Multiple virus infections occur in individual polygyne and monogyne Solenopsis invicta ants

Concurrent infections of Solenopsis invicta colonies with S. invicta virus 1 (SINV-1), SINV-2, and SINV-3 has been reported. However, whether individual ants were capable of supporting multiple virus infections simultaneously was not known, nor whether the social form of the colony (polygyne or monogyne) had an influence on the occurrence of multiple infection rates in individual ants. S. invicta field populations were sampled sequentially to establish whether multiple virus infections co-occurred in individual worker ants. In addition, the intra-colony virus infection rates were compared in monogyne and polygyne field colonies to determine whether social form played a role in the viral infection prevalence. All combinations of virus infection (SINV-1, SINV-2, or SINV-3 alone, SINV-1 & SINV-2, SINV-1 & SINV-3, SINV-2 & SINV-3, and SINV-1, SINV-2 & SINV-3) were detected in individual worker ants as well as queens in the field. Thus, individual S. invicta ants can be infected simultaneously with all combinations of the S. invicta viruses. Colony social form did have an influence on the intra-colony prevalence of multiple S. invicta virus infections. Polygyne colonies exhibited significantly greater intra- and inter-colony single and multiple virus infections compared with monogyne colonies.     

Infection characteristics of Solenopsis invicta virus 2 in the red imported fire ant, Solenopsis invicta

Solenopsis invicta virus 2 (SINV-2) is the second virus identified from the fire ant, S. invicta Buren. SINV-2 is unique among positive-strand RNA viruses from insects by possessing four cistrons in a monopartite genome. Fire ant colonies testing positive for SINV-2 by RT-PCR did not exhibit any discernable symptoms. RT-PCR-based surveys for SINV-2 among 688 fire ant mounds in Alachua County, Florida, sampled during the period January, 2006 through December, 2007 showed that the prevalence of SINV-2 among nests ranged from 1.6% to 16.4%. Unlike S. invicta virus 1, no seasonal-associated prevalence was observed with regard to SINV-2 infection among fire ant colonies. No social form specificity was evident; SINV-2 was found in both monogyne and polygyne S. invicta ants. Real-time quantitative PCR experiments showed that SINV-2 genome equivalents per individual ant ranged from 1.9 × 10⁷ in. pupae to 4.3 × 10¹¹ in. inseminated queens. The SINV-2 infection was detected in all ant stages examined (eggs, larvae, pupae, workers, and queens). Tissue tropism studies indicated that the alimentary canal (specifically the midgut) is most likely the susceptible tissue. SINV-2 was successfully transmitted to uninfected S. invicta ants by feeding a partially purified homogenate of SINV-2-infected ants. The SINV-2 transmission rate ranged from 30% to 80%, and both positive (genomic) and negative (replicative) SINV-2 RNA strands accumulated in recipient ants over the course of the experiment. These results indicated that SINV-2 replicates within S. invicta.     

Examination of host genome for the presence of integrated fragments of Solenopsis invicta virus 1

A series of oligonucleotide primer pairs covering the entire genome of Solenopsis invicta virus 1 (SINV-1) were used to probe the genome of its host, S. invicta, for integrated fragments of the viral genome. All of the oligonucleotide primer sets yielded amplicons of anticipated size from cDNA created from an RNA template from SINV-1. However, no corresponding amplification was observed when genomic DNA (from 32 colonies of S. invicta) was used as template for the PCR amplifications. Host DNA integrity was verified by amplification of an ant-specific gene, SiGSTS1. The representation of fire ant colonies included both social forms, monogyne and polygyne, and those infected and uninfected with SINV-1. Furthermore, no amplification was observed from genomic DNA from ant samples collected from Argentina or the US. Thus, it appears that SINV-1 genome integration, or a portion therein, has not likely occurred within the S. invicta host genome.     

