Spatio-Temporal Population Sampling of a Fire Ant Parasitoid Donald C

DOI: 10.1111/j.1570-7458.2008.00763.x

Blackwell Publishing Ltd Spatio-temporal population sampling of a fire ant parasitoid Donald C. Henne*, William S. Hilbun & Seth J. Johnson Department of Entomology, Louisiana State University Agricultural Center, 404 Life Sciences Building, Baton Rouge, LA 70803, USA Accepted: 18 June 2008

Key words: Pseudacteon tricuspis, Solenopsis invicta, Thelohania solenopsae, SADIE, red imported ﬁre ant, decapitating ﬂy, Phoridae, Diptera, Hymenoptera, Formicidae, Microsporidia, Thelohaniidae

Abstract The spatial and temporal population structure of Pseudacteon tricuspis Borgmeier (Diptera: Phori- dae), an introduced parasitoid of the invasive red imported fire ant, Solenopsis invicta Buren (Hymenoptera: Formicidae), was investigated at three locations in Louisiana, USA, during the fall of 2005. Georeferenced disturbed fire ant mounds that attracted adult P. tricuspis were analyzed using Spatial Analysis by Distance IndicEs to model the population structure of this parasitoid on a local spatial scale, and relate spatial abundance to host colony social form and infection with the fire ant pathogen, Thelohania solenopsae Knell, Allen and Hazard (Microsporidia: Thelohaniidae). Spatial associations between P. tricuspis counts and host S. invicta colonies infected with T. solenopsae were not detected. Overall, spatio-temporal count patterns of P. tricuspis were largely random. However, high P. tricuspis counts were significantly associated with areas of polygyne host colonies. Occasional aggregations of P. tricuspis populations into discrete spatial patches and gaps were also found. The implications of this study are that population surveys of P. tricuspis should be conducted in several areas at a location to account for local patchiness in fly populations and/or varying densities and social form of host ant colonies.

Analyzing counts of organisms that are based on tradi- Introduction tional methods of sample variance–mean relationships It is a trivial fact that populations of organisms occupy and derivatives thereof (see Taylor, 1984) only provide certain areas at certain times, and some species are more information on the numeric properties of the underlying mobile than others. Understanding the processes that frequency distribution, and therefore have limited affect the spatial distribution of plant and animal popu- capability to describe the essential spatial information of lations is an important focus in ecology (Tuda, 2007). counts (Perry & Hewitt, 1991; Perry et al., 1999). For example, For mobile animal species, spatial information is often a highly skewed series of counts (e.g., negative binomial) restricted to trap counts at speciﬁc locations (Perry, 1995). with one or more very large values relative to the others These count locations may be spatially arranged in regular, may nevertheless be completely random in space (Perry & random, or clustered patterns without regard to the properties Dixon, 2002). Current analyses of insect spatial and tem- of the underlying count frequency distribution (Ferguson poral distributions utilize location information of sample et al., 2000). Given that population sampling relies on units to detect and measure degree of non-randomness spatial dispersion patterns (Southwood, 1978), determining in the spatial pattern in two-dimensional space (Perry, 1998a) the spatial and temporal population structure of organisms and spatial association between data sets collected on is an important objective toward understanding their different occasions (Winder et al., 2001). This methodology population dynamics. However, the difﬁculty of directly has been applied to the study of several host–parasitoid studying movement of individual animals, particularly systems (e.g., Ferguson et al., 2000, 2006; Weaver et al., small and numerous species such as insects, presents 2005). Therefore, key insights into spatial dynamics of special problems for ecologists (Perry, 1998a). host–parasitoid systems, such as spatial aggregations of parasitoids and their hosts, may be achieved by incorpo- *Correspondence: E-mail: [email protected] rating spatial information of count data into analyses.

