Mammals in a fragmented savannah landscape in south-western Brazil

Abstract

Mammal abundance, richness, and community structure were examined in fragments of savannah and in gallery forests of south-western Brazil in order to assess the influence of fragment size and habitat on the communities. Five savannah fragments and two gallery forests were sampled. Within the savannah fragments, environmental parameters such as tree and shrub densities were measured. Live-traps and pitfalls were used for trapping small mammals; larger mammals were recorded from footprints and sightings. Seventeen species of small mammal and 15 species of large mammal were recorded. Small mammals were affected by fragment size and isolation, whereas larger mammals were not. Communities of small mammals were better structured in the largest fragments, and those in smaller fragments were subsets of the communities in larger fragments. Gallery forests revealed a rather different assemblage of mammals, but also contained species common to both gallery and large savannah remnants. Data are discussed relating mammal community structure to area size and vegetation structure, in an attempt to understand the present conservation status of this fragmented landscape.

Keywords:

• biogeography
• Cerrado
• community structure
• conservation
• fragmentation
• isolation
• species richness
• matrix
• nestedness subsets

Acknowledgements

Our thanks to FUNDECT/MS/Brazil for financial support, and to the Universidade Federal do Mato Grosso do Sul (PPG-Ecologia e Conservação/CCBS) and the owners and staff of the farms São Cristóvão (Mr. Aldino Girotto), Santa Helena (Mr. Preto), Nova Aliança (Mr. Carlos Gandini), and Cachoeirão (Mr. Iasuo Kasai) for logistic support. We thank A.P. Carmignotto (Museum of Zoology/USP; UFSCAR) and A.R. Percequillo (ESALQ/USP) for helping in the species identifications, and L.Z. Prates, C.F. Vargas, and C.S. Goulart for helping in the field. We thank the anonymous referees for their critical reviews of the manuscript, R.D. Loyola for helping in the NODF analysis, and Janet W. Reid for the English improvement. NCC is a CNPq research fellow.