The influence of ambient conditions and space on the phenological patterns of a Solenopsis phorid guild in an arid environment

Biological Control 42 (2007) 262–273 www.elsevier.com/locate/ybcon

The influence of ambient conditions and space on the phenological patterns of a Solenopsis phorid guild in an arid environment Patricia J. Folgarait a, Richard J.W. Patrock a

a,b,* ,

Lawrence E. Gilbert

b

Centro de Estudios e Investigaciones, Universidad Nacional de Quilmes, Roque Saenz Pen˜a 180, B1876BXD Bernal, Buenos Aires, Argentina b Section of Integrative Biology and Brackenridge Field Laboratory, University of Texas, Austin, TX 78712, USA Received 26 May 2006; accepted 25 April 2007 Available online 25 May 2007

Abstract We observed the diurnal distribution of a phorid parasitoid guild of Solenopsis fire ants across five sites in an arid region of western Argentina over 17-months. We found a rich assembly of 15 taxa, of which 7 species were found each month of the year and over most times of the day. The majority of species were found most frequently in the evening. A Canonical Correspondence Analysis of the hourly abundances of the flies in relation to field meteorological conditions suggested that two broad groups of species existed, one of which had flight periods associated with hotter, drier conditions than the second. The first group was most commonly represented by Pseudacteon tricuspis, the P. obtusus complex and P. cultellatus, while some members of the second group, such as the P. nocens complex and P. litoralis were the most abundant and commonly found flies. The range of conditions in which these flies were found suggests that all of the common taxa represent populations that might be suitable for introduction into similarly arid environments of Texas.  2007 Elsevier Inc. All rights reserved. Keywords: Argentina; Climate; Fire ants; Parasitoid guild; Niche partitioning; Seasonal activity; Microselia aduncus; Pseudacteon borgmeieri; Pseudacteon bulbosus; Pseudacteon comatus; Pseudacteon convexicauda; Pseudacteon cultellatus; Pseudacteon curvatus; Pseudacteon litoralis; Pseudacteon nocens; Pseudacteon nr. nocens; Pseudacteon nudicornis; Pseudacteon obtusus; Pseudacteon nr. obtusus; Pseudacteon solenopsidis; Pseudacteon tricuspis; Solenopsis electra; Solenopsis macdonaghi; Solenopsis invicta; Solenopsis richteri; Biological control

1. Introduction Decisions involving the selection of a natural enemy for employment in a classical biological control program center around issues of safety, efficacy and practicality (Carruthers and D’Antonio, 2005). These concerns have certainly been in the foreground for the biological control of imported fire ants in the southern United States, specifically with respect to the phorid parasitoids in the genus Pseudacteon (Porter et al., 2004; Gilbert and Patrock, 2002; Graham et al., 2003; Vazquez et al., 2005; Vogt * Corresponding author. Address: Centro de Estudios e Investigaciones, Universidad Nacional de Quilmes, Roque Saenz Pen˜a 180, B1876BXD Bernal, Buenos Aires, Argentina. Fax: +54 11 4365 7182/7101. E-mail address: [email protected] (R.J.W. Patrock).

1049-9644/$ - see front matter  2007 Elsevier Inc. All rights reserved. doi:10.1016/j.biocontrol.2007.04.020

and Streett, 2003). There is an abundance of species choices associated in the native range of these pest ants in South America where over 20 species of Pseudacteon utilize Solenopsis fire ants as hosts in this region (Disney, 1994; Porter and Pesquero, 2001; Folgarait et al., 2005a). To help ensure on-target management of the imported species, host specificity tests have exploited phylogenetic relationships within the host genus (Porter and Gilbert, 2004 and references therein) whereby populations of Pseudacteon species that do not discriminate between the South American saevissima complex and the North American geminata complex of the Solenopsis fire ant group (Trager, 1991) are considered too host generalist to safely admit for releasing. The best taxonomic focal unit for introduction is considered to be the population (Folgarait et al., 2006) because there is extensive known inter-populational variation in host

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

choice or preference (Porter and Briano, 2000; Patrock et al., 2006) and because there is evidence for populations of size related morphs or cryptic species for at least two species, P. nr. obtusus (Folgarait et al., 2005b; Kronfurst et al., submitted) and P. nr. nocens (Folgarait et al., 2006). In addition, there is informed suspicion (Gilbert and Patrock, 2002; Folgarait et al., 2003, 2006) that species ranging from the wet tropics to dry savannahs probably vary across populations with respect to physiological tolerances (Folgarait et al., 2005a). Local guilds, or communities of fire ant Pseudacteon are typically composed of 5– 10 species (Orr et al., 1997; Folgarait et al., 2003) so there can be abundant choices of taxa. We use the term ‘guild’ here because of its connotation of functional similarity (Root, 1967), as well as given the fact the larger local community of phorids attacking ants includes many other taxa of both the flies and host ants. The investigation of the temporal activity patterns is one research approach that Folgarait et al. (2003) used to explore interspecific variation of these flies and they laid out their assumptions about how this information could be used to help decide upon candidates. Their basic premise was and is that there should be a firm understanding of the fit, as well as mismatches in abiotic conditions between source and potential release sites of candidates. Their explicit assumption was that historical climatic conditions experienced by the populations could be used as guidelines for predicting possible tolerances and activity patterns of the flies under conditions that might be encountered in the adopted range. We continue this approach in this study and document the activity patterns of the guild of Pseudacteon in an arid area of western Argentina, Santiago del Estero. Our intended release areas for candidates found in this study are in central and south Texas with a focus on the south Texas plains phytogeographic region. Gilbert and Patrock (2002) discuss reasons why this area climatically matches well with that in Santiago del Estero, namely, correlations with low rainfall and high temperatures. Several factors would seem to tightly tie activity patterns of these flies with ambient conditions. They are soft bodied and tiny, for the most part being a few mm in length and must struggle with the demands of surface–volume conditions in this size range. Their adult life span is short (around a week or less, Fadamiro et al., 2005) and the presence of female flies attacking the ants indicates an active ovipositional phase that is constrained by its host’s activities. Diurnal activities of the flies are tied to solar period in two respects; adult emergence is just before sunrise or during the morning (Wuellner et al., 2002; Folgarait, personal observation) and oviposition extends only while there is available light (Pesquero et al., 1996; Morrison et al., 1999; Orr et al., 1995) and if temperatures are suitable (Fowler et al., 1995; Morrison et al., 1999, 2000; Wuellner et al., 2003; Folgarait et al., 2003). Our goals were to describe the daily, seasonal and spatial phenological patterns of a Pseudacteon guild from an arid environment in relationship to measured meteorologi-

