The abundance of invertebrates on H. collinsiana ? differed significantly between habitats (x2=5.07x2=5.07, df=1df=1, P=0.02P=0.02) and among invertebrate species within habitat (x2=844.1x2=844.1, df=62df=62, P<0.001P<0.001). Overall, the abundance of invertebrates was greater in riparian vegetation than in forest gaps; there were significantly greater number of individuals within the orders Araneae, Hymenoptera, Lepidoptera and Stylommatophora (terrestrial snails and slugs) in forest gaps than in riparian vegetation ( Fig.?1). For H. latisphata ?, the abundance of invertebrates was significantly different between habitats (x2=x2= 16.18, df=1df=1, P<0.001P<0.001) and among invertebrate species within habitat (x2=379.1x2=379.1, <a href="http://www.selleckchem.com/products/AZD0530.html
">click here</a> df=52df=52, P<0.001P<0.001). Contrary to H. collinsiana, invertebrate abundance was significantly greater in forest gaps than in road edges; whereas the number of individuals within the orders Araneae, Coleoptera, Hymenoptera, Isoptera, Lepidoptera and Stylommatophora was greater in forest gaps than on road edges ( Fig.?1). The insect orders with significantly greater number of individuals in human disturbed habitats were Diptera for H. collinsiana <a href="http://en.wikipedia.org/wiki/ALOX15
">ALOX15</a> in riparian vegetation and Hemiptera for H. latispatha on road edges. Furthermore, habitat condition influenced the composition of invertebrate communities in the studied heliconias. The invertebrate communities of H. collinsiana and H. latispatha in forest gaps were more similar in species composition (Bray�CCurtis index 0.80) than those in road edges and riparian vegetation (Bray�CCurtis index 0.56). The most dissimilar invertebrate community <a href="http://www.selleckchem.com/products/wnt-c59-c59.html
">beta-catenin inhibitor</a> comparing the three habitats was that inhabiting H. latisphata on road edges. Within species similarity was, 0.68 for H. latispatha between forest gaps and road edges, and 0.71 for H. collinsiana between forest gaps and riparian vegetation. The completeness of the inventories was, 89% in gaps and 85% in road edges for H. latisphata; whereas for H. collinsiana, 89% in gaps and 85% in riparian vegetation. Comparing the species rarefaction curves at the lowest abundance value of 212 individuals, species richness was significantly greater for H. collinsiana in riparian vegetation than in forest gaps ( Fig.?2); while no significant difference in the richness of invertebrate species was found between forest gaps and road edges in H. latispatha. Furthermore, species diversity was very similar between habitats for the two Heliconia species, where: H. latisphata, 31.35 and 29.74 effective species in forest gaps and road edges, respectively; and in H. collinsiana 28.42 and 31.51 effective species in gaps and riparian vegetation, respectively ( Fig.?2). We observed that regardless of plant species (H. collinsiana ? and H. latisphata ?) or habitat (natural and human disturbed), all our four Heliconia ?�Cinvertebrate networks exhibited a nested pattern of interactions (Mean?��?SD. NODF ?: 39.02?��?4.25; all pp-values <0.05).