Exploitation of a high genomic mutation rate in Solenopsis invicta virus 1 to infer demographic information about its host, Solenopsis invicta

The RNA-dependent RNA polymerase (RdRp) region of Solenopsis invicta virus 1 (SINV-1) was sequenced from 47 infected colonies of S. invicta, S. richteri, S. geminata, and S. invicta/richteri hybrids collected from across the USA, northern Argentina, and northern Taiwan in an attempt to infer demographic information about the recent S. invicta introduction into Taiwan by phylogenetic analysis. Nucleotide sequences were calculated to exhibit an overall identity of >90% between geographically-separated samples. A total of 171 nucleotide variable sites (representing 22.4% of the region amplified) were mapped across the SINV-1 RdRp alignment and no insertions or deletions were detected. Phylogenetic analysis at the nucleotide level revealed clustering of Argentinean sequences, distinct from the USA sequences. Moreover, the SINV-1 RdRp sequences derived from recently introduced populations of S. invicta from northern Taiwan resided within the multiple USA groupings implicating the USA as the source for the recent introduction of S. invicta into Taiwan. Examination of the amino acid alignment for the RdRp revealed sequence identity >98% with only nine amino acid changes observed. Seven of these changes occurred in less than 4.3% of samples, while 2 (at positions 1266 and 1285) were featured prominently. Changes at positions 1266 and 1285 accounted for 36.2% and 34.0% of the samples, respectively. Two distinct groups were observed based on the amino acid residue at position 1266, Threonine or Serine. In cases where this amino acid was a Threonine, 90% of these sequences possessed a corresponding Valine at position 1285; only 10% of the Threonine¹²⁶⁶-containing sequences possessed an Isoleucine at the 1285 position. Among the Serine¹²⁶⁶ group, 76% possessed an Isoleucine at position 1285, while only 24% possessed a Valine. Thus, it appears that the Threonine¹²⁶⁶/Valine¹²⁸⁵ and Serine¹²⁶⁶/Isoleucine¹²⁸⁵ combinations are predominant phenotypes.     

Detection and quantitation of Solenopsis invicta virus-2 genomic and intermediary replicating viral RNA in fire ant workers and larvae

Quantitative real-time PCR (QPCR) was used to quantify the genome of Solenopsis invicta virus-2 (SINV-2) from infected individual ants of S. invicta. Strand-specific cDNA synthesis oligonucleotide primers and RNase digestion after cDNA synthesis allowed quantification of plus (genomic) and minus (replicative) strands of the SINV-2 genome. Both strands were detected in adult workers and larval fire ants indicating that the virus was replicating within the ant. The differences between the genomic to replicative strand ranged from 199-fold in larvae to 479-fold in workers with an average ratio of 339:1.     

Solenopsis invicta virus-1A (SINV-1A): distinct species or genotype of SINV-1?

We have cloned and sequenced a 2845 bp cDNA representing the 3'-end of either a new picorna-like virus species or genotype of Solenopsis invicta virus-1 (SINV-1). Analysis of the nucleotide sequence revealed 1 large open reading frame. The amino acid sequence of the translated open reading frame was most identical to structural proteins of SINV-1 (97%), followed by the Kashmir bee virus (KBV, 30%), and acute bee paralysis virus (ABPV, 29%). A PCR-based survey for SINV-1 and the new species or genotype (tentatively named S. invicta virus-1A, SINV-1A) using RNA extracts of S. invicta collected around Gainesville, Florida, revealed a mean colony infestation rate of 25% by SINV-1 and 55% by SINV-1A. Both SINV-1 and SINV-1A were found to co-infect 17.5% of the nests surveyed. Although the data preclude definitive species or genotype assignment, there is no doubt that SINV-1A is distinct from SINV-1, identifiable, and infects S. invicta. We provide a simple RT-PCR technique capable of discerning SINV-1 and SINV-1A infection of S. invicta.     

Phenology, distribution, and host specificity of Solenopsis invicta virus-1

Studies were conducted to examine the phenology, geographic distribution, and host specificity of the Solenopsis invicta virus-1 (SINV-1). Two genotypes examined, SINV-1 and -1A, exhibited similar seasonal prevalence patterns. Infection rates among colonies of S. invicta in Gainesville, Florida, were lowest from early winter (December) to early spring (April) increasing rapidly in late spring (May) and remaining high through August before declining again in the fall (September/October). Correlation analysis revealed a significant relationship between mean monthly temperature and SINV-1 (p<0.0005, r=0.82) and SINV-1A (p<0.0001, r=0.86) infection rates in S. invicta colonies. SINV-1 was widely distributed among S. invicta populations. The virus was detected in S. invicta from Argentina and from all U.S. states examined, with the exception of New Mexico. SINV-1 and -1A were also detected in other Solenopsis species. SINV-1 was detected in Solenopsis richteri and the S. invicta/richteri hybrid collected from northern Alabama and Solenopsis geminata from Florida. SINV-1A was detected in S. geminata and Solenopsis carolinensis in Florida and the S. invicta/richteri hybrid in Alabama. Of the 1989 arthropods collected from 6 pitfall trap experiments from Gainesville and Williston, Florida, none except S. invicta tested positive for SINV-1 or SINV-1A. SINV-1 did not appear to infect or replicate within Sf9 or Dm-2 cells in vitro. The number of SINV-1 genome copies did not significantly increase over the course of the experiment, nor were any cytopathic effects observed. Phylogenetic analyses of SINV-1/-1A nucleotide sequences indicated significant divergence between viruses collected from Argentina and the U.S.     