Fire ant parasitoid dynamics 133

The red imported fire ant, Solenopsis invicta Buren P. tricuspis to host social form and presence/absence of (Hymenoptera: Formicidae), is an exotic invasive insect in T. solenopsae. the USA and is regarded as a significant economic pest (Lofgren, 1986; Porter et al., 1992). Management of Materials and methods S. invicta in the USA includes biological control through importation of natural enemies from the indigenous range Study locations of S. invicta in South America, particularly parasitic flies The P. tricuspis populations sampled in this study of the genus Pseudacteon Coquillet (Diptera: Phoridae). originated from a release conducted 17 km northeast of These flies are solitary endoparasitoids of Solenopsis fire Covington, St. Tammany Parish, LA, USA (30°36′N, ants (Morrison & Porter, 2005) and adults of one intro- 90°01′W) from 8–13 September 1999 (2 165 flies released) duced Pseudacteon species, Pseudacteon tricuspis Borgmeier, (see Henne et al., 2007). Populations of P. tricuspis were are attracted to alarm pheromones emitted by S. invicta sampled weekly at three widely separated locations in (Morrison & King, 2004). Pseudacteon females insert a Washington Parish, LA, USA, between 21 September and single egg into the host ants’ thorax, and the maggot 19 October 2005. The GPS coordinates for each study migrates to the hosts’ head capsule, consumes the contents location were recorded with a MagellanTM GPS 315/320 of the head over a 2-week period, and eventually decapitates (Magellan, Santa Clara, CA, USA). The first study location the host (Porter, 1998). was 8 km south of Bogalusa (30°41′N, 89°53′W; hereafter Phorid fly population dynamics are not well-understood called Farm 2), and was approximately 70 × 70 m in an (Morrison, 2000). In particular, almost no information is unmaintained cattle pasture. The second study location available concerning the spatial and temporal dynamics of was 14 km southwest of Franklinton (30°48′N, 90°18′W; various Phoridae (Disney, 1994). Populations of Pseu- hereafter called Farm 4), and was approximately 40 × 60 m dacteon parasitoids of the tropical fire ant, Solenopsis geminata in a mowed cattle pasture. The third study location was (Fabricius), in central Texas were characterized as having 16 km west of Bogalusa (30°46′N, 90°01′W; hereafter significant variations in abundance, both spatially and called Farm 7), and was approximately 170 × 100 m in an temporally (Morrison et al., 1999). Morrison & King unmaintained cattle pasture. (2004) evaluated phorid abundance at disturbed S. invicta At each study site, all active S. invicta colonies (usually mounds at multiple sites in North Central Florida over 15–25) were located in an arbitrarily defined area and time and determined that abundance deviated signifi- permanently marked with a barcode to enable future cantly from a regular distribution. However, neither of identification. The barcode consisted of a high durability these studies evaluated phorid abundance at the same weather-resistant polyester barcode label adhered to a individual mounds over time nor modeled abundance in 10 × 5 cm piece of rigid plastic sheeting that was anchored a spatio-temporal context. Additionally, no studies have flush with the ground with 10-cm long wire staple. The evaluated potential spatial patterns of phorid flies in barcode label assigned a unique number to each S. invicta relation to host colony social forms (monogyne vs. polygyne) mound that was retained throughout the study phase. and/or infection with microsporidian disease. Two social A Symbol® MC 3000 batch mobile computer (Motorola, forms of S. invicta, monogyne (single-queen) and polygyne Holtsville, NY, USA) was used to scan and record each (multiple-queens) occur in the USA (Glancey et al., 1973). barcode’s unique value into a database. The two forms are regulated by a single gene that is Decimal-degree locations of S. invicta mounds were homozygous (Gp-9B) in monogyne queens and hetero- taken with a GPS to obtain x, y-coordinates of sampling zygous (Gp-9B and Gp-9b) in polygyne queens (Ross & locations. Generally, at least 10 m separated S. invicta Keller, 1998; Krieger & Ross, 2002). mounds. At each location, weekly mean soil moisture The microsporidian entomopathogen, Thelohania levels were obtained using a Lincoln soil moisture meter solenopsae Knell, Allen and Hazard (Microsporidia: (Forestry Suppliers Part No. 3052; Lincoln Irrigation, Thelohaniidae), has been isolated from polygynous Lincoln, NE, USA), driven into the soil to a depth of 10 cm. S. invicta colonies in both North and South America, Ten measurements at 5 m intervals were made along a causing significant mortality in infected colonies (Oi & transect through each study area. Williams, 2002). However, it is unknown if S. invicta social form or infection with T. solenopsae influence P. tricuspis Sampling Pseudacteon tricuspis spatial distributions. Therefore, the objectives of this study Sampling Pseudacteon parasitoids of Solenopsis involved were to (1) characterize the number of attracted P. tricuspis simply disturbing host nests and awaiting arrival of flies. to disturbed S. invicta mounds in a spatio-temporal way Two observers disturbed individual S. invicta mounds in at three study sites, and (2) relate the abundance of the following manner: mounds were vigorously disturbed