263

cal conditions. These results could be used as guidelines for species husbandry and facilitating collections of species for rearing or post-release monitoring. In addition, we wished to compare the activity patterns of the flies in this area with those found for the Argentine S. richteri host guild (Folgarait et al., 2003) to better understand interpopulational variation in climatic tolerances of the different fly species, specifically with respect to being able to match these populations with our field sites in the United States. 2. Materials and methods We studied the Solenopsis fire ant phorid guild in the outskirts of the capital of the Argentina province of Santiago del Estero, near Brea Pozo (S 28.25, W 63.95). The area lies in the phytogeographical region of the dry western Chaco, (Cabrera and Willink, 1980), characterized by xerophyllic trees such as the quebrachos Schinopsis and Aspidosperma, along with brea (Cercidium praecox (R. et P.) Harms) and various Prosopis such as itin, P. kuntzei Harms. We designated five sites for observation in this area because they offered us a useful combination of host-parasitoid availability, habitat variation and logistical access. Sites were located in two distinct localities, one that was composed of four patches of fire ant mounds (Sites 1–4) and the other a single patch (Site 5), each varying according to vegetation cover and proximity to irrigation. Sites 1 and 4 were separated by approximately 300 m, while Sites 2 and 3 were in the approximate midrange of these two and separated by about 10 m and a tree line. Site 5 was approximately 3 km from these. We found two host species, S. interrupta Santschi and S. invicta Buren existing in mosaics within each of these sites. Solenopsis invicta was almost the exclusive Solenopsis fire ant in Site 4 and the only one sampled, while S. interrupta was more common in the other sites and the only one sampled in Site 1. Identification of S. saevissima group spp. can be difficult in their native range and so in addition to employing the keys of Pitts (2002) and Trager (1991), we had specimens identified using the professional services of J. Trager. From January 2003 through July 2004, we monitored each of these sites on a monthly basis utilizing a 10-min sampling-without-replacement protocol, repeated throughout the day. Sampling typically began in the first or second hour following sunrise and continued at least until sunset. The 10 min sample would include walking along a transect of 5–6 disturbed mounds per site and looking for phorids hovering over, or attacking the fire ants. These would be aspirated, identified to sex or species (Pesquero and Porter, 2001) and after each sampling day returned to the field. Females could normally be identified in the field using a 15–20· magnifying lens but we were unable to discriminate males by species. When identification of females was unclear, we would bring the specimens back to the laboratory in Buenos Aries for confirmation. We disturbed mounds by digging a small hole in its side and after the ants emerged, we placed a small plug of tuna fish inside the hole

264

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

to feed and maintain their presence above ground. Periodically, the mound was redisturbed using a stick. Meteorological conditions were monitored in the field by (1) placing HOBO dataloggers (ONSET Corp.) under shade covers and (2) measuring tri-hourly soil temperatures at 10 cm depth with temperature probes in full sun, full shade and partial sun-shade conditions. These observations allowed us to obtain minimum, mean and maximum temperatures and humidity (vapor pressure deficits) that varied positionally within site. Soil temperatures are considered the best indicators of host fire ant foraging activities (Porter and Tschinkel, 1987) and the other conditions were considered likely to have a strong influence on flight activity levels of these tiny flies (Folgarait et al., 2003). The observation points were changed on semi-hourly basis according to the solar direction. Historical climatic data was obtained from the nearest available weather station at the airport in Santiago del Estero. We sampled at various times during the day over the period of 17 months. Chronological times are not equivalent across seasons at the latitude and longitude of Santiago del Estero with respect to the length of solar days, so we standardized our hours across seasons by using Temporal Hours (Terwilliger and Sawyer, 1996). These units have the same relative values for sunrise, solar noon and sunset for each day of the year but differ with respect to their lengths, being relatively longer in the summer than in the winter. To determine Temporal Hours (TH), we found the proportion of the day that elapsed between sunset and sunrise when each sample was taken; that is, we divided the difference between sample time and sunrise time by the daylength for each sample date. These decimal fractions were then placed in one of 12 equal sized numerical bins to obtain our Temporal Hours (e.g. values between 0 and 0.083 were assigned 0100 TH). Times for sunrise, solar noon, sunset, civil and nautical twighlights at the latitude and longitude of Brea Pozo, Argentina (S 28.25, W 63.95) were determined using the U.S. Naval Observatory converter at http://aa.usno.navy.mil/data/docs/RS_OneYear.html. With respect to our use of Temporal Hours instead of either civil or Babylonian hours, such as used by Pesquero et al. (1996) in their diurnal study of Brazilian phorid activity, we find two related benefits. First, we can describe hours of the day as scalars that are independent of the calendar and second, Temporal Hours allowed us to make across seasons ANOVA comparisons of time in a scale (solar) that is more likely encountered by the flies than that dictated by a clock. Seasons were defined by date of the sample (e.g. Spring is March 21st–June 20th). We were interested in examining the overall effects of time and space and the contribution of sampling on variation in phorid diversity and abundance. We compared species richness with respect to sites and times, using our samples, averaged by hour as a replicate, with a threeway ANOVA (Site, Season, Time of Day). Because our sample sizes were unbalanced with respect to the combination of Site by Month by Hour, we used seasons as defined

above and a broader categorization of time of day (Morning = 0000–0259, Noon = 0300–0559, Afternoon = 0600– 0859, Evening = 0900–1200). Species accumulation curves can be useful in selecting collection strategies (Longino, 2000), particularly since they give estimates of the extent to which additional sampling extends, or increases the number of species found (Gotelli and Colwell, 2001). To better understand the influence of length of time and time of day in the field on our daily estimates of species richness we generated our observed species accumulation curves based on hourly periods according to our sampling protocol. To make species richness comparisons across seasons and time periods, we generated Coleman rarefaction richness values and their 95% confidence intervals for each day-site-time period using EstimateS 7.5 (Colwell, 2005). We tested for species richness differences across seasons, time periods and sites by examining overlapping confidence intervals among the means of the final species richness values (Payton et al., 2003). Previous studies indicate that Pseudacteon exhibit climatic activity profiles (Folgarait et al., 2003; Morrison et al., 1998). We therefore asked the extent to which our guild might have been structured by ambient meteorological conditions. To address this multivariate question, we used Canonical Correspondence Analysis (CCA, Legendre and Legendre, 1998). CCA is a direct extension of multiple regression but differs in that its mathematical algorithms force the ordination axes to represent relationships between the matrices, represented in our case by taxa abundances and the environmental data. As such it is has been found to be instructive for directly addressing questions such as how multispecies abundance data are influenced by environmental gradients. We examined this question at 3 scales: the hour, day and month. Our sample sizes were 486, 73, and 18 for the hour, day and month analyses, respectively. For all CCA analyses, our field abundance data formed the basis for our species matrix and these differed only by the summations of the means. The meteorological matrices differed, however, among our analyses. For the hour CCA, only field meteorological data (hourly means) was used in our secondary, or explanatory matrix. In the day and month datasets we used the daily or monthly field means along with rainfall data, specifically rainfall sums and event counts from 14 or 60 days preceding the field work obtained from the meteorological station. In addition we used the temperature range for the day or month. All meteorological variables were Z-value transformed because of their different units and scales. The abundance data were transformed to have Chord distances since we had several uncommon taxa, using the application ‘Transformation’ designed for this purpose by Legendre and Gallagher (2001). Since CCA is related to multiple regression, we addressed the problem of multicollinearity of our environmental matrix by removing variables with VIF scores greater than 10 (Belsley et al., 1980).