Solenopsis invicta virus-1 tissue tropism and intra-colony infection rate in the red imported fire ant: a quantitative PCR-based study

Quantitative real-time PCR was employed to measure the Solenopsis invicta virus 1 (SINV-1) load in tissues, individuals, and among colonies of the red imported fire ant, S. invicta Buren. Among tissues examined from SINV-1-infected adults and larvae, the alimentary canal (specifically the midgut) consistently had the highest number of SINV-1 genome copies (91.1 and 99.9%, respectively). Negative staining of a supernatant of the gut homogenate demonstrated the presence of spherical virus particles with a diameter of 30-35 nm, consistent with SINV-1. The number of SINV-1 genome copies in infected larvae and workers from the same queenright colonies were similar to each other. In other words, the infection rate was consistent among both developmental stages. No significant differences were observed in SINV-1 genome copy number among infected colonies sampled during the winter and summer. Although the SINV-1 infection rate of summer-collected mounds was previously shown to be six-times higher than winter-collected mounds, the intra-colony infection rate appears to be unaffected by season. Perhaps less inter-mound interaction during the winter months among S. invicta restricts spread of the virus. A positive correlation between intra-colony infection rate and mean SINV-1 genome copy number per ant was also observed. Based on these results, it is likely that SINV-1 replicates in gut epithelia of S. invicta and virus is shed into the gut lumen where it may be transmitted to nestmates by trophallaxis.     

Expressed sequence tags from the red imported fire ant, Solenopsis invicta: annotation and utilization for discovery of viruses

An expression library was created and 2304 clones sequenced from a monogyne colony of Solenopsis invicta. The primary intention of the project was to utilize homologous gene identification to facilitate discovery of viruses infecting this ant pest that could potentially be used in pest management. Additional genes were identified from the ant host and associated pathogens that serve as an important resource for studying these organisms. After assembly and removal of mitochondrial and poor quality sequences, 1054 unique sequences were yielded and deposited into the GenBank database under Accession Nos. EH412746 through EH413799. At least nine expressed sequence tags (ESTs) were identified as possessing microsatellite motifs and 15 ESTs exhibited significant homology with microsporidian genes. These sequences most likely originated from Thelohania solenopsae, a well-characterized microsporidian that infects S. invicta. Six ESTs exhibited significant homology with single-stranded RNA viruses (3B4, 3F6, 11F1, 12G12, 14D5, and 24C10). Subsequent analysis of these putative viral ESTs revealed that 3B4 was most likely a ribosomal gene of S. invicta, 11F1 was a single-stranded RNA (ssRNA) virus contaminant introduced into the colony from the cricket food source, 12G12 appeared to be a plant-infecting tenuivirus also introduced into the colony as a field contaminant, and 3F6, 14D5, and 24C10 were all from a unique ssRNA virus found to infect S. invicta. The sequencing project illustrates the utility of this method for discovery of viruses and pathogens that may otherwise go undiscovered.     

Molecular diversity of the microsporidium Kneallhazia solenopsae reveals an expanded host range among fire ants in North America

Kneallhazia solenopsae is a pathogenic microsporidium that infects the fire ants Solenopsis invicta and Solenopsis richteri in South America and the USA. In this study, we analyzed the prevalence and molecular diversity of K. solenopsae in fire ants from North and South America. We report the first empirical evidence of K. solenopsae infections in the tropical fire ant, Solenopsis geminata, and S. geminata × Solenopsis xyloni hybrids, revealing an expanded host range for this microsporidium. We also analyzed the molecular diversity at the 16S ribosomal RNA gene in K. solenopsae from the ant hosts S.invicta, S. richteri, S. geminata and S. geminata × S. xyloni hybrids from North America, Argentina and Brazil. We found 22 16S haplotypes. One of these haplotypes (WD_1) appears to be widely distributed, and is found in S. invicta from the USA and S. geminata from southern Mexico. Phylogenetic analyses of 16S sequences revealed that K. solenopsae haplotypes fall into one of two major clades that are differentiated by 2-3%. In some cases, multiple K. solenopsae haplotypes per colony were found, suggesting either an incomplete homogenization among gene copies within the 16S gene cluster or multiple K. solenopsae variants simultaneously infecting host colonies.     