134 Henne et al.

> with spades (5–10 s) and ants were crushed to release regular pattern of counts, while Ia 1 indicates aggregation semiochemicals to attract P. tricuspis (Morrison & King, of observed counts into clusters. The probability, Pa, that 2004). Previous mean fly counts made by the two observers the observed data is more aggregated than expected from (Donald C. Henne and Seth J. Johnson) were not a random permutation of the counts is significant at significantly different (P>0.05) (DC Henne & SJ Johnson, P<0.05. Here, 39 permutations were performed in SADIE. unpubl.). The number of P. tricuspis adults that arrived at The P. tricuspis spatial count data were also analyzed disturbed mounds during the ensuing 5 min was counted. with SADIE to determine the degree of clustering. SADIE Fly surveys were conducted between 11:00 and 17:00 hours calculates an overall mean of the sampled population (CDT) and when ambient temperatures were warm enough and then assigns an index of clustering (v) to each sample for Pseudacteon fly activity (>20 °C) (Morrison et al., 1999). location. Sample locations that have counts greater than

the sample mean are classiﬁed as positive (vd) ‘donor’ units, Determination of host colony social form and presence of while sample locations that have counts less than the sample

Thelohania solenopsae mean are classiﬁed as negative (vr) ‘receiver’ units. Clusters

A sample of approximately 1 g of S. invicta workers was of count data can occur as either patches (vd), areas of obtained from each mound on 21 September 2005 by relatively high counts that are close together, or gaps (vr), plunging a 20-ml glass scintillation vial into the mound. areas of relatively low counts that are close together (Perry,

The vials were coated with Fluon® to keep S. invicta inside 1995). Expected clustering indices approach 1 (vd or vr) for > the vials. After collection, the vials were ﬁlled with 95% random counts, vd 1 for counts that belong to a patch, and < ethanol. Social form was determined by the multiplex vr –1 for counts that belong to a gap. To test for non- polymerase chain reaction (PCR) methods as described in randomness, the mean value of the clustering index over the

Valles & Porter (2003). Presence of T. solenopsae in the patch units (Vd) is compared with its expected value of 1. samples was determined by multiplex PCR methods as Similarly, the mean value of the clustering index over the described in Valles et al. (2002). gap units (Vr) is compared with its expected value of –1.

Significance levels for vd and vr are established through a Analysis of Pseudacteon tricuspis spatial distributions two-tailed test by the 95th percentiles of the randomized > < The spatial patterns of P. tricuspis counts were analyzed distributions, where vd 1.5 and vr –1.5 are considered using Spatial Analysis by Distance IndicEs (SADIE) significant at the 0.025 and 0.975 levels, respectively. (SADIEShell, version 1.22; Rothamsted Experimental Finally, spatially referenced data that share the same Station, Harpenden, Herts, UK; http://www.rothamsted. coordinates can be analyzed for spatial association (i.e., bbsrc.ac.uk/pie/sadie/SADIE_home_page_1.htm) (Perry, similarity in the spatial patterns of two data sets) (Perry 1995, 1998a,b). SADIE is a spatial analysis software et al., 1999; Perry & Dixon, 2002). SADIE can be used to program designed for use in situations where species are measure the time evolution of change in population struc- patchily distributed into discrete aggregations (Winder ture when the same species is sampled at the same spatial et al., 2001), and is appropriate to situations such as coordinates over time, and can detect spatial association counting numbers of adult P. tricuspis appearing at between two species (Winder et al., 2001; Perry & Dixon, disturbed S. invicta mounds. SADIE compares the degree 2002). To compare consecutive weekly spatial patterns of of spatial pattern in the observed count arrangement to the P. tricuspis, an overall spatial association index, Χ (upper minimum effort that individuals in that sample would case chi), was calculated by SADIE based on the similarity need to expend to move to a regular arrangement. This of local clustering indices (vi and vj above) from the is also called the distance to regularity (D) where consecutive weekly distributions (Perry, 1998b). Similar abundance would be equivalent in each sample unit (Perry association indices were computed for P. tricuspis vs. et al., 1999). Thus, spatially aggregated counts will have T. solenopsae and P. tricuspis vs. S. invicta social form (except higher values of D than counts that are regular (Perry & for Farm 7, which had only two T. solenopsae-infected Dixon, 2002). The spatial pattern is quantified by mounds and was almost entirely polygyne). Spatially randomly permuting the observed set of counts among the associated distributions will have values of X>0 for patch sample unit locations, thereby generating expected coinciding with a patch, a value of X = 0 for random, < distances to regularity (Ea) to test the null hypothesis that and values of X 0 or dissociated distributions for patch counts are randomly arranged with respect to one another coinciding with a gap. Significance of X is tested against (Perry et al., 1999). An overall aggregation index is a random X derived from a randomization procedure of = computed as Ia D/Ea for each sample data set to establish the clustering indices, and incorporates an adjustment = the dispersion pattern (Perry, 1995, 1998b). Values of Ia 1 for small-scale spatial autocorrelation in the data sets < indicate randomly arranged counts, Ia 1 indicates a (Dutilleul, 1993; Perry & Dixon, 2002). The null hypothesis Fire ant parasitoid dynamics 135

