Phylogenetic and morphological relationships between nonvolant small mammals reveal assembly processes at different spatial scales

Multiple interacting factors affect seed predation in an African savanna small mammal community

Multiple factors affect seed predation, including seed traits, habitat type, seed predator community composition, predation risk, and seasonality. How all these factors and their interactions simultaneously influence seed predation has rarely been tested experimentally in situ. Here, we assessed the relative contribution of the factors driving seed predation in an African savanna rodent community, comprising six ecologically similar species. We first conducted seed preference tests under semicaptive conditions to determine which seed trait (size, shell hardness, nutritional content) influenced seed predation. Then we performed in situ experiments to establish whether rodent community composition (diversity and abundance), seed type, habitat type, seasonality, predation risk, and their interactions affected seed predation. Semicaptive experiments showed that rodents preferred smaller, lighter seeds, containing relatively high water content. In situ experiments showed that predation risk was an important factor influencing seed predation, with rodents removing considerably more seeds in areas where predation risk was lower. Habitat type also affected seed predation, but its effects were strongly linked to predation risk. In areas where predation risk was higher, rodents removed more seeds in more heterogeneous habitats, whereas in areas where predation risk was lower, rodents removed more seeds in less heterogeneous habitats. Seasonality was the least influential factor shaping seed predation. Rodents removed more seeds in winter compared to other seasons, but only in areas where predation risk was low. We provide experimental evidence for a multifaceted approach to understanding the relative contribution of the different factors driving variation in seed predation in natural communities and show that these factors are likely hierarchically arranged.

A global database on non-volant small mammal composition in natural and human-modified habitats

Non-volant small mammals, which include small-bodied representatives from several mammal orders, have been used as a model group to test the effects of habitat conversion and edge creation on biodiversity. Small mammals occupy a large variety of habitat types and vegetation strata, and have varied lifestyles and diets. They include species with slow-to fast-life history (the Etruscan shrew Suncus etruscus and European Hare Lepus europaeus, respectively) and with very specialized to very generalist habits and diets (the Atlantic bamboo rat Kannabateomys amblyonyx and house mouse Mus musculus, respectively). There are no databases with global coverage focusing on small mammal composition in natural and human-modified habitats and that include neglected natural habitats (e.g. grasslands and savannas). Here, peer-reviewed articles were searched in the primary literature to synthesize almost half century (1973–2017) of research on small mammal composition in natural forests, grasslands and their natural edges, and in five types of human-modified habitats (human-induced forest edges, human-induced grassland edges, crop fields, clear-cuts and tree plantations). The complete database includes information from 199 peer-reviewed articles. Presence data were obtained for 534 species (including 30 unidentified) in 551 sites distributed in 45 countries, 92 ecoregions, 10 biomes and six realms. Measurements of sampling effort and number of species records (number of individuals, captures) per habitat were also obtained, from which researchers can calculate a measure of abundance standardized by the sampling effort. The database will be useful for researchers interested in local-to broad-scale patterns of alpha- and beta-diversity in natural and human-modified habitats.

Environmental filtering is the main assembly rule of ground beetles in the forest and its edge but not in the adjacent grassland

In a fragmented landscape, transitional zones between neighboring habitats are common, and our understanding of community organizational forces across such habitats is important. Edge studies are numerous, but the majority of them utilize information on species richness and abundance. Abundance and taxonomic diversity, however, provide little information on the functioning and phylogeny of the co-existing species. Combining the evaluation of their functional and phylogenetic relationships, we aimed to assess whether ground beetle assemblages are deterministically or stochastically structured along grassland-forest gradients. Our results showed different community assembly rules on opposite sides of the forest edge. In the grassland, co-occurring species were functionally and phylogenetically not different from the random null model, indicating a random assembly process. Contrary to this, at the forest edge and the interior, co-occurring species showed functional and phylogenetic clustering, thus environmental filtering was the likely process structuring carabid assemblages. Community assembly in the grassland was considerably affected by asymmetrical species flows (spillover) across the forest edge: more forest species penetrated into the grassland than open-habitat and generalist species entered into the forest. This asymmetrical species flow underlines the importance of the filter function of forest edges. As unfavorable, human-induced changes to the structure, composition and characteristics of forest edges may alter their filter function, edges have to be specifically considered during conservation management.

Conversion from environmental filtering to randomness as assembly rule of ground beetle assemblages along an urbanization gradient

Urbanization fragments, isolates or eliminates natural habitats, and changes the structure and composition of assemblages living in the remaining natural fragments. Knowing assembly rules is necessary to support and/or maintain biodiversity in urban habitats. We hypothesized that forest communities in rural sites are organized by environmental filtering, but this may be changed by urbanization, and in the suburban and urban forest fragments replaced by randomly organized assemblages, influenced by the colonization of species from the surrounding matrix. Evaluating simultaneously the functional and phylogenetic relationships of co-existing species, we showed that at the rural sites, co-existing ground beetle species were functionally and phylogenetically more similar than expected by chance, indicating that environmental filtering was the likely process structuring these communities. Contrary to this, in urban and suburban sites, the co-occurring species were functionally and phylogenetically not different from the null model, indicating randomly structured assemblages. According to our findings, changes in environmental and habitat characteristics accompanied by urbanization lead to assemblages of randomly colonized species from the surrounding matrix, threatening proper ecosystem functioning. To reassemble stochastically assembled species of urban and suburban fragments to structured, properly functioning communities, appropriate management strategies are needed which simultaneously consider recreational, economic and conservation criteria.