Loss of microbial (pathogen) infections associated with recent invasions of the red imported fire ant Solenopsis invicta

Loss of natural enemies during colonization is a prominent hypothesis explaining enhanced performance of invasive species in introduced areas. Numerous studies have tested this enemy release hypothesis in a wide range of taxa but few studies have focused on invasive ants. We conducted extensive surveys for the presence of six microbes in recently established populations (California, Australia, New Zealand, Taiwan, and China) of the invasive fire ant Solenopsis invicta. These microbes include Wolbachia, two microsporidia (Kneallhazia solenopsae and Vairimorpha invictae) and three RNA viruses (SINV-1, -2 and -3), all of which previously have been reported in native South American populations of S. invicta. These surveys showed that the total number of enemy species is lower in the recently invaded areas compared with both South American and US populations. Only two microbes were found in any of these recently invaded areas: SINV-1 was detected in all surveyed populations except Australia and New Zealand, and SINV-2 was detected in California and Taiwan only. These results support the general prediction that invasive species lose many of their natural enemies during invasion. Further, the conspicuous absence of some of these microbes in these areas may result from strong selection against founders due to fitness costs associated with harboring detrimental infections rather than the alternative hypothesis that they simply were absent among the original founders. While the successful invasion of S. invicta in these recently invaded areas may be explained partly by the absence of natural enemies, other factors likely have been important as well.     

Culture-independent identification of gut bacteria in fourth-instar red imported fire ant, Solenopsis invicta Buren, larvae

Red imported fire ants (RIFA), Solenopsis invicta Buren, are medical, urban, and agricultural pests from South America. They are successful invaders due to their preference for disturbed habitats, high reproductive rates, and the ability to feed on a wide variety of food items (omnivorous). Fourth-instar larvae are used by the colony to digest solid food and then regurgitate it for consumption by workers and queens. Larvae are an ideal source of investigations of endosymbiotic bacteria possibly involved in nutrient distributions. Our study utilized 16S rDNA sequencing to describe the composition of the bacterial community in fourth-instar ant larvae in order to identify possible endosymbiotic bacteria present therein. The 16S rRNA gene was directly amplified from mixed-population DNA of whole fire ant larval guts and cloned into Escherichia coli. Bacterial communities from three geographically separated RIFA colonies were examined. Sequenced bacterial clones from guts were determined to be predominantly from the phylum Proteobacteria and the family Enterobacteriaceae. Our results did not detect the presence of endosymbiotic bacteria in the guts of RIFA larvae among the colonies. In addition, minimal species overlap was found when bacterial inventories were compared among colonies. Thus, bacteria coadapted with red imported fire ant larvae were not detected. Identified bacteria were not closely affiliated with endosymbiotic bacteria common in other insect species. Bacteria communities appeared to be unique to each geographical location and were determined by the foods consumed by the ants.     

Phorid fly, Pseudacteon tricuspis, response to alkylpyrazine analogs of a fire ant, Solenopsis invicta, alarm pheromone