Figure 1 (A) Locations of Solenopsis invicta mounds infected with Thelohania solenopsae at Farm 2 (1 = presence, 0 = absence); (B) S. invicta social form (P = polygyne, M = monogyne); (C-G) time series of Pseudacteon tricuspis abundance at disturbed S. invicta mounds, September–October 2005 [C (21 September), D (28 September), E (5 October), F (12 October), and G (19 October)]. NA, no information.

is that there is no spatial association between the two sets Results of spatially referenced data. A t-test was also performed on abundances of P. tricuspis Host colony disease prevalence and Pseudacteon tricuspis in relation to host social form for each sample date Presence of T. solenopsae at Farms 2, 4, and 7 are mapped for Farms 2 and 4 only (S. invicta at Farm 7 was 95% in Figures 1A, 2A, and 3A, respectively. Few S. invicta polygyne). Spatial maps of P. tricuspis counts at individual mounds were found to be infected with T. solenopsae S. invicta mounds, S. invicta social form, and presence of (Farm 2, 25%; farm 4, 13%; and Farm 7, 10%). No T. solenopsae were generated using S-Plus™ 7.0 (Insightful, signiﬁcant spatial associations were found between Seattle, WA, USA). Contour maps showing signiﬁcant P. tricuspis abundance and mounds infected with gap and patch indices were generated using 3DField 2.9 T. solenopsae (two-tailed test: P>0.025 for positive and (http://3dfmaps.com © Vladimir Galouchko). P<0.975 for negative spatial associations). 136 Henne et al.

Table 1 Spatial associations between Pseudacteon tricuspis abundance vs. Solenopsis invicta social form and presence of

Thelohania solenopsae. Association index values (X) are given with their probability (Pa) in parentheses

Location Dates Social form association index value Thelohania association index value Farm 2 21/9 0.2726 (0.17) 0.5245 (0.04) 28/9 0.1206 (0.39) –0.0064 (0.38) 5/10 0.4544 (0.03) 0.1418 (0.29) 12/10 0.4776 (0.004) 0.5128 (0.04) 19/10 0.2757 (0.18) 0.2970 (0.14) Farm 4 21/9 0.1777 (0.35) 0.1126 (0.27) 28/9 0.4088 (0.09) –0.1921 (0.56) 5/10 0.1305 (0.37) 0.3402 (0.11) 12/10 0.4449 (0.07) 0.3051 (0.14) 19/10 0.2694 (0.23) 0.0767 (0.36) < Value in bold indicates signiﬁcance (PI 0.025 for positive spatial association.

Social form and Pseudacteon tricuspis tended to increase during the first 2–3 weeks of this study, Social form of S. invicta at Farms 2, 4, and 7 are mapped in and declined thereafter (Table 3). The percentage of Figures 1B, 2B, and 3B, respectively. The majority of disturbed S. invicta mounds that attracted P. tricuspis also S. invicta mounds sampled at each location were polygyne varied at each study location from week to week, and was (Farm 2, 67%; Farm 4, 63%; and Farm 7, 95%). A generally highest during maximum soil moisture levels significant (P<0.05) positive spatial association between (Table 3). Dynamic patterns of soil moisture levels at all P. tricuspis abundance and the polygyne S. invicta social three locations were identical. Soil moisture readings form was found at Farm 2 on 12 October, and was increased during the first 2 weeks of the survey, were marginally significant on 5 October (Table 1). Significantly highest during the 28 September survey, and declined (P<0.05) more flies were often associated with disturbed during each successive survey. polygyne than monogyne mounds (Table 2). Overall, P. tricuspis abundance showed a random spatial

distribution pattern, with Ia values close to 1 (Table 3). Spatio-temporal population structure of Pseudacteon tricuspis However, significant (P<0.05) aggregated spatial distribu- Weekly counts of P. tricuspis at individual disturbed tion patterns were detected at Farm 2 on 21 September, S. invicta mounds at Farms 2, 4, and 7 are shown in 5 October, and 19 October and at Farm 7 on 5 October. Figures 1C–G, 2C–G, and 3C–G, respectively. Extremely No significant temporal associations of P. tricuspis spatial variable counts of P. tricuspis were typical of sample patterns were detected for all weekly comparisons (two-tailed locations. In general, fly abundance at Farms 2 and 7 test: P>0.025 for positive and P<0.975 for negative spatial