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

Two of the three assumptions for interpreting the biological importance of the results of this test (Legendre and Legendre, 1998) were examined prior to analysis. The first assumption is that the species have unimodal distributions along the environmental variables. To do this, we examined which of the meteorological distributions for the common species (greater than 100 observations) were not significantly playkurtic since a bimodal distribution is considered to be an extreme type of platykurtosis. The second assumption is that the environmental gradients are long enough to allow each species to have an optimal frequency and less than optimal tails. To understand the limits of this assumption, we looked at the parameter ranges for all the variables and their coenoclines (Gauch and Whittaker, 1972) for each of the common species. The third assumption, here uncontested, is that the species distributions are under environmental control. We tested, using PC-ORD 4 (Ellison, 2000), the null hypothesis that there was no structure to the species abundance matrix and therefore no relationship between the datasets. To meet the assumptions of the different statistical analyses we transformed our data before analysis. All species data was first adjusted for sampling effort. Species counts were square root transformed prior to ANOVA and Bonferroni’s method was applied across all analyses to adjust the P-values for concurrently run tests. Other statistical programs used included JMP 6.0 (ANOVA and MANOVA, Lehamn et al., 2005), Excel 2003 with PopTools 2.6.9 (for Monte-Carlo Estimates, Hood, 2005) and Statview 5.0. Filemaker 7.0 was used to join datasets for other data manipulations. Voucher specimens for the species are kept at the Universidad Nacional de Quilmes.

265

3. Results We sampled the five sites over the course of 17 months, making 3473 10-min samples within 703 civil hours (668 Temporal hours). We found 15 phorid species in two genera in our sites: (1) Microselia aduncus Borgmeier, (2) Pseudacteon borgmeieri Schmitz, (3) P. bulbosus Brown, Folgarait and Gilbert, (4) P. comatus Borgmeier, (5) P. convexicauda Borgmeier, (6) P. cultellatus Borgmeier, (7) P. curvatus Borgmeier, (8) P. litoralis Borgmeier, (9) P. nocens Borgmeier, (10) P. nr. nocens, (11) P. nudicornis Borgmeier, (12) P. obtusus Borgmeier, (13) P. nr. obtusus, (14) P. solenopsidis (Schmitz) and (15) P. tricuspis Borgmeier. Three species were found every month of the survey (Table 1, Fig. 1, P. nocens, P. nr. nocens and P. obtusus) and an additional five species were found every calendar month of the year (P. cultellatus, P. litoralis, P. nudicornis, P. nr. obtusus and P. tricuspis). Other species were rare being collected either only once (P. bulbosus and P. comatus), on two separate days (M. aduncus) or on scattered days through the study (P. borgmeieri, 8 d, 4 months, P. solenopsidis (6 d, 4 months, summer and fall). Intermediate in occurrence between these groups were species that were conspicuously absent during the winter (P. curvatus, 11 months and P. convexicauda, 7 months). Males, which are likely to be P. tricuspis and/or P. obtusus (Calcaterra et al., 2005) were found each month sampled (Table 1). We found the fewest (6–8 species) during the winter (July–Sept, Fig. 1). The difference in species richness was accounted for by the absence of the rare and intermediate collected species during these months. The highest species richnesses on a monthly basis were found between February and April 2004 (12–13 species). On a matched monthly

Table 1 Occurrences (expressed as percentages) of the different taxa within the Solenopsis phorid parasitoid guild during the study period (2003–2004) Taxa

P. nocens Males P. litoralis P. cultellatus P. obtusus P. tricuspis P. nr. nocens P. nr. obtusus P. nudicornis P. curvatus P. convexicauda P. borgmeieri P. solenopsidis M. aduncus P. bulbosus P. comatus Total number

Percentage of time founda,b Month

Day

Hour

Morning

Noon

Afternoon

Evening

100 100 82.4 94.1 100 88.2 100 88.2 88.2 64.7 41.2 29.4 23.5 11.8 5.9 5.9 17

83.6 68.5 68.5 61.6 67.1 57.5 60.3 61.6 25.7 23.3 19.2 11.0 8.2 2.7 1.4 1.4 74

49.0 34.0 28.5 28.0 25.0 23.8 18.3 17.4 15.2 4.8 3.8 2.6 1.7 0.5 0.2 0.2 668

48.6 16.5 20.2 17.4 13.8 5.5 16.5 9.2 8.5 0.9 11.0 1.8 0.9 0.0 0.0 0.0 109

44.6 33.2 25.0 26.6 29.4 22.3 9.2 14.1 17.1 3.3 2.2 2.2 0.5 1.1 0.0 0.0 184

39.0 36.3 21.4 28.6 23.1 30.2 11.0 15.9 15.9 6.6 1.1 2.2 2.2 0.6 0.0 0.6 182

63.0 42.7 43.2 34.9 29.2 29.7 34.9 26.6 18.3 6.8 3.7 3.7 2.6 0.0 0.5 0.0 192

Species are ordered by their relative occurrence on an hourly basis. a Percentage was calculated as the number of months, days, hours or time periods each taxa was observed divided by the total number of possible observations for that interval. b Fonts for percentages that have been boldfaced or italicized represent the period of the day with the highest or lowest percentage observations, respectively for each of the taxa.

266

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

Phorid relative abundance

4

6 11

3

7

P. nudicornis P. nr. obtusus P. nr. nocens P. tricuspis P. obtusus P. cultellatus P. litoralis P. nocens

12 8

9

10

10 13

2

8 10

1

5 10

6

12

7

8

NS 0 1

2 Su

3

4

5

6

7

8 W

F

9

10

11 Sp

2003

12

1

2

3

4

Su

5 F

6

7 W

2004

Fig. 1. Monthly mean numbers of the common Solenopsis phorids on a per sample basis across the study period from January 2003 to August 2004. The graph is stacked with species graphed in descending order of overall abundance from the bottom of the graph. Numbers above the lines represent the numbers of all species found in that month. NSSampling not done in April and May 2003.

basis, there were fewer species and generally fewer flies/species found in 2003 than in 2004, though P. nr. nocens was a clear exception to this observation (Fig. 1). The relationship between species richness and abundance was not significant (r = 0.33, P = 0.20). All except the rare species were found in all sites. With respect to species richness by site, all species were found in site 1, where also M. aduncus, P. bulbosus and P. comatus were only found. All species except these were found at site 5 and with the added exception of P. solenopsidis, all other species were found at the other sites. Most of the common species could be found broadly through the day with absences only in the early morning (Table 1, Fig. 2). Four species were found at all temporal hours (P. litoralis, P. nocens, P. nr. nocens, and P. obtusus). Absent only in the first hour around sunrise were P. borgmeieri, P. cultellatus, P. nudicornis, and P. nr. obtusus, as well as males. Pseudacteon curvatus and P. tricuspis were absent only during the first 2 h following sunrise, while P. convexicauda and P. solenopsidis were also found throughout the day, though they occurred at scattered intervals. The most common species found at any time of the day was P. nocens, which was 44.1% more likely to be found on an hourly basis than the second ranked species, P. litoralis. Evening hours were generally the best time to collect flies with males and ten of the 15 species more likely to be found in this time frame than any other (Table 1). Mornings were generally the poorest times to collect with nine species least likely to be found in this time frame. The species exception to this morning deficit was P. convexicauda, which had its highest incidence early in the day. The distribution of phorid abundances, however, was bimodal with the highest peak in the evening and a smaller peak in the morning. These modes were largely the result of increased numbers of P. nocens at these times, and to a lesser extent, P. nr. nocens and P. litoralis (Fig. 2a). Species richness on a mean hourly basis differed according to time of day, season and site sampled in our 3-way

