Biogeography and macroecology of phorid flies that attack fire ants in south-eastern Brazil and Argentina

Aim Saevissima group fire ants, Solenopsis richteri and S. invicta, have become serious pests when introduced from Argentina and Brazil to other continents. In South America, Solenopsis are distributed across a great variety of habitats and climates. In North America, S. invicta, introduced free of phorids, now ranges from coast to coast in the south. Success in introducing particular Pseudacteon as agents for the biological control of fire ants has varied across climatic zones. We aimed at assembling all the information about fire ant phorids from Argentina and Brazil, to estimate their richness and geographical ranges, to perform a climatic analysis for these distributions, to define groups and climate‐based communities, and to test and elucidate Rapoport's biogeographical rule. Location Argentina and Brazil (South America). Methods From field and museum collections and historical records, we developed a database of fire ant‐specific phorids throughout their known geographical range. A total of 123 sites with values for 15 climatic variables were mapped between 10° and 38° SL and between 35° and 65° C WL for the presence/absence of phorids. We calculated species richness across all sites combined, and for each phytogeographical region, using rarefaction curves, and ICE and Mmean estimators. We calculated mid‐latitudinal points, geographical ranges and areas for each species. The correlation between mid‐latitudinal point and ranges/areas was tested against a null model generated from the randomization of the raw distributional data. We used several types of multivariate analyses to distinguish groups of phorids by phytogeographical regions, hosts and climate, to find gradients of climate throughout the studied area, to define phorid communities in terms of their relationships with gradients of climate, and to test a mechanism for Rapoport's rule. Results Richness estimations using ICE and Mmean estimators were similar or higher than the observed values depending on the phytogeographical region. Cluster multivariate analyses based on climatic, phytogeographic and host data revealed distinct groupings of Pseudacton. The ‘cerrado’ group was confined to tropical savanna areas. A more ‘widespread’ group included ‘Chaco’ and ‘Maritime’ subgroups defined by their respective association with extreme temperatures or precipitation. Ordination multivariate analyses showed (1) two climatic gradients throughout the study area: one of temperature and the other of precipitation, and (2) that climatic variables significantly explained the observed assemblages of phorids. Positive and negative signs of the eigenvalues from the main axes of a canonical correspondence analysis allowed us to define eight communities whose geographical distribution resembled that of phytogeographical regions. We found a significant and positive correlation between geographical areas and mid latitudinal points, and furthermore, the Mantel test based on climatic variables suggested a mechanism for Rapoport's rule applying in the case of Pseudacteon. Main conclusions Pseudacteon species with greater mid‐latitudinal points occupy broader geographical areas and confront more stressful environmental conditions. Because the composition of Pseudacteon communities is largely determined by climatic variables, the correspondence between climates at sites of origin vs. sites of release should be an important consideration in choosing specific phorids for biocontrol efforts.