Table 2 Numbers of Pseudacteon tricuspis (mean ± SE) associated with Solenopsis invicta social form. Number of mounds in parentheses

No. Pseudacteon tricuspis Location Date Monogyne Polygyne P-values Farm 2 21/9 2.7 ± 0.8 (7) 4.5 ± 1.0 (13) 0.16 28/9 2.7 ± 0.7 (7) 3.9 ± 1.5 (13) 0.41 5/10 3.4 ± 3.1 (7) 11.6 ± 2.1 (14) <0.001 12/10 0.0 ± 0.0 (7) 2.5 ± 0.8 (13) 0.007 19/10 0.0 ± 0.0 (5) 1.2 ± 0.6 (12) 0.04 Farm 4 21/9 1.0 ± 0.5 (5) 1.7 ± 0.7 (10) 0.22 28/9 0.5 ± 0.5 (2) 4.7 ± 1.6 (9) – 5/10 1.6 ± 0.7 (5) 2.1 ± 0.7 (10) 0.30 12/10 1.3 ± 0.9 (6) 7.0 ± 1.8 (10) 0.007 19/10 3.5 ± 1.3 (4) 6.9 ± 2.1 (10) 0.10 Fire ant parasitoid dynamics 137

Figure 2 (A) Locations of Solenopsis invicta mounds infected with Thelohania solenopsae at Farm 4 (1 = presence, 0 = absence); (B) S. invicta social form (P = polygyne, M = monogyne); (C-G) time series of Pseudacteon tricuspis abundance at disturbed S. invicta mounds, September–October 2005 [C (21 September), D (28 September), E (5 October), F (12 October), and G (19 October)]. NA, no information.

associations). Clustering of P. tricuspis into significant gaps scale, and provided a detailed and informative perspective < (vr –1.5) was identified at Farm 2 on 5 October, and of P. tricuspis population structure on a local scale. The > patches (vd 1.5) at Farm 2 on 5 October and Farm 7 on 5 overall results suggest that P. tricuspis spatial patterns are October. The cluster indices for these locations and dates generally random, but aggregations of high abundance are contour mapped in Figure 4A,B (Farm 2, Farm 7, occasionally occur in space. In this study, the significant respectively). positive spatial association found between P. tricuspis and polygyne social form at Farm 2 is difficult to interpret in biological terms. It may simply be a function of the dense Discussion population structure of polygyne S. invicta (Macom & This study addressed several questions regarding spatial Porter, 1996). The number of host workers in an individual patterns of abundance of adult P. tricuspis on a spatio-temporal mound is not correlated with P. tricuspis abundance 138 Henne et al.

Figure 3 (A) Locations of Solenopsis invicta mounds infected with Thelohania solenopsae at Farm 7 = = (1 presence, 0 absence); (B) S. invicta social form (P = polygyne, M = monogyne); (C-G) time series of Pseudacteon tricuspis abundance at disturbed S. invicta mounds, September–October 2005 [C (21 September), D (28 September), E (5 October), F (12 October), and G (19 October)]. NA, no information.

(Morrison & King, 2004). However, Morrison & Porter several hundred meters (DC Henne & SJ Johnson, unpubl.) (2005) found a positive correlation between mound area and can therefore respond to host semiochemicals over (m2/ha) and P. tricuspis abundance in North Central considerable distances. Our study sites were relatively Florida. Consequently, high host population density may homogeneous pastures with no evident patchiness in translate into higher abundance of parasitoids in a direct vegetation or other potential heterogeneities that could density-dependent way. account for the significant gaps and patches of P. tricuspis Spatial theory predicts that patchy population distribu- that were identified at Farms 2 and 7. Other than the tions can arise even in continuous habitats through limited aggregated densities of polygyne S. invicta, there were no dispersal combined with host–parasitoid interactions apparent landscape features or environmental heterogeneity (Maron & Harrison, 1997). Certainly, S. invicta nest sites that would have potentially influenced abundance of have limited dispersal, generally moving only a few meters P. tricuspis. Another study (DC Henne, unpubl.) found no (personal observation). Conversely, P. tricuspis can disperse correlation between P. tricuspis abundance and grass Fire ant parasitoid dynamics 139