ANOVA (Site by Season by Time of Day ANOVA, Table 2). We found significant Site by Season and Season by Time of Day interactions but not a significant Site by Time of Day interaction, so we split our next analyses by the common interacting factor, Season, to examine how site and time of day varied across the year. The Site and Time of Day interaction was non-significant so we looked at time and spatial components separately. With respect to Sites, significant differences were found in three seasons, spring, fall and winter. Site 4 had higher species richness on an hourly basis in all of these seasons than Site 2 and greater richness than site 5 in Winter, while Site 3 had greater species richness than Sites 2 and 5 in the Spring and Winter (all adjusted P values < 0.03, respectively, Fig. 4). All other site-season contrasts were not significant at P = 0.05. With respect to time of day, significant differences were seen in all seasons (Fig. 3). A significantly greater species richness on an hourly basis was found in the Evening than in Morning (Fall), Noon (Spring and Winter) and Afternoon (Spring and Summer). Afternoon had a significantly greater species richness on an hourly basis than Noon (Fall and Winter), as well as Morning in Fall. Morning and Noon had significantly greater values than Afternoon in the Summer (with all adjusted P-values < 0.023, Fig. 3). The observed species accumulation curves indicated a mean increase of around 0.65–1 species found per hour of additional field observation. Evening hours, again, found the most species of flies, according to this measure. Sampling done during the morning, noon, afternoon or evening found on average approximately 33.0, 54.2, 51.3 and 76.3% of all species collected on a daily basis, respectively. Our analysis of mean species richnesses (Coleman estimates) across seasons supported this assessment of quality later collecting, in that evening had significantly greater species richnesses in each season than the other time periods (Table 3). The morning hours were generally the least productive using these estimates with the exception of the summer in which we found the fewest number of species in the afternoon.

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

Table 2 Three-way ANOVA Table for species richness based on the mean hourly number of species observed

3.0

2.0

P. litoralis P. cultellatus P. nr. nocens P. nocens

1.0

0.5

TH-2

TH-4

TH-6

TH-8

TH-10

Variables

DF

SS

F-Value

P-Value

Power

Site (ST) Season (SN) Time of day (TD) ST*SN ST*TD SN*TD ST*SN*TD Residual

4 2 3 8 12 6 24 483

8.45 2.22 18.42 4.82 2.27 19.67 6.85 108.69

9.39 4.92 27.29 2.68 0.84 14.57 1.27

<0.0001 0.0077 <0.0001 0.0069 0.6072 <0.0001 0.1788

1.000 0.814 1.000 0.937 0.495 1.000 0.920

TH-12

Mean no.species observed

Phorid relative abundancea

0

267

P. nr. obtusus P. obtusus P. tricuspis Males

0.25

Morning

2

Noon Afternoon Evening

1.5

1

0.5

0 Spring

Summer

Fall

Winter

Season

0 TH-4

TH-6

TH-8

TH-10

TH-12

Fig. 3. Hourly mean number of species (±standard errors) observed per sample for the four seasons at different times of the day across the study period.

0.1

P. borgmeieri P. convexicauda P. curvatus P. nudicornis

0.05

0.0 TH-2

Morning

TH-4

TH-6

Noon

TH-8

Afternoon

TH-10

TH-12

Evening

Temporal Hour Fig. 2. Hourly mean numbers of the most common Solenopsis phorids on a per sample basis for temporal hours across the study period from January 2003 through August 2004. (a) Data for the P. nocens complex, P. cultellatus and P. litoralis. (b) Data for the P. obtusus complex, males and P. tricuspis. (c) Data for P. borgmeieri, P. convexicauda, P. curvatus and P. nudicornis. The graphs are stacked with species graphed in descending numerical order within each graph. aPlease note that the Y-axis scales of these graphs are substantially different. Refer to Table 1.

With respect to the Coleman estimates by site, Site 1 consistently had the highest Coleman values in each season (Table 3) while the rankings changed among the other sites according to season. Site 2 had significantly higher Coleman estimates than sites 3, 4 and 5 in Summer and Sites

Mean no.species observed

TH-2

1.0

Spring Summer Fall Winter

0.8

0.6

0.4

0.2 1

2

3

4

5

Site Fig. 4. Hourly mean numbers of species observed per sample (square-root transformed) at each field site by season.

3 and 4 in Fall but this site was ranked lowest in the Winter. Site 3 and Site 4 had intermediate Coleman values relative to other sites. Site 5 was ranked last among the sites in Spring and Summer but it was second in Fall and Winter. The results from Table 3 suggested that temperature was also important with respect to species richness because of the typically lower Coleman values associated with higher and lower temperatures in the summer and winter, respectively. To examine this observation, we ran a step-wise

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

Table 3 Estimated species richness (maximum Coleman estimate ±1 standard deviation of the mean) for seasons by time of day and site Time of day

Spring

Morning Noon Afternoon Evening

Summer

No. species ± SD 4.7 ± 0.24 5.8 ± 0.12 4.1 ± 0.23 8.4 ± 0.11

i g i d

Site

No. species ± SD

1 2 3 4 5

10.0 ± 0.07 c 9.0 ± 0.06 d 10.0 ± 0.08 c 8.0 ± 0.01 e 6.0 ± 0.1 g

Morning Noon Afternoon Evening

6.8 ± 0.34 6.6 ± 0.19 0.6 ± 0.05 8.8 ± 0.17

f f j c

1 2 3 4 5

12.0 ± 0.11 11.0 ± 0.12 10.0 ± 0.11 10.0 ± 0.08 9.0 ± 0.06

b b c c d

Fall

Morning Noon Afternoon Evening

4.9 ± 0.24 i 8.0 ± 0.12 e 9.5 ± 0.11 b 10.0 ± 0.1 a

1 2 3 4 5

13.0 ± 0.06 11.0 ± 0.01 10.0 ± 0.06 10.0 ± 0.01 12.0 ± 0.06

a b c c b

Winter

Morning Noon Afternoon Evening

1 2 3 4 5

11.0 ± 0.07 4.0 ± 0.14 6.0 ± 0.02 7.0 ± 0.01 8.0 ± 0.08

b h g f e

0.1 ± 0.2 k 5.1 ± 0.19 h 5.7 ± 0.24 gh 6.6 ± 0.12 f

Differences in letters within column represent means with non-overlapping 95% confidence intervals.

regression of our daily field meteorological data, along with the difference between the beginning and final daily temperatures (thermal amplitude) against species counts. Daily accumulated temperature was the only explanatory variable accepted into the model (F1, 72 = 24.5, P < 0.0001). The results of our CCA analyses were mixed. The first three axes of the hour and day but only the first axis of the month CCA were significant (PC ORD Monte-Carlo estimates, P < 0.04 in all cases). With respect to the hour analysis, the first three axes explained 85.9% of the variation in the phenological data (68.4%, 10.2% and 7.3% for the first, second and third axes, respectively, Fig. 5). Since the amount of variation explained by the day and month CCA analyses was either low (15.6) or moderate (34.9 for day and month, respectively) we discuss the results of the hour CCA. The first axis splits most of the common species into two groups, on one hand P. cultellatus, P. obtusus, P. nr. obtusus and P. tricuspis and on the other, P. litoralis, P. nocens, P. nr. nocens. The first group we consider to be more heat-dry tolerant (Table 4) than the second ‘temperate’ group based upon (1) the high positive intra and interset correlations and biplot scores (all equal or greater than 0.6, Fig. 5) of each of the meteorological variables with the first CCA axis, (2) the significant Species-Environment correlation with this axis (0.67, P = 0.001, PC-ORD MonteCarlo estimates) and (3) the mean temperatures and vapor pressure deficits present when the taxa were observed (Table 4). The position of time near the origin indicates that the variable Time was equivocal with respect to assigning taxa on this axis. The position of males was nearest to P. tricuspis (Fig. 5).

