Figure 1 in First characterization of a taxonomically well-resolved trophic network composed by host plants and gall midges (Diptera: Cecidomyiidae) in the Neotropical region

Figure 1. Bipartite network of host plants and gall midge species at the Restinga of Barra de Maricá (Maricá, RJ, Brazil). Lower bars represent host plant species and upper bars represent gall-midge species; grey bars represent interactions. Bar thickness is proportional to the number of interactions of each species. ble interactions. The observed connectance (C = 0.028) decades), which enhances the sampling of rarer interacwas lower than expected from null model values (Null tions, and consequently increases the specialization of C = 0.032 ± 0.001, p <0.001). Similarly, the observed num- the network. However, this fact reinforces the relevance ber of links per species (L = 0.608) also was lower than of the observed patterns, because even with such a long expected by chance (Null L = 0.710 ± 0.020, p <0.001). sampling, only species-specific plant-galling interactions The observed modularity for plant-galling network was were registered. very high (M = 0.958), but did not differ from null model The structure of the network formed by the gall midgvalues (Null M = 0.959 ± 0.001, p> 0.05). Robustness ob- es and their host plants proved to be highly specialized. served was relatively low (R = 1.343), but was higher than The connectance observed in the present study (2.8%) expected by chance (Null R = 1.334 ± 0.232, p <0.001). was low as compared to other plant-phytophagous networks (review in Araújo et al., 2015). However, comparing with other networks of galling arthropods, the val- DISCUSSION ue observed here was higher than observed by Araújo