Table 3 Summary of Pseudacteon tricuspis spatial and temporal abundance at three locations (Farms 2, 4, and 7), aggregation indices (Ia), gap (vr), and patch (vd) indices

% 2 3 4 Sample No. mounds mounds Average no. flies Soil moisture SADIE Ia SADIE vr SADIE vd Location date sampled with flies at mounds1 probe reading (P-value) (P-value) (P-value) Farm 2 21/9 21 81 5.1 ± 0.7 0.7 1.21 (0.05) –1.305 (0.08) 1.132 (0.31) 28/9 21 81 4.3 ± 1.1 2 0.79 (0.85) –0.723 (0.85) 0.889 (0.82) 5/10 22 73 12.9 ± 1.9 0.8 1.93 (0.03) –1.992 (<0.00001) 2.073 (<0.00001) 12/10 21 43 3.7 ± 0.9 0.3 1.05 (0.41) –1.184 (0.03) 0.996 (0.41) 19/10 18 22 3.5 ± 1.2 0.1 1.44 (0.03) –1.480 (0.03) 1.433 (0.03) Farm 4 21/9 17 65 2.6 ± 0.6 1.6 0.86 (0.59) –0.994 (0.31) 0.613 (0.85) 28/9 13 85 7.7 ± 2.4 5 0.81 (0.69) –0.786 (0.67) 0.878 (0.62) 5/10 17 65 3.5 ± 0.7 3 0.65 (0.89) –0.658 (0.90) 0.705 (0.87) 12/10 18 72 8.0 ± 1.6 1.8 1.30 (0.18) –1.346 (0.33) 1.478 (0.26) 19/10 15 87 6.5 ± 1.6 0.3 0.92 (0.49) –0.974 (0.49) 0.822 (0.56) Farm 7 21/9 21 43 2.0 ± 0.3 0.5 1.48 (0.10) –1.435 (0.15) 1.460 (0.10) 28/9 21 90 5.2 ± 0.9 3.3 0.92 (0.56) –0.952 (0.46) 1.073 (0.33) 5/10 21 76 7.1 ± 1.7 1.7 1.81 (0.03) –1.691 (0.08) 1.853 (0.025) 12/10 21 81 6.5 ± 1.4 0.6 0.67 (0.92) –0.684 (0.85) 0.567 (0.97) 19/10 21 43 5.0 ± 1.3 0.3 1.00 (0.44) –0.960 (0.36) 1.275 (0.08) 1Only mounds at which flies were observed. 2 < Ia values in bold indicate aggregated spatial pattern (P 0.05). 3 < Vr value in bold indicates significant presence of gaps (P 0.025). 4 < Vd values in bold indicate significant presence of patchiness (P 0.025).

heights or soil moisture levels that were measured adjacent for P. tricuspis, approximately 38 days at 27 °C (Folgarait to disturbed S. invicta mounds. If landscape features were et al., 2002). important, then we would expect to ﬁnd some consistent The fact that P. tricuspis were collected at disturbed spatio-temporal structure in P. tricuspis populations. mounds containing T. solenopsae-infected workers sug- Instead, spatio-temporal structures of local P. tricuspis gests that flies may not be able to differentiate between populations in homogeneous pastures are inconsistent. infected vs. non-infected ants from long-distance cues. This may be a consequence of very short adult lifespans of However, because only a few S. invicta colonies were this parasitoid (Porter et al., 1995). infected with T. solenopsae, a more detailed spatial repre- The phenology of Pseudacteon parasitoids of S. geminata sentation was not possible. Studies to compare attack in central Texas was studied by Morrison et al. (1999, 2000). rates and/or survival of P. tricuspis on T. solenopsae- They found that phorid abundance varied seasonally and infected hosts have not been done, but would be rainfall patterns were possibly linked to abundance. interesting. Similar patterns have been observed with P. tricuspis in The results of this study have given several insights Florida (Morrison & Porter, 2005) and Louisiana (DC into the spatial structure and dynamics of P. tricuspis Henne & SJ Johnson, unpubl.). The last signiﬁcant rainfall populations in a homogeneous habitat. First, it indicates at the study sites prior to initiation of these surveys occurred that P. tricuspis counts have a random spatial distribution, on 29 August 2005, when Hurricane Katrina passed but spatial aggregations manifest when phorid popula- through southeastern Louisiana. Morrison et al. (2000) tions are high. Second, sampling P. tricuspis populations also determined that soil moisture levels were often a good should be done with the distribution of S. invicta predictor of phorid abundance. Soil moisture levels at populations in mind, particularly where polygyne popula- 10-cm depth did not peak until late September, a lag tions occur. Consequently, any survey of P. tricuspis of approximately 1 month after Hurricane Katrina. should attempt to sample a representative portion of Abundance of P. tricuspis also appeared to positively the area. In a related study (DC Henne & SJ Johnson, respond to this rain event with a lag of approximately unpubl.), it was determined that a minimum of 15 S. invicta 1 month, corresponding to the development time required mounds should be sampled to achieve an estimate of 140 Henne et al.