bl cm

1.5

Hourly CCA Axis 2

Season

(10.2)

268

1

a

S (7.3%)

nd 0.5

0

s

cul cn

NL

cur Time br

-0.5

(68.4)

M

n2

-0.5

T3

O V2

o2

T2 Tr

ST T1 V1

0

0.5

1

1.5

Fig. 5. Canonical Correspondence Analysis (CCA) biplot between abundances of phorid species and hourly field meteorological conditions. Species abbreviations are a, M. aduncus; br, P. borgmeieri; bl, P. bulbosus; cm, P. comatus; cn, P. convexicauda; cl, P. cultellatus; cur, P. curvatus; L, P. litoralis; N, P. nocens; n2, P. nr. nocens; nd, P. nudicornis; O, P. obtusus; o2, P. nr. obtusus; s, P. solenopsidis; Tr, P. tricuspis and M, Males. Abbreviations for meteorological conditions used were ST or hourly maximum soil temperature (C), T1, T2 and T3 or minimum, mean and maximum air temperatures (C) and V1 and V2 or mean and maximum vapor pressure deficits (kPa). Time is the temporal hour of the observation. The most common taxa abbreviations are boldened.

4. Discussion This study documents the richest local fire ant phorid guild known to date. More species were found at site 1, in an area of about 1/4 hectare, than are known for all the described Pseudacteon (12) in North America. In fact, the median daily species richness for the study (5.5) is as large as that known for the entire Nearctic phorid guild hosted by the Solenopsis geminata complex (5 species). Outside of the context of ant–parasitoid interactions, this is the largest documented contingent, to our knowledge, of a single genus of parasitoids known for a host locality (Krombein et al., 1979). This unprecedented find in species richness might be partially explained by looking at host, spatial and climatic variation, as well as our sampling effort. One factor most likely underlying the richness of Pseudacteon is the radiation of the host Solenopsis in South America, where most of the Pseudacteon are found. The majority of the fire ant species are found here (17 of 20 species) and there is a growing awareness that Solenopsis invicta has undergone a rapid radiation on its own, in that there are cryptic species hidden within its southern South American populations (Ahrens et al., 2005). Host species diversification might help explain the baseline of the greater number of Pseudacteon in South America than other regions. Host variation might also be a factor in why we found more than a doubling of species relative to that of Folgarait et al. (2003), in a study on S. richteri with a similar time frame of sampling. There is no evidence for a western to eastern cline in Pseudacteon diversity across Argentina since (1) we found no significant differences in species rich-

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

269

Table 4 Means and standard deviations for vapor pressure deficits (VPD); air and soil temperatures recorded when individuals of each of the species were present Species

No. of hours

VPD (kPa)

Air temperature (C)

Soil temperature (C)

Mean rank

M. aduncus P. borgmeieri P. bulbosus P. comatus P. convexicauda P. cultellatus P. curvatus P. litoralis P. nocens P. nr. nocens P. nudicornis P. obtusus P. nr. obtusus P. solenopsidis P. tricuspis Males

3 17 1 1 25 192 31 197 344 127 106 169 122 11 161 224

2.3 ± 0.1 2.9 ± 1.3 0.8 1.3 2.0 ± 1.2 2.5 ± 1.6 3.2 ± 1.9 2.1 ± 1.4 2.1 ± 1.4 1.9 ± 1.3 2.0 ± 1.4 2.9 ± 1.9 3.1 ± 1.9 1.3 ± 1.7 3.3 ± 1.8 3.0 ± 1.9

34 ± 0.7 26.1 ± 6.7 26.2 26.2 26.1 ± 7.9 26.6 ± 10.1 28.7 ± 10.4 24.1 ± 8.7 24 ± 9.0 23.8 ± 8.2 23.8 ± 9.7 27.9 ± 9.6 27.3 ± 10.5 23 ± 8.7 28 ± 11.2 27.5 ± 10.2

25.6 ± 1.6 18.4 ± 8.9 16.2 15.3 20.5 ± 7.1 19.7 ± 8.8 18.6 ± 10.1 17.5 ± 8.9 17 ± 8.5 17.8 ± 9.4 16.5 ± 8.9 20.4 ± 9.1 21.6 ± 9.4 16.2 ± 9.5 21.6 ± 8.8 20.6 ± 8.9

3.6 7.0 13.1 13.2 9.0 7.1 4.1 11.7 11.6 12.1 12.3 5.3 4.3 14.4 2.6 4.6

Mean rank is the average rank for all meteorological variables held by each of the taxa or males.

ness/locality among the regions reported in Calcaterra et al. (2005) (One-way ANOVA F2, 27 = 4.52, P = 0.11), (2) their data suggested that Northwestern Argentina generally had lower richness than Eastern Argentina (2.8 vs. 4.8 species/locality) and (3) c-diversity of the regions also indicates the reverse trend, though this is most certainly due to less historical sampling in the west (Folgarait et al., 2003). In the present study, we monitored an area with a complex bi-species mosaic of fire ants (S. invicta and S. interrupta) that might have led to an increase in richness over single host species guilds, such as that surveyed by Folgarait et al. (2003). Two other species, S. electra and S. macdonaghi are known from the region, as well (Pitts, 2002). The rarest taxa, M. aduncus, P. bulbosus and P. comatus, for example were only associated with S. interrupta since they were found at Site 1 where we only found this ant species, while the other species were found associated with both ant taxa. Most studies of Solenopsis phorid guilds, however, also examined fire ant mosaics, either with S. invicta and other S. saevissima (i.e. Orr et al., 1997) or species within the S. geminata complex (Morrison et al., 1999) and given the overlapping distribution patterns of Solenopsis fire ants in South America (Pitts, 2002; Trager, 1991) fire ant community mosaics are probably common in most areas of Argentina outside of the southern provinces where only S. richteri is known. Sampling effort clearly is a factor in our finding of so many species, particularly since three of the taxa were encountered only once to thrice and four others were collected sparsely. Pseudacteon bulbosus is known only from the province of Santiago del Estero and has never been found common at any locality (Brown et al., 2003; Calcaterra et al., 2005, this study). The other sparse species are common components of their Solenopsis guilds in the eastern part of their range, including M. aduncus in southern Brazil on S. saevissima (Fowler et al., 1995), P. convexicauda, P. comatus and P. borgmeieri in Buenos Aires on S. richterii (Bruzzone, 2004; Folgarait et al., 2005a,b) and

