Fig 1.
Presence records of Discocyrtus dilatatus plotted over relevant ecoregions in subtropical and temperate South America.
Yellow dots: records used to calibrate the models; white dots (in Paraguay): three pre-1950 records excluded in this research. Dashed line: Tropic of Capricorn (Trop. Capr, 23°26’14”S). Nomenclature of ecoregions [14]: HCh: Humid Chaco, DCh: Dry Chaco, Yu: Southern Andean Yungas, FS: Paraná flooded savanna, Es: Espinal, HP: Humid Pampas, MS: Southern Cone Mesopotamian savanna, UrS: Uruguayan savanna, APAF: Alto Paraná Atlantic forests, AMF: Araucaria moist forests, SM: Serra do Mar coastal forests, Pa: Pantanal, Ce: Cerrado. Inset: sector represented in South America, displaying the ecoregions that form the ‘dry diagonal’ of open vegetation (Dry Chaco-yellow, Cerrado-grey, Caatinga-blue). Maps were designed using free spatial data available at http://www.diva-gis.org/Data, and https://www.worldwildlife.org/publications/terrestrial-ecoregions-of-the-world.
Table 1.
Localities of Discocyrtus dilatatus studied for COI, with geographical coordinates, haplotypes detected, collection identifiers (CDA-F id#) and GenBank accession numbers.
Fig 2.
Distribution models of Discocyrtus dilatatus for current climate and projection on different Quaternary conditions.
(a) Current climate; (b) Holocene climatic optimum (HCO, -6k); (c) Last Glacial Maximum (LGM, -21k); (d) Last inter-glacial (LIG, -130k). Maps (a) and (d) display the mean grid for the 30-replicates run, whereas (b) and (c) represent de averaged grid of the three GCM simulations (CCSM, MIROC and MPI, each one obtained as the mean grid for the 30-replicates run). Scale of suitability (logistic), from light to dark red: 0.4269 (threshold), 0.550, 0.660, 0.760. In the case of (b) and (c), the yellowish area around the prediction shows the maximal span of predictions with all three GCMs together. Training AUC for the 30-replicates run: average 0.969 (0.963–0.972). In maps (b) and (d) the approximate extent of the marine transgressions during -6k (+6 m) and LIG (+9 m) is also displayed. Abbreviations for Argentinean Provinces: Ju: Jujuy, Sa: Salta, T: Tucumán, Cat: Catamarca, Cba: Córdoba, SF: Santa Fe, Ch: Chaco, Fo: Formosa, Mis: Misiones, Cor: Corrientes, ER: Entre Ríos. Map was designed using free spatial data available at http://www.diva-gis.org/Data.
Table 2.
Changes of range size in Discocyrtus dilatatus, as modeled in different chronologies / Global Climate Models (GCM) in the Upper Quaternary.
Fig 3.
Overlay of the binary predictions of Discocyrtus dilatatus for three LGM simulations (CCSM, MIROC and MPI).
The overlap area of three, two or one models is displayed in decreasing intensities of red. Within the three-GCM-overlap, the ‘stable area since -21k’ (i.e., the grid-cells shared by all current, -6k and LGM models) is identified with blue.
Fig 4.
Paleodistributional shifts of Discocyrtus dilatatus during the Upper Quaternary, as predicted on three different GCM simulations.
(a) CCSM, (b) MIROC and (c) MPI. Grey: areas of the precedent chronology lost during each transition. Color: new range (lighter, areas gained via expansion; darker, areas shared with previous stage). In the case of LIG to LGM transitions, the very few grid-cells shared are in red, to improve contrast. Maps were designed using free spatial data available at http://www.diva-gis.org/Data.
Table 3.
COI polymorphism in Discocyrtus dilatatus for the complete data set and two data partitions.
Table 4.
Hierarchical partition of the variance components for haplotypes of Discocyrtus dilatatus under different hypotheses.
Fig 5.
Haplotypes of Discocyrtus dilatatus: Distribution and network.
(A) Geographical distribution of haplotypes. Circles sizes are proportional to the number of specimens examined in each locality; its internal divisions represent the number of haplotypes per locality. Dots indicate known records of D. dilatatus; those with thick outline are the localities studied in the phylogeographic analysis. Widespread haplotypes (i.e., shared by two or more sectors of the species range) are identified with a number and the color key in the inset; other haplotypes just use the color standardized in the inset of B. Map was designed using free spatial data available at http://www.diva-gis.org/Data. (B) Median-joining haplotype network. The number of mutational steps between adjacent haplotypes is represented by line marks. The size of circles corresponds to the frequency of each haplotype, and the number of localities where a given haplotype occurs is displayed as internal divisions. Colors indicate the geographical sectors recognized in the species range, as defined in the inset and in Table 1.
Fig 6.
Calibrated maximum-clade-credibility tree obtained with beast.
Estimated ages (large numbers) are given only for nodes with posterior probability >0.67 (indicated with open dots; the higher support, the larger dots); small numbers display the 95% credibility interval. Supported clades are identified by letters (A to G). Color squares refer to the geographical areas predefined in Fig 5, and circles point out the ancestral haplotypes, as recognized by their position in the median-joining network; their chronological span in the tree is indicated with a distinct color in branches (grey, blue, brown). Stars indicate presumed expansion events affecting NWA haplotypes. Crosses at some nodes display the position in the phylogeny of the hypothetical nodes recovered in the haplotype network. Pleistocene glacial-interglacial cycles, with ages in kya, are represented by vertical stripes; Last Glacial Maximum (LGM) in darker blue.
Fig 7.
Bayesian skyline plot, depicting the demographic history of Discocyrtus dilatatus.
The y axis represents the increase of the effective population size and the x axis represents the time in million years. The solid line indicates the mean value of the population size over time; the shaded area displays the 95% confidence interval.
Table 5.
Major biogeographical scenarios proposed for the Upper Quaternary history of Discocyrtus dilatatus, based on the integrated molecular and paleodistributional evidence.