determine the impact of intra-specific aggression on survival of the three ant forms, the percentage of resident or intruder ants that were alive 24 h after aggression tests were compared among the three forms at the 1-1 and 5-5 battles. The percentages of alive resident ants were significantly higher for S. richteri than for S. invicta and hybrids at both the 1-1 (χ22 = 23.19, P < 0.0001) and 5-5 (χ22 = 52.57, P < 0.0001) battles (Fig. 3). Similar results were recorded among intruder ants of the three forms: percentage of intruder ants alive after 24 h was significantly greater for S. richteri at both battles (1-1 battles: χ22 = 50.91, P < 0.0001; 5-5 battles: χ22 = 56.63, P < 0.0001) (Fig. 3). These results showing that S. richteri suffered the least mortality 24 h after intra-specific encounters, are not surprising given the comparatively lower intra-specific aggression ratings recorded for this form. No significant effect of group size (battles) was recorded on the survival of S. invicta and hybrids. However, lower survival rates were recorded for S. richteri at the 5-5 compared with the 1-1 battles (Fig. 3). In general, no significant differences were recorded in the survival of resident versus intruder workers at the 1-1 and 5-5 battles. However, the survival of resident workers was greater than that of intruders at the 5-1 battles, which was not surprising.
Significant differences were recorded in inter-specific aggression ratings among the six pairwise combinations of S. invicta (R for red), S. richteri (B for black), and hybrid (H) ants at the 5-5 battles (χ25 = 84.92, P < 0.0001). The highest inter-specific aggression ratings were recorded for the two pairwise combinations of S. invicta and S. richteri (R–B and B–R pairwise combinations), suggesting intense competition between both forms (Fig. 4b). In contrast, the least inter-specific aggression ratings were recorded for the two pairwise combinations of S. richteri and hybrids (B–H and H–B pairwise combinations) (Fig. 4b). However, these differences were not significant at the 1-1 battles (χ25 = 0.37, P = 0.99; Fig. 4a). Further analysis of the aggression data for the 1-1 battles by comparing the percentage of attacks initiated by intruder ants between the two pairwise combinations of the same two ant forms, revealed some important differences (Fig. 5). For the pairwise combinations of S. invicta and S. richteri (R–B versus B–R), the percentage of attacks initiated by intruder ants was significantly higher when S. invicta workers were the intruders than when S. richteri were the intruders (χ21 = 3.97, P = 0.04; Fig. 5). However, no such differences were recorded in the interactions between S. invicta and hybrids (R–H versus H–R) (χ21 = 0.43, P = 0.51; Fig. 5) or between S. richteri and hybrids (B–H versus H–B) (χ21 = 0.006, P = 0.94; Fig. 5). These results further confirmed that S. invicta workers were generally more aggressive towards S. richteri workers than they were to hybrid workers. In general, the levels of inter-specific aggression recorded among the three forms were similar to their levels of intra-specific aggression.
Significant differences in the post-aggression survival of residents versus intruders were recorded in the pairwise combinations of S. invicta versus hybrids, and S. richteri versus hybrids at the 1-1 and 5-5 battles (Fig. 6a,b). In the interactions between S. invicta and hybrid ants, significantly lower survivals were recorded for hybrid ants, irrespective of whether they were the residents (at 1-1 battles: t = 3.61, P= 0.001; at 5-5 battles: t = 4.58, P< 0.0001) or the intruders (at 1-1 battles: t = 3.01, P= 0.004; at 5-5 battles: t= 8.97, P< 0.0001) (Fig. 6a,b). Significant differences in survival of residents versus intruders were also recorded in the 5-5 interactions between S. invicta versus S. richteri. When S. invicta was the resident, survival of S. richteri intruders was significantly lower than survival of S. invicta residents (Fig. 6b).
