Tropical rain forest fragmentation affects biotic interactions in distinct ways. Little is known, however, about how fragmentation affects animal trophic guilds and their patterns of interactions with host plants. In this study, we analyzed changes in biotic interactions in forest fragments by using a multitrophic approach. For this, we classified arthropods associated with Heliconia aurantiaca herbs into broad trophic guilds (omnivores, herbivores and predators) and assessed the topological structure of intrapopulation plant-arthropod networks in fragments and continuous forests. Habitat type influenced arthropod species abundance, diversity and composition with greater abundance in fragments but greater diversity in continuous forest. According to trophic guilds, coleopteran herbivores were more abundant in continuous forest and overall omnivores in fragments. Continuous forest showed a greater diversity of interactions than fragments. Only in fragments, however, did the arthropod community associated with H aurantiaca show a nested structure, suggesting novel and/or opportunistic host-arthropod associations. Plants, omnivores and predators contributed more to nestedness than herbivores. Therefore, Heliconia-arthropod network properties do not appear to be maintained in fragments mainly caused by the decrease of herbivores. Our study contributes to the understanding of the impact of fragmentation on the structure and dynamics of multitrophic arthropod communities associated with a particular plant species of the highly biodiverse tropical forests. Nevertheless, further replication of study sites is needed to strengthen the conclusion that forest fragmentation negatively affects arthropod assemblages.

We evaluated the structure of intrapopulation howler monkey-plant interactions by focusing on the plant species consumed by different sex and age classes in continuous and fragmented forests in southern Mexico. For this we used network analysis to evaluate the impact of fragmentation on howler population traits and on resource availability and food choice. A total of 37 tree and liana species and seven plant items (bark, immature fruits, flowers, mature fruits, immature leaves, mature leaves and petioles) were consumed, but their relative consumption varied according to sex and age classes and habitat type. Overall, adult females consumed the greatest number of plant species and items while infants and juveniles the lowest. For both continuous and fragmented forests, we found a nested diet for howler monkey-plant networks: diets of more selective monkeys represent subsets of the diets of other individuals. Nestedness was likely due to the high selectivity of early life stages in specific food plants and items, which contrasts with the generalized foraging behaviour of adults. Information on the extent to which different plant species and primate populations depend on such interactions in different habitats will help to make accurate predictions about the potential impact of disturbances on plant-animal interaction networks.

We analyzed the influence of natural and anthropogenic habitat disturbance on the structure of invertebrate communities living on two species of Heliconia herbs. We compared the invertebrate community structure associated to both species growing in natural forest gaps, on road edges for H. latispatha, and in riparian vegetation for H. collinsiana. We assessed the topological structure of individual-based Heliconia–invertebrate networks. Species richness was greater in H. collinsiana inhabiting riparian vegetation but no differences were found in the diversity of invertebrates for any Heliconia species and habitat. Invertebrate abundance was greater in gaps for H. latispatha and in riparian vegetation for H. collinsiana showing a species turnover in human disturbed habitats. The invertebrate community was not randomly assembled but highly nested, revealing a structured pattern for all habitat conditions. Heliconia–invertebrate network properties appear to be maintained in human disturbed habitats, despite differences in species richness, abundance and composition and host number and quality. Our study contributes to the understanding of the structure of ecological interactions in contrasting habitats. Because they provide food and habitat for the associated fauna and several microhabitats for colonization, heliconias could be used as habitat elements for invertebrate conservation in human impacted landscapes

Habitat loss and fragmentation are recognized as affecting the nature of biotic interactions, although we still know little about such changes for reptilian herbivores and their hindgut nematodes, in which endosymbiont interactions could range from mutualistic to commensal and parasitic. We investigated the potential cost and benefit of endosymbiont interactions between the spur-thighed tortoise (Testudo graecaL.) and adult oxyurid nematodes (Pharyngodonidae order Oxyurida) in scrublands of southern Spain. For this, we assessed the association between richness and abundance of oxyurid species with tortoises’ growth rates and body traits (weight and carapace length) across levels of habitat loss (low, intermediate and high). Furthermore, by using an intrapopulation ecological network approach, we evaluated the structure and diversity of tortoise–oxyurid interactions by focusing on oxyurid species infesting individual tortoises with different body traits and growth rates across habitats. Overall, tortoise body traits were not related to oxyurid infestation across habitats. Oxyurid richness and abundance however, showed contrasting relationships with growth rates across levels of habitat loss. At low habitat loss, oxyurid infestation was positively associated with growth rates (suggesting a mutualistic oxyurid–tortoise relationship), but the association became negative at high habitat loss (suggesting a parasitic relationship). Furthermore, no relationship was observed when habitat loss was intermediate (suggesting a commensal relationship). The network analysis showed that the oxyurid community was not randomly assembled but significantly nested, revealing a structured pattern for all levels of habitat loss. The diversity of interactions was lowest at low habitat loss. The intermediate level, however, showed the greatest specialization, which indicates that individuals were infested by fewer oxyurids in this landscape, whereas at high habitat loss individuals were the most generalized hosts. Related to the latter, connectance was greatest at high habitat loss, reflecting a more uniform spread of interactions among oxyurid species. At an individual level, heavier and larger tortoises tended to show a greater number of oxyurid species interactions. We conclude that there is an association between habitat loss and the tortoise–oxyurid interaction. Although we cannot infer causality in their association, we hypothesize that such oxyurids could have negative, neutral and positive consequences for tortoise growth rates. Ecological network analysis can help in the understanding of the nature of such changes in tortoise–oxyurid interactions by showing how generalized or specialized such interactions are under different environmental conditions and how vulnerable endosymbiont interactions might be to further habitat loss.