The phorid fly, Pseudacteon tricuspis Borgmeier, is a parasitoid of the red imported fire ant, Solenopsis invicta Buren. This fly has been reported to use fire ant chemicals, specifically venom alkaloids and possibly alarm pheromone to locate its host. A recent study identified 2-ethyl-3,6-dimethyl pyrazine as a component of the alarm pheromone of S. invicta. To determine the possible involvement of this fire ant alarm pheromone component in mediating fire ant-phorid fly interactions, we tested electroantennogram (EAG) and behavioral responses of P. tricuspis females to the commercially available mixture of 2-ethyl-3,6-dimethyl pyrazine and its 3,5-dimethyl isomer, as well as six structurally related alkylpyrazine analogs at varying doses. Pseudacteon tricuspis females showed significant EAG response to 2-ethyl-3,6(or 5)-dimethyl pyrazine (herein referred to as pheromone-isomer) at all doses, 0.001-10 μg. Among the tested alkylpyrazine analogs, 2,3-diethyl-5-methyl pyrazine showed significant EAG activity at 0.1 and 1 μg. 2,3-dimethyl pyrazine also showed significant EAG activity at 0.1 μg. Results of four-choice olfactometer bioassays demonstrated significant attraction of P. tricuspis females to the pheromone-isomer (2-ethyl-3,6(or 5)-dimethyl pyrazine) at all tested doses (0.01, 0.1, 1 and 10 μg). The analogs, 2,3-diethyl-5-methyl pyrazine and 2,3-dimethyl pyrazine were significantly better than the control at the higher doses (0.1, 1 and 10 μg). The pheromone-isomer was significantly better than both analogs at two doses, 0.1 and 1 μg. These results confirm that the reported fire ant alarm pheromone component plays a role in mediating attraction of phorid flies to host workers. Venom alkaloids were previously shown to attract P. tricuspis; therefore, we propose that fire ant alarm pheromones may act in tandem or synergistically with venom alkaloids to attract phorid fly parasitoids to fire ant workers.     

Fire ant alarm pheromone and venom alkaloids act in concert to attract parasitic phorid flies, Pseudacteon spp.

Pseudacteon tricuspis, Pseudacteon obtusus and Pseudacteon curvatus are three species of parasitic phorid flies (Diptera: Phoridae), which have been introduced as classical biological control agents of imported, Solenopsis fire ants (Hymenoptera: Formicidae) in the southern USA. Previous studies demonstrated the behavioral response of P. tricuspis to the venom alkaloids and alarm pheromone of the fire ant, S. invicta. In the present study, we compared the responses of P. tricuspis, P. obtusus and P. curvatus to Solenopsis invicta alarm pheromone, venom alkaloids, or a mixture of both chemicals in four-choice olfactometer bioassays. The main hypothesis tested was that the fire ant alarm pheromone and venom alkaloids act in concert to attract Pseudacteon phorid flies. Both sexes of all three Pseudacteon species were attracted to low doses of the fire ant alarm pheromone or venom alkaloids (i.e. 1 ant worker equivalent) alone. However, the flies were significantly more attracted to a mixture of both chemicals (i.e., 1:1 mixture of alarm pheromone + alkaloids) than to either chemical. The results suggest an additive rather than a synergistic effect of combining both chemicals. Comparing the fly species, P. tricuspis showed relatively greater attraction to cis alkaloids, whereas the alkaloid mixture (cis + trans) was preferred by P. obtusus and P. curvatus. In general, no key sexual differences were recorded, although females of P. tricuspis and P. obtusus showed slightly higher response than conspecific males to lower doses of the alarm pheromone. The ecological significance of these findings is discussed, and a host location model is proposed for parasitic phorid flies involving the use of fire ant alarm pheromone and venom alkaloids as long range and short range attractants, respectively.     

Evaluating biological control of fire ants using phorid flies: effects on competitive interactions

The effects of the parasitic phorid fly, Pseudacteon tricuspis Borgmeier, on the competitive interactions between the red imported fire ant, Solenopsis invicta Buren, and a native North American ant, Forelius pruinosus (Roger), were investigated in the laboratory. P. tricuspis is a highly host-specific endoparasitoid of S. invicta workers that is currently being reared and released as a biological control agent of S. invicta in the US. We tested the effect of P. tricuspis on the colony growth rate of S. invicta when S. invicta was forced to compete with F. pruinosus for a protein resource (freeze-killed crickets) in laboratory competition arenas. In addition to colony growth rate, we quantified the effect of the phorid flies on the foraging rate of S. invicta. Though S. invicta significantly reduced its foraging rate in the presence of the phorid flies, we did not detect an effect of the flies on colony growth rate. Possible explanations for these results include behavioral compensation by S. invicta for the presence of the flies. We present these laboratory results in light of a literature search indicating that laboratory tests of biological control agent efficacy are good predictors of field efficacy. We conclude that P. tricuspis alone is unlikely to suppress S. invicta populations in the field by reducing their competitive ability.     