The results of the intra-specific and inter-specific aggression bioassays showed that S. invicta was the most aggressive of the three forms, whereas S. richteri was the least aggressive. Workers of S. invicta were not only highly aggressive towards their congeners, but also displayed high aggressive responses towards non-nestmate conspecifics, suggesting a low aggression threshold. These results were confirmed by our data showing significantly lower post-intra-specific aggression mortality for S. richteri than for S. invicta or hybrids: lower aggression resulted in higher survival. Data from the inter-specific aggression tests showed intense aggression between S. invicta and S. richteri with most of the attacks initiated by the former, irrespective of which form the intruder was, and may suggest that S. invicta workers are generally more aggressive toward S. richteri workers than they are to hybrid workers. Alternatively, S. richteri workers may have a greater tendency than hybrids to escalate interactions with S. invicta. The data also showed that in the interactions between S. invicta and S. richteri or hybrids, S. invicta workers suffered the least mortality as intruders or residents.
Our results are consistent with those reported by Jones and Phillips (1987). The authors compared the aggressive behaviour of S. invicta and three native ant species and reported that S. invicta was more aggressive than two of the native species. Similarly, S. invicta also showed greater aggressive interactions than the Argentine ant, Linepithema humile (Mayr) (Kabashima et al., 2007). However, our results are not consistent with those from an earlier study by Obin and Vander Meer (1989b) in which no differences were detected among S. invicta, S. richteri and their hybrids in aggression towards heterospecific intruders. The differences between our data and those of Obin and Vander Meer (1989b) may be as a result of several factors including differences in ant populations and handling protocols. For instance, the ant colonies tested by Obin and Vander Meer (1989b) were obtained from variable sources: S. invicta workers were reared from newly mated queens, whereas workers of S. richteri and the hybrid were field collected. Second, the colonies had been maintained in the laboratory 1–2 years before the tests. In contrast, monogyne colonies of the three forms collected around the same time and held in the laboratory for a similar length of time before testing, were tested in the present study.
Many authors have reported on the use of cuticular hydrocarbons as nestmate recognition cues by ants (Provost et al., 1994; Lahav et al., 1999; Lenoir et al., 1999; Liang & Silverman, 2000; Wagner et al., 2000; Akino et al., 2004; Ozaki et al., 2005; Vasquez et al., 2008). While this study was not designed to identify nestmate recognition cues, our data suggest that cuticular hybrocarbons are unlikely to be the sole or main nestmate recognition cue in imported fire ants. This supposition is based on our data which showed that all three fire ant forms displayed no aggression towards nestmates, but high aggression towards non-nestmate conspecifics. However, we did not detect significant variations in the cuticular hydrocarbon profiles of conspecific fire ants from different colonies, a finding in agreement with a previous report that cuticular hydrocarbon profile was a poor predictor of aggressive behaviour in a laboratory colony of S. invicta (Obin, 1986). Nevertheless, these results must be interpreted with caution in the light of a few recent papers, which suggest that nestmate discrimination in ants may be mediated by even subtle differences in the ratios of saturated and unsaturated components of their cuticular hydrocarbons (Akino et al., 2004; Vasquez et al., 2008). It is plausible that nestmate recognition in imported fire ants is mediated (either solely or in combination with cuticular hydrocarbons) by non-cuticular hydrocarbon contact chemicals possibly from the nest, which are shared by nestmates, but which vary between colonies and species (Wilson, 1971; Hölldobler & Michener, 1980; Vander Meer, 1988; Howard, 1992). Further studies are necessary to determine the cues mediating nestmate discrimination in imported fire ants.
In general, similar aggression and survival rates were recorded in the 1-1 and 5-5 resident–intruder battles, although differences between forms and residents versus intruders were slightly amplified in the 5-5 battles. These results are consistent with those of a recent paper on the effect of bioassay on aggressive behaviour in L. humile, which showed that the live 1-1 and 5-5 battles correlated with each other and provided consistent results (Roulston et al., 2003). However, the 5-1 battles offered an advantage to resident ants and may not represent a fair measure of aggression. The lower survival recorded for S. richteri at the 5-5 compared with the 1-1 battles was a direct reflection of their relatively higher aggression score at the 5-5 battles, and may suggest that this form displayed higher aggression as a group than as individuals. A similar result was reported for the ant, Formica xerophila Smith, whose workers acted more aggressively towards a competitor when in a large group than as isolated individuals (Tanner, 2006). The author reported in a follow-up study that the correlation between group size and aggressive behaviour in F. xerophila is coordinated by direct nestmate contact (Tanner, 2008).