P. solenopsidis in Brazil on S. invicta (Orr et al., 1995). Pseudacteon curvatus is apparently common on a wide variety of Solenopsis fire ants across its range but we found it infrequently on both hosts. It is likely that our study coincided with a lull in its population dynamics since it was extremely common in the area before this study and then following probably related with atypical harsh weather conditions (Folgarait et al., Unpublished data). The other infrequent species, P. convexicauda is not considered to be a Solenopsis parasitoid but instead a parasitoid of Paratrechina spp (Porter and Gilbert, 2004). We include it in our dataset because it was repeatedly found attacking and not just hovering over Solenopsis ants at mounds in the clear absence of its other putative hosts. It was also found attacking Paratrechina sp. at experimental baits at Site 5 (Patrock, Pers. Obs.) and in Buenos Aires (Folgarait, Pers. Obs.). Whether it develops on either of these ant genera is not known but its recurrent recorded proclivity for both of these taxa and others (Pheidole and Camponotus, Bruzzone, 2004) needs a fuller examination than it has been given since other Pseudacteon are not found repeatedly making the same ‘‘host identification’’ mistake. In our study seven species were found every calendar month of the year while Folgarait et al. (2003) found five of six species with intermittent temporal distributions and only one species, P. borgmeieri found specifically during the winter, as well as the rest of the year. One explanation for this contrast in temporal distributions might be that meteorological conditions in Buenos Aires were more severe than what was experienced during our study. The opposite is actually the case (Fig. 6). There was less precipitation and higher vapor pressure deficits in Santiago del Estero than Buenos Aires across the year (Fig. 6a) and the minimum temperatures during the winter were as low or lower in the west than around Buenos Aires during the study periods (Fig. 6b). The historical records describe a similar pattern in that Santiago del Estero tended to have more stressful summer and winter temperature and humid-

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

Mean daily VPd (KPa)

270

Santiago del Estero 1

3

11 1

1

5 9

11 3 9 7

7

5

Buenos Aires

0 0

5

10

Mean monthly rainfall (cm) 35

Mean daily temperature (oC)

Santiago del Estero

30 Maximum

25 Buenos Aires

20 15

Santiago del Estero

20 15 Minimum

10 Buenos Aires

5 0

J

F

Summer

M

A

M

Fall

J

J

A

Winter

S

O

N

D

Spring

Fig. 6. Comparative climatographs of the meteorological conditions present during the current and the Buenos Aires study (Folgarait et al., 2003). (a) The mean monthly rainfall sums (cm) plotted against the daily mean vapor pressure deficits (kPa) for months of the year. (b) Illustrates the mean daily minimum and maximum temperatures (C) during the two studies plotted by the month of year.

ity conditions than would be seen at Costanera Sur in Buenos Aires (Prohaska, 1976). Generally, this will be the case for meteorological comparisons between deeply internal and coastal areas at the same latitude. The contrasts in climate suggests that populations of the set of overlapping species differ in their respective climatic tolerances, with those in the west probably being more stress tolerant. In both studies, fewer species and abundances were found in the winter months than otherwise, suggesting that seasonal patterns of the flies vary by degree for all the taxa. Santiago del Estero, however, did have a significantly higher daily thermal amplitude than Buenos Aires during the study times (Fig. 6b, ANOVA, F1, 2314 = 1985.9, P < 0.0001) and in every season (Fig. 6b, ANOVA, F3, 2314 = 42.0, P < 0.0001). This set of results, along with our analysis of thermal amplitude and species richness suggests that the lower temperature thresholds for some of the species may not have been reached in Folgarait et al’s. (2003) study but they were in the present study. The continual presence of the eight common species across seasons during the study suggests that in this area, these species do not exhibit diapause or that the necessary triggers for its induction were not presented during the time interval observed. The latter possibility or the possibility that there exists intra-populational variation in diapause

is suggested by our rearing data, as well. Progeny of Pseudacteon tricuspis flies collected in this area in May 2001 and reared in the laboratory in Austin, TX did not emerge until November 2001, 5 months later. Given that (1) we have long experience with such international shipments and have not observed before such a long delay in emergence and (2) the 2001 meteorological conditions were not substantively different from that of the current study, it is more likely that we observed intra-populational variation in diapause tendencies of P. tricuspis in this population. As for the less common species, the long absences are ambiguous with respect to diapause. One ‘less common’ species, P. curvatus has been collected frequently outside of the study period, suggesting that our 18-month cross-section of time was too brief to capture its broader populations dynamics. If so, explanations for absences for the less common species are confounded by standard sampling issues at low population levels. The temporal distribution patterns that we saw for the taxa with respect to diurnal (Table 1) and temperature relations (Table 3) are otherwise similar to that outlined in the multivariate approach of Folgarait et al. (2003) and the descriptive works of Pesquero et al. (1996) and Calcaterra (2005), although we observed a broader range of conditions for our taxa and we add new information for other species. In general, richness and abundance increased over the course of the day. We found that evening hours, seasonally adjusted were generally the best time to collect an assortment of species under the conditions of the study, though colder conditions in this period, such as normally found on winter days could limit collecting during this time period (Folgarait and Gilbert, 1999; Morrison et al., 1999; Wuellner and Saunders, 2003). These trends might be explained by temperature and the emergence behavior of the flies. First, the coldest time of the day is at sunrise (Aguado and Burt, 2004) and temperatures increase temporally with insolation. Accordingly, we found that the thermal amplitude sampling was our best meteorological variable in explaining daily species richness, perhaps because of the wider range of temperature options available for the species. Emergence profiles of the flies also are a factor since our rearing data of the taxa from this area places emergence late in the morning, biasing the presence of these short-lived species later into the day. We were not able to discriminate activity levels according to time and meteorological conditions, according to our CCA analyses, however. Such a relative diminishment in activities, though, is suggested in Fig. 3 where the number of species found on winter evenings and afternoons were not significantly different. We concur with the observations of Calcaterra et al. (2005) that Pseudacteon nocens and P. litoralis will be found attacking ants at dusk and we add P. obtusus to this list. We would often require a flashlight to see into the mounds to detect the flies. Turning on and off the light did not noticeably affect their behavior in the holes (Orr, 1992), though the flies, specifically P. nocens, could be found up until nautical sunset (ca. 0.01–0.003 foot-candles,

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

Leibowitz (1987) with their departure being more gradual than that described by Calcaterra et al. (2003). These species are apparently better dark adapted than other Pseudacteon. The temporal schedules of these flies as seen here may fit effectively into that of the activity patterns of their hosts. Foraging of S. invicta, for instance, is periodic; there is often a lull during the heat of the day, followed by a resurgence of activity in the late afternoon, with the evening and nocturnal contribution to foraging predominate over other times of the day. The fact that we were most likely to find more species of flies during the normal times of resumption of ant foraging and its buildup, as well as these three species of flies found nearly to the limit of available light, suggests the possibility of a fortuitous synchrony between host and parasitoid activity amplitudes. We found significant differences in species counts across our sites according to season. The species counts varied according to the scale, however with sample measurements (averaged by hour) and species accumulations (Coleman estimates) suggesting different contrasts that were sometimes contradictory (Fig. 4, Table 3). For instance, at the finest scale, Site 2 had consistently lower species richness than Site 4 whereas its accumulated species richness was significantly higher than site 4 in each season except winter. The opposing trends suggest that while there were more species on a per sample basis in some sites versus others, there was also more redundancy in which species appeared in those sites over the course of the day. In general, the pattern that we saw with respect to differences among the sites on the within-hour scale appeared to be related to the degree of vegetation cover with Sites 2 and 5 having relatively fewer species than the others and also fewer shrubs and trees. This was most pronounced in the contrast between Sites 2 and 3 that were about a decameter apart. The tree-lined site 3 was significantly richer three of four seasons than site 2 on this scale, which was largely barren ground during the study. Vegetation cover was also thought to be related to the population dynamics of P. curvatus in release and establishment efforts of this species (Graham et al., 2003). The results of our CCA analysis and the follow-up examination of thermal-moisture relations for the individual taxa suggests that almost all of the species with flight periods best suited for the hot-dry conditions in south Texas are already being released or are under culture, including P. tricuspis, P. obtusus and P. cultellatus with the exclusion only of P. nr. obtusus (Table 3). However, all of the eight most common taxa (Table 1) were active regularly under hot and dry conditions even though thetemperate group species (P. nocens complex and P. litoralis) were apparently more likely to hide away until cooler conditions prevailed. From another perspective, however, P. nocens and P. litoralis were the most common species, suggesting that even though the flight periods of the ‘temperate’ may not be as relatively common in hot periods, their population numbers were not adversely affected by the extended periods of the hot-dry weather prevalent in