Phenology and parasitism rates in introduced populations of <Emphasis Type="Italic">Pseudacteon tricuspis</Emphasis>, a parasitoid of <Emphasis Type="Italic">Solenopsis invicta</Emphasis>

We documented patterns of seasonal abundance and rates of parasitism in introduced populations of Pseudacteon tricuspis Borgmeier, a phorid parasitoid of the red imported fire ant, Solenopsis invicta Buren. Adult P. tricuspis populations were censused at monthly intervals for 1 year at three sites in northern Florida. Censuses were conducted by aspirating phorids attracted to disturbed S. invicta mounds. Pseudacteon tricuspis adults were present in every month at all sites, although abundances varied greatly among sites and over time. The highest densities of flies (up to 453 censused at 10 disturbed S. invicta mounds in 30 min) were observed in November, and changes in abundance over time were positively correlated among sites. Sex ratios were usually male biased. Parasitism rates were evaluated by collecting workers from field colonies and monitoring them in the laboratory for evidence of parasitism. Parasitism rates were very low – always less than 1%. The average parasitism rate per colony over 16 colonies and 2 years was 0.058%. No pupariation occurred within the first 8 days of collection, suggesting parasitism by P. tricuspis induced behavioral changes in parasitized workers that precluded such workers from our collections. If so, true field parasitism rates may be several times higher than measured here, yet still low in an absolute sense. These low parasitism rates can be reconciled with observed adult phorid densities by considering the large number of host ants present at the study sites.     

Complete genome sequence of a novel picorna-like virus isolated from Spodoptera exigua

The complete genome sequence and the gene organization of a novel insect picorna-like virus, Spodoptera exigua virus (SeV), were determined. The genomic RNA of the SeV was 9501 nt in length excluding the poly(A) tail and contained a single, large open reading frame (nt 392–9424) encoding a 3010 aa polyprotein. Sequence comparisons with other viral polyproteins revealed that the consensus sequences for picornavirus RNA helicase, cysteine protease, and RNA-dependent RNA polymerase (RdRp) proteins are found on the genome in that order from the 5′ to the 3′ end. In terms of sequence similarity, identity, and genome organization, SeV resembled insect picorna-like viruses belonging to the genus Iflavirus. A phylogenetic analysis based on the eight conserved domains in the RdRp sequence showed that SeV was most closely related to the Perina nuda virus and Ectropis obliqua picorna-like virus, suggesting that these three insect picorna-like viruses might share a common ancestor.     

Defensiveness of the fire ant, Solenopsis invicta, is increased during colony rafting

Colonies of the fire ant, Solenopsis invicta, can survive flood conditions by forming a raft of ants that floats on the water’s surface until the flood recedes or higher ground is found. Having been forced from the protection of their subterranean nests, rafting colonies are totally exposed and are without retreat. I tested the hypothesis that rafting S. invicta colonies would compensate for their elevated vulnerability by increasing their defensiveness. I measured defensiveness using the amount of venom workers delivered per sting (venom dose), since the repellent effects (i.e., pain and tissue damage) of fire-ant venom are dose-dependent. In the laboratory I assayed colony defensiveness before and after flooding colonies from their nests with water. Colonies were consistently and significantly more defensive while rafting (i.e., each colony’s workers delivered higher venom doses when their colony was rafting than they did when it was assayed pre-flood). The larger venom doses of rafting colonies may reduce their chances of being damaged by encounters with other animals by reducing the duration of such encounters through increased repellency. Encounters with S. invicta during flood conditions have the potential to be unusually dangerous; large concentrations of workers are exposed and available for defense, and they deliver significantly larger venom doses when they sting. Received 29 March 2005; revised 20 June 2005; accepted 24 June 2005.     

Zombie fire ant workers: behavior controlled by decapitating fly parasitoids

Laboratory observations were conducted on four separate red imported fire ant, Solenopsis invicta, colonies that contained workers parasitized by the decapitating fly, Pseudacteon tricuspis. Parasitized S. invicta workers remained inside the nest during parasitoid larval development and left the nest approximately 8 – 10 hours before decapitation by the parasitoid. When parasitized ants left the nest, they were highly mobile, were responsive to tactile stimuli, and showed minimal defensive behavior. Ants ultimately entered into a grass thatch layer, where they were decapitated and the fly maggots pupariated. This study reveals that parasitized ants exhibit behaviors that are consistent with host manipulation to benefit survival of the parasitoid. Received 9 November 2006; revised 26 January 2007; accepted 7 February.