Acknowledgements Thanks are extended to Lacey Inmon (LSU Agricultural Center) for assistance with PCR methods. Valuable comments were provided by J. Cronin, G. Henderson, and T. Schowalter. Thanks also to K. Smith, N. Smith, and R. Jones for allowing access to their properties to conduct this research. Approved for publication by the Director, Louisiana Agricultural Experiment Station as Manuscript number 2008-234-1438.

References Disney RHL (1994) Scuttle Flies: The Phoridae. Chapman & Hall, London, UK. Dutilleul P (1993) Testing spatial segregation using a nearest neighbor contingency table. Ecology 75: 1940–1948. Ferguson AW, Barari H, Warner DJ, Campbell JM, Smith ET et al. (2006) Distributions and interactions of the stem miners Psylliodes chrysocephala and Ceutorhynchus pallidactylus and their parasitoids in a crop of winter oilseed rape (Brassica napus). Entomologia Experimentalis et Applicata 119: 81–92. Ferguson AW, Klukowski Z, Walczak B, Perry JN, Mugglestone MA et al. (2000) The spatio-temporal distribution of adult Ceutorhynchus assimilis in a crop of winter oilseed rape in relation to the distribution of their larvae and that of the parasitoid Trichomalus perfectus. Entomologia Experimentalis et Applicata 95: 161–171. Folgarait PJ, Bruzzone OA, Patrock RJW & Gilbert LE (2002) Developmental rates and host specificity for Pseudacteon parasitoids (Diptera: Phoridae) of fire ants (Hymenoptera: Formicidae) in Argentina. Journal of Economic Entomology 95: 1151–1158. Glancey BM, Craig CH, Stringer CE & Bishop PM (1973) Multiple fertile queens in colonies of the imported fire ant, Solenopsis invicta. Journal of Entomological Science 8: 127–238. Henne DC, Johnson SJ & Porter SD (2007) Status of Pseudacteon tricuspis Borgmeier (Diptera: Phoridae) releases in Louisiana. Figure 4 Contour maps showing areas of significant patches Florida Entomologist 90: 565–569. (>1.5, lighter grey) and gaps (<–1.5, darker grey) of Krieger MJB & Ross KG (2002) Identification of a major gene Pseudacteon tricuspis at disturbed S. invicta mounds at (A) Farm regulating complex social behavior. Science 295: 328–332. 2 (5 October 2005) and (B) Farm 7 (5 October 2005). Sample Lofgren CS (1986) The economic importance and control of locations denoted by a white X. imported fire ants in the United States. Economic Impact and Control of Social Insects (ed. by SB Vinson), pp. 227–255. Praeger, New York, NY, USA. Macom TE & Porter SD (1996) Comparison of polygyne and monogyne red imported fire ant (Hymenoptera: Formicidae) population densities. Annals of the Entomological Society of the phorid population mean that is reasonably close America 89: 535–543. to the true population mean. At peak populations, Maron JL & Harrison S (1997) Spatial pattern formation in an significant patches and gaps in P. tricuspis abundance at insect host-parasitoid system. Science 278: 1619–1621. S. invicta mounds can occur. Therefore, phorid sampling Morrison LW (2000) Biology of Pseudacteon (Diptera: Phoridae) should be conducted in several widely separated ant parasitoids and their potential to control imported locations to ensure that spatial heterogeneity in phorid Solenopsis fire ants (Hymenoptera: Formicidae). Recent populations can be accounted for. Research Developments in Entomology 3: 1–13. Fire ant parasitoid dynamics 141