271

the dry Chaco. These species should all do well as candidates for incorporation into a classical biological control program for the region of the South Texas plains with its similar climatic conditions to the dry Chaco. In addition, there may be a payoff in adding physiologically differentiated populations of the other common taxa that already have been released but with little success (Gilbert and Patrock, 2002) in more arid areas of the fire ant’s acquired distributional range. The benefit in such a ‘redundant’ release effort can be seen across all temporal scales in that flies from this population were found over almost all times of the day, under a wide range of meteorological conditions and across all the months of the year under conditions that could be seen normally in south Texas. We suggest that these results indicate that the flies are more likely to establish in this area than might other populations from other, perhaps less severe areas of South America (Folgarait et al., 2005a). In addition, if these activity schedules held after establishment, there would be little enemy time-free space during daylight for the fire ants. This type of temporal packing is important when the presumed mechanism underlying the use of these flies is related to their impact on the ants’ resource acquisition (Feener, 1981) more than on the mortality they inflict (Gilbert and Patrock, 2002). One caveat should be noted, however. We have not observed a single instance of a ‘swarm’ of phorids, with dozens to hundreds of flies at a mound (this study, Folgarait and Patrock, Unpub. Data,), as is commonly seen with P. tricuspis and P. curvatus in other localities (Porter et al., 1995, 2005), including Buenos Aires in Argentina (PJF, Pers. Obs.). Our observations in Santiago del Estero indicate a rich guild of many commonly occurring taxa. With respect to the species abundances, there appears to be a greater level of evenness than seen in other studies and perhaps generally lower overall population levels in this area, apparently independent of host densities, for the two focal taxa currently released in Texas. This observation suggests that there are likely costs associated with stress tolerances in these populations that might affect the population rates of increase for some of the species. Acknowledgments Gustavo Azzimonti did virtually all of the fieldwork. Rodolfo Carrara helped him on occasion. We are indebted to the staff of the Direccio´n de Fauna Silvestre de Santiago del Estero who provided permits to perform this research. This research was supported by the Lee and Ramona Bass Foundation, the Helen C. Kleberg and Robert J. Kleberg Foundation and The State Of Texas Fire Ant Project. P.J.F. thanks CONICET for their oversight and the Universidad Nacional de Quilmes for use of their facilities. References Aguado, E., Burt, J.E., 2004. Understanding Weather and Climate. Prentice Hall, Upper Saddle River, NJ.

272

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273

Ahrens, M.E., Ross, K.G., Shoemaker, D.D., 2005. Phylogeographic structure of the fire ant, Solenopsis invicta in its native South American range: roles of natural barriers and habitat connectivity. Evolution 59, 1733–1743. Belsley, D.A., Kuh, E., Welsch, R.E., 1980. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. Wiley, New York. Brown, B.V., Folgarait, P.J., Gilbert, L.E., 2003. A new species of Pseudacteon Coquillet (Diptera: Phoridae) attacking Solenopsis fire ants (Hymenoptera: Formicidae) in Argentina. Sociobiology 41, 685–688. Bruzzone, O.A., 2004. Ana´lisis de patrones de distribucio´n de los dı´pteros parasitoides del ge´nero Pseudacteon: una aproximacio´n a trave´s de tres escalas diferentes. PhD thesis, Universidad de Buenos Aires, Argentina. Cabrera, A.L., Willink, A., 1980. Biogeografı´a de Ame´rica Latina. Secretarı´a General de la Organizacı´on de los estados Americanos, Programa Regional de Dearrollo Cientı´fico y Tecnolo´gico, Washington, D.C. Calcaterra, L.A., Porter, S.D., Briano, J.A., 2005. Distribution and abundance of fire ant decapitating flies (Diptera: Phoridae: Pseudacteon), in three regions of southern South America. Ann. Entomol. Soc. Am. 98, 85–95. Colwell, R.K., 2005. EstimateS: statistical estimation of species richness and shared species from samples. User’s guide and computer program, version 7.5, published at http://purl.oclc.org/estimates. Carruthers, R., D’Antonio, C., 2005. Science and decision-making in biological control of weeds: benefits and risks of biological control. Biol. Control 35, 181–182. Disney, R.H.L., 1994. Scuttle Flies: The Phoridae. Chapman & Hall, London. Ellison, A.M., 2000. PC-ORD Version 4 Review. Bull Ecol. Soc. Am. 81, 127–128. Fadamiro, H.Y., Chen, L., Graham, L.F., 2005. Lifespan and patterns of accumulation and mobilization of nutrients in the sugar-fed phorid fly, Pseudacteon tricuspis. Physiol. Entomol. 30, 212–224. Folgarait, P.J., Bruzzone, O.A., Gilbert, L.E., 2003. Phenological patterns for parasitoids specialized attacking Solenopsis fire ants at the Reserva Eco´logica Costanera Sur, Argentina. Biol. Control 28, 368–378. Folgarait, P.J., Bruzzone, O.A., Porter, S.D., Pesquero, M.A., Gilbert, L.E., 2005a. Biogeography and macroecology of phorid flies that attack fire ants in south-eastern Brazil and Argentina. J. Biogeogr. 32, 353–367. Folgarait, P.J., Chirino, M.G., Patrock, R.J.W., Gilbert, L.E., 2005b. Development of Pseudacteon obtusus (Diptera: Phoridae) on Solenopsis invicta and Solenopsis richteri fire ants (Hymenoptera: Formicidae). Environ. Entomol. 34, 308–316. Folgarait, P.J., Gilbert, L.E., 1999. Phorid parasitoids affect foraging activity of Solenopsis richteri under different availability of food in Argentina. Ecol. Entomol. 24, 163–173. Folgarait, P.J., Patrock, R.J.W., Gilbert, L.E., 2006. Development of Pseudacteon nocens (Diptera: Phoridae) on Solenopsis invicta and Solenopsis richteri fire ants (Hymenoptera: Formicidae). J. Econ. Entomol. 99, 295–307. Fowler, H.G., Pesquero, M.A., Campiolo, S., Porter, S.D., 1995. Seasonal activity of species of Pseudacteon (Diptera: Phoridae) parasitoids of fire ants (Solenopsis saevissima) (Hymenoptera: Formicidae) in Brazil. Cientifica (Jaboticabal) 23, 367–371. Gauch Jr., H.G., Whittaker, R.H., 1972. Coenocline simulation. Ecology 53, 446–451. Gilbert, L.E., Patrock, R.J.W., 2002. Phorid flies for the biological suppresion of imported fire ant in Texas: region specific challenges, recent advances and future prospects. Southwest Entomol. Suppl. 25, 7–17. Gotelli, N., Colwell, R.K., 2001. Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecol. Lett. 4, 379–391.