No significant size differences were recorded among the major workers of the three ant forms, suggesting that the superior aggressiveness of S. invicta and hybrids is not because of their larger size. The data which showed relatively lower intra-specific aggression for S. richteri are consistent with a previous report of lower genetic diversity in this form compared with S. invicta and hybrids (Ross et al., 1987a, b). Higher genetic variability may thus explain the higher intra-specific aggression recorded in the present study for S. invicta and hybrids. Our data showed that the hybrid is intermediate between both parental forms in aggressiveness and post-aggression survival, and thus is not entirely in consonant with the ‘hybrid superiority’ theory proposed by Moore (1977).
Although the results of this laboratory aggression study may not entirely reflect field situations, they may provide insights into the interactions between S. richteri and S. invicta in the early years after their introductions into Alabama in the early 1900s (Wilson, 1958). The aggressiveness and superior combative ability of S. invicta and hybrids may contribute, at least in part, to the continuing range expansion of both forms in southern U.S. In contrast, the relatively lower aggression and competitive ability of S. richteri may, in part, explain the range restriction and dwindling populations of this form, which is now limited to a relatively small area in northwest Alabama, northern Mississippi and southern Tennessee (Streett et al., 2006). It has been proposed that inter-specific competition in ants may be explained by Lanchester’s laws of combat (Lanchester, 1916), which describe casualty rates in two opposing armies as functions of the numbers of individuals and their fighting abilities (McGlynn, 2000). However, a recent study by Plowes and Adams (2005) did not provide strong support for this proposition. The authors reported that causalities incurred by two fighting groups of S. invicta were not strongly affected by group sizes and that individual prowess was the main factor which determined the outcomes of the battles (Plowes & Adams, 2005). Other factors such as exploitative competition may also contribute to the ecological dominance of S. invicta, and the range restriction of S. richteri in southern U.S. In addition to inter-specific aggression (Holway, 1999; Rowles & O’Dowd, 2007), exploitative competition was reported as a major mechanism underlying the displacement of native ants by L. humile (Holway, 1999). However, the role of this mechanism in mediating shifts in the population densities and distributions of the three forms of imported fire ants in southern U.S. remains uninvestigated. Climatic requirements may also play a role: S. richteri has a more temperate distribution than S. invicta in its native South America (Shoemaker et al., 1996), and was found to be more cold-tolerant than S. invicta in a laboratory test (James et al., 2002). Climatic conditions in the native range of S. richteri are similar to those in northern Alabama, Mississippi and southern Tennessee, where this form is currently restricted in the U.S. The combination of climatic requirements, inferior aggression and competitive ability will make it difficult for S. richteri to compete with S. invicta in the warmer parts of their introduced range and may explain the current restricted range of S. richteri.
These results suggest that the hybrid form, which has an overlapping range with its parental forms in Alabama and Mississippi (Streett et al., 2006), is likely to continue to expand its range northward and southward and may pose a threat to the survival of S. richteri in southern U.S. Based on their recent data showing major shifts in spatial distributions of the three ant forms in Mississippi, Streett et al. (2006) predicted that S. richteri will eventually be replaced by S. invicta and hybrid populations in the state and perhaps in the entire country. While temperature and other climatic factors (e.g. isoclines) may prevent a total replacement of S. richteri by its congeners in the northernmost part of their range, these results suggest the likelihood of a more restricted and narrower range for S. richteri in the future.
We thank Clement Akotsen-Mensah for technical support. Drs. Fudd Graham, Robert Vander Meer and J. T. Vogt provided valuable review of an earlier version of the manuscript. This research was supported by the Alabama Fire Ant Management Program and the Alabama Agricultural Experiment Station. Xiaofang He was supported in part by a grant from the National Natural Science Foundation of China (No. 30571242).