Morrison LW, Kawazoe EA, Guerra R & Gilbert LE (1999) tating flies (Diptera: Phoridae) that parasitize Solenopsis fire Phenology and dispersal in Pseudacteon flies (Diptera: Phoridae), ants (Hymenoptera: Formicidae). Florida Entomologist 81: parasitoids of Solenopsis fire ants (Hymenoptera: Formicidae). 292–309. Annals of the Entomological Society of America 92: 198– Porter SD, Fowler HG & MacKay WP (1992) Fire ant mound den- 207. sities in the United States and Brazil (Hymenoptera: Formicidae). Morrison LW, Kawazoe EA, Guerra R & Gilbert LE (2000) Journal of Economic Entomology 85: 1154–1161. Ecological interactions of Pseudacteon parasitoids and Solenopsis Porter SD, Pesquero MA, Campiolo S & Fowler HG (1995) ant hosts: environmental correlates of activity and effects on Growth and development of Pseudacteon phorid fly maggots competitive hierarchies. Ecological Entomology 25: 433–444. (Diptera: Phoridae) in the heads of Solenopsis fire ant workers Morrison LW & King JR (2004) Host location behavior in a (Hymenoptera: Formicidae). Environmental Entomology 24: parasitoid of imported fire ants. Journal of Insect Behavior 17: 475–479. 367–383. Ross KG & Keller L (1998) Genetic control of social organization Morrison LW & Porter SD (2005) Phenology and parasitism in an ant. Proceedings of the National Academy of Sciences of rates in introduced populations of Pseudacteon tricuspis, a the USA 95: 14232–14237. parasitoid of Solenopsis invicta. Biocontrol 50: 127–141. Southwood TRE (1978) Ecological Methods: With Particular Oi DH & Williams DF (2002) Impact of Thelohania solenopsae Reference to the Study of Insect Populations, 2nd edn. Chapman (Microsporidia: Thelohaniidae) on polygyne colonies of red & Hall, New York, NY, USA. imported fire ants (Hymenoptera: Formicidae). Journal of Taylor LR (1984) Assessing and interpreting the spatial distributions Economic Entomology 95: 558–562. of insect populations. Annual Review of Entomology 29: Perry JN (1995) Spatial analysis by distance indices. Journal of 321–357. Animal Ecology 64: 303–314. Tuda M (2007) Understanding mechanism of spatial ecological Perry JN (1998a) Measures of spatial pattern for counts. Ecology phenomena: a preface to the special feature on ‘Spatial 79: 1008–1017. statistics’. Ecological Research 22: 183–184. Perry JN (1998b) Measures of spatial pattern and spatial Valles SM, Oi DH, Perera OP & Williams DF (2002) Detection of association for counts of insects. Population and Community Thelohania solenopsae (Microsporidia: Thelohaniidae) in Ecology for Insect Management and Conservation. Proceedings Solenopsis invicta (Hymenoptera: Formicidae) by multiplex of the Ecology and Population Dynamics Section of the 20th PCR. Journal of Invertebrate Pathology 81: 196–201. International Congress of Entomology, Florence, Italy, 25–31 Valles SM & Porter SD (2003) Identification of polygyne and August 1996 (ed. by J Baumgartner, P Brandmayr & BFJ monogyne fire ant colonies (Solenopsis invicta) by multiplex Manly), pp. 21–33. Balkema, Rotterdam, The Netherlands. PCR of Gp-9 alleles. Insectes Sociaux 50: 199–200. Perry JN & Dixon P (2002) A new method for measuring spatial Weaver DK, Nansen C, Runyon JB, Sing SE & Morrill WL association in ecological count data. Ecoscience 9: 133–141. (2005) Spatial distributions of Cephus cinctus Norton Perry JN & Hewitt M (1991) A new index of aggregation for (Hymenoptera: Cephidae) and its braconid parasitoids animal counts. Biometrics 47: 1505–1518. in Montana wheat fields. Biological Control 34: 1–11. Perry JN, Winder L, Holland JM & Alston RD (1999) Red-blue Winder L, Alexander CJ, Holland JM, Woolley C & Perry JN plots for detecting clusters in count data. Ecology Letters 2: (2001) Modelling the dynamic spatio-temporal response of 106–113. predators to transient prey patches in the field. Ecology Letters Porter SD (1998) Biology and behavior of Pseudacteon decapi- 4: 568–576.