Graham, L.C., Porter, S.D., Pereira, R.M., Dorough, H.D., Kelley, A.T., 2003. Field releases of the decapitating fly Pseudacteon curvatus (Diptera: Phoridae) for control of imported fire ants (Hymenoptera: Formicidae) in Alabama, Florida, and Tennessee. Flor. Entomol. 86, 334–339. Hood, G., 2005. PopTools, 2.6.9. Available from: . CSIRO, Canberra. Krombein, K.V., Hurd Jr., P.D., Smith, D.R., Burks, B.D., 1979. Catalog of Hymenoptera in America North of Mexico. Smithsonian Institution Press, Washington, DC. Lehamn, A., O’Rourke, N., Stepanski, L., 2005. JMP For Basic Univariate And Multivariate Statistics. A Step-by-step Guide. SAS Institute, Inc. Carey, NC. Legendre, P., Gallagher, E., 2001. Ecologically meaningful transformations for ordination of species data. Oecologia 129, 271–280. Leibowitz, H.W. 1987. Ambient illuminance during twighlight and from the moon. In: Johnson, C.A. Leibowitz, H.W. (Eds.), Night Vision Current Research and Future Directions. Symposium Proceedings. Springer, Berlin. pp. 19–22. Longino, J.T., 2000. What to do with the data. In: Agosti, D., Majer, J.D., Alonso, L.E., Schultz, T.R. (Eds.), Ants. Standard Methods for Measuring and Monitoring Biodiversity. Smithsonian Institute Press, Washington, DC, pp. 186–203. Morrison, L.W., Kawazoe, E.A., Guerra, R., Gilbert, L.E., 1999. Phenology and dispersal in Pseudacteon flies (Diptera: Phoridae), parasitoids of Solenopsis fire ants (Hymenoptera: Formicidae). Ann. Entomol. Soc. Am. 92, 198–207. Morrison, L.W., Kawazoe, E.A., Guerra, R., Gilbert, L.E., 2000. Ecological interactions of Pseudacteon parasitoids and Solenopsis ant hosts: environmental correlates of activity and effects on competitive hierarchies. Ecol. Entomol. 25, 433–444. Orr, M.R., 1992. Parasitic flies (Diptera: Phoridae) influence foraging rhythms and caste division of labor in the leaf-cutter ant, Atta cephalotes (Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 30, 395–402. Orr, M.R., Seike, S.H., Benson, W.W., Gilbert, L.E., 1995. Flies suppress fire ants. Nature 373, 292–293. Orr, M.R., Seike, S.H., Gilbert, L.E., 1997. Foraging ecology and patterns of diversification in dipteran parasitoids of fire ants in South Brazil. Ecol. Entomol. 22, 305–314. Patrock, R.J.W., Albioni Montenegro, G.E., Gilbert, L.E., Folgarait, P.J., 2006. Possible identification of cryptic species of S. invicta using parasitoid flies: parasitoid host discrimination is population dependent. In: Proceedings of the XV Congress of the IUSSI, Washington, DC, pp. 164–349. Payton, M.E., Greenstone, M.H., Schenker, N., 2003. Overlapping confidence intervals or standard error intervals: What do they mean in terms of statistical significance? J. Ins. Sci. 3, 1–6. Pesquero, M.A., Campiolo, S., Fowler, H.G., Porter, S.D., 1996. Diurnal patterns of ovipositional activity in two Pseudacteon fly parasitoids (Diptera: Phoridae) of Solenopsis fire ants (Hymenoptera: Formicidae). Flor. Entomol. 79, 455–457. Pitts, J.P., 2002. A cladistic analysis of the Solenopsis saevissima speciesgroup (Hymenoptera: Formicidae). Ph.D. Dissertation, University of Georgia, Athens, p. 266. Porter, S.D., Briano, J., 2000. Parasitoid-host matching between the little decapitating fly Pseudacteon curvatus from Las Flores, Argentina and the black fire ant Solenopsis richteri. Flor. Entomol. 83, 422–427. Porter, S.D., Gilbert, L.E., 2004. Assessing host specificity and field release potential of fire ant decapitating flies. In: Van Driesche, R.G., Reardon, R. (Eds.), Assessing Host Ranges of Parasitoids and Predators Used for Classical Biological Control: A guide to Best Practice. Forest Health Technology Enterprise Team, Morgantown, WV, pp. 152–176. Porter, S.D., Nogueira de Sa´, L.A., Morrison, L.W., 2004. Establishment and dispersal of the fire ant decapitating fly Pseudacteon tricuspis in North Florida. Biol. Control 29, 179–188.

P.J. Folgarait et al. / Biological Control 42 (2007) 262–273 Porter, S.D., Pesquero, M.A., 2001. Illustrated key to Pseudacteon decapitating flies (Diptera: Phoridae) that attack Solenopsis saevissima complex fire ants in South America. Flor. Entomol. 84, 691–699. Porter, S.D., Tschinkel, W.R., 1987. Foraging in Solenopsis invicta (Hymenoptera: Formicidae): effects of weather and season. Environ. Entomol. 16, 802–808. Porter, S.D., Vander Meer, R.K., Pesquero, M.A., Campiolo, S., Fowler, H.G., 1995. Solenopsis (Hymenoptera: Formicidae) fire ant reactions to attacks of Pseudacteon flies (Diptera: Phoridae) in Southeastern Brazil. Ann. Entomol. Soc. Am. 88, 570–575. Prohaska, F., 1976. The climate of Argentina, Paraguay and Uruguay. In: Schwerdtfeger, W. (Ed.), Climates of Central and South America. Elsevier Scientific, New York, pp. 13–112. Root, R.B., 1967. The niche exploitation pattern of the blue-grey gnatcatcher. Ecol. Monogr. 37, 317–350. Terwilliger, R., Sawyer, F., 1996. The dialist’s companion, MS-DOS program and manual. The American Sundial Society, Rockville, MD.

273

Trager, J.C., 1991. A revision of the fire ants, Solenopsis geminata group (Hymenoptera: Formicidae: Myrmicinae). J. N.Y. Entomol. Soc. 99, 141–198. Vazquez, R.J., Porter, S.D., Briano, J.A., 2005. Field release and establishment of the decapitating fly Pseudacteon curvatus on red imported fire ants in Florida. Bio. Control 51, 207–216. Vogt, J.T., Streett, D.A., 2003. Pseudacteon curvatus (Diptera: Phoridae) laboratory parasitism, release and establishment in Mississippi. J. Entomol. Sci. 38, 317–320. Wuellner, C.T., Porter, S.D., Gilbert, L.E., 2002. Eclosion, mating, and grooming behavior of the parasitoid fly Pseudacteon curvatus (Diptera: Phoridae). Flor. Entomol. 85, 563–566. Wuellner, C.T., Saunders, J.B., 2003. Circadian and circannual patterns of activity and territory shifts: comparing a native ant (Solenopsis geminata, Hymenoptera: Formicidae) with its exotic, invasive congener (S. invicta) and its parasitoids (Pseudacteon spp., Diptera: Phoridae) at a central Texas site. Ann. Entomol. Soc. Am. 96, 54–60.