Abstracts

THE ROLE OF GALLERY FORESTS IN THE DISTRIBUTION OF CERRADO MAMMALS

JOHNSON, M. A., SARAIVA, P. M. and COELHO, D.

Departamento de Ecologia, Universidade de Brasília, C.P. 04631, CEP 70919-970, Brasília, DF, Brazil

Correspondence to: Marc A. Johnson, Departamento de Ecologia, Universidade de Brasília, C.P. 04631,

CEP 70919-970, Brasília, DF, Brazil, e-mail: johnsonm@guarany.unb.br

Received May 21, 1998 – Accepted October 21, 1998 – Distributed September 10, 1999

(With 1 figure)

ABSTRACT

The Cerrado biome contains a rich mammal community, with an influence from the Amazonian and Atlantic rainforests, principally observed in the gallery forests. In this paper, through literature review, it is shown that the non-volant mammal community of the gallery forests is distinct from the mammal communities of any other physionomy of the Cerrado. Additionally, the gallery forests contain twice as many species common to the rainforests when compared to all the other physiognomies of the Cerrado combined. The gallery forests appear to provide habitat within the Cerrado for rainforest mammals, increasing the biodiversity of this biome. As such, gallery forests may function as dispersion corridors for these species.

Key words: cerrado, dispersion corridors, gallery forests, mammals.

RESUMO

O papel das matas de galeria na distribuição dos mamíferos do cerrado

O bioma dos cerrados contém uma rica comunidade de mamíferos, com influência das matas Amazônica e Atlântica, principalmente observada nas matas de galeria. Neste trabalho, por revisão da literatura, foi demonstrado que a comunidade de mamíferos não voadores das matas de galeria no cerrado é distinta das comunidades de mamíferos de qualquer outro tipo de fisionomia do cerrado. Além disso, as matas de galeria contêm duas vezes mais espécies comuns às matas úmidas que às outras fisionomias do cerrado (sensu latu) reunidas. As matas de galeria parecem fornecer hábitat dentro do cerrado (sensu latu) para mamíferos das matas úmidas, aumentando a biodiversidade deste bioma. Assim, as matas de galeria podem funcionar como corredores de dispersão para estas espécies.

Palavras-chave: cerrado, corredores de dispersão, mamíferos, matas de galeria.

INTRODUCTION

The Cerrado (sensu latu) is the second largest vegetal formation in Brazil, covering an area of approximately 2.0 million km² (Ab’Saber, 1971; Eiten, 1972; Felfili & Silva, 1993; Oliveira-Filho et al., 1989). It is composed of by phytogeographically distinct areas, from open formations such as campos limpos, campos sujos and campos cerrados, to more closed formations such as cerrado (sensu stritu), cerradão and gallery forests (Eiten, 1972). With the vegetation of the Cerrado being markedly heterogeneous, the fauna associated with the various phytophysiognomies of the Cerrado is likewise distributed in a hetergeneous manner (Alho, 1993).

The mammal community of the Cerrado gallery forests appears to be strongly influenced by the Amazonian and Atlantic rainforests (Bishop, 1984; Fonseca & Redford, 1984; Mares et al., 1986; Nitikman & Mares, 1988; Redford & Fonseca, 1986). The term “community” is used here to describe the set of species known to occur within a particular habitat. Several authors have even suggested that the Cerrado gallery forests function as dispersion corridors between the Atlantic and Amazonian rainforests (Mares et al., 1989; Redford & Fonseca, 1986).

The first objective of this paper is to evaluate the degree to which the Cerrado gallery forest mammal community differs from mammal communities of other phytophysiognomies of the Cerrado. Secondly, the respective influences from the Amazonian and Atlantic rainforests over the Cerrado gallery forest mammal community are considered. In so doing, it may be seen as to whether the Cerrado gallery forests actually serve as extentions of suitable habitat for rainforest mammals.

METHODOLOGY

Using information from the literature, a list of non-volant, Cerrado mammal species was constucted (^Appendix Appendix ). As indicated by the literature, species were listed according to their occurrence in the following phytophysiognomies of the Cerrado: gallery forests, cerradão, cerrado (sensu stritu) and campos.

To assess the relative dissimilarity of the gallery forest mammal community in relation to those of other phytophysiognomies of the Cerrado, the occurrence data were analyzed with a Jaccard similarity index, using the program SYSTAT (Wilkinson et al., 1992).

In a second analysis, the role of the gallery forests for mammals common to the rainforests was examined by listing the Cerrado species which occur in either the Amazonian or Atlantic rainforests (see ^Appendix Appendix ). The number of species common to these forests were compared between two groups: the Cerrado gallery forests and all other phytophysiognomies of the Cerrado combined. The data from the cerradão, cerrado (sensu stritu) and campos were combined so as to reduce the effect of incongruencies in the literature as well as a lack of information for some phytophysiognomies.

RESULTS

The Jaccard similarity index revealed values less than 0.50 when the gallery forest mammal community was compared with any other mammal community of the Cerrado (Table 1). The phytophysiognomy most similar to the gallery forests in relation to species composition appeared to be the cerrado (sensu stritu), with a similarity value of 0.462.

Matrix of Jaccard Similarity Index values, comparing non-volant mammal species composition for each of the various phytophysiognomies of the Cerrado (sensu latu)

The gallery forests listed 73 species common to one or both of the humid forests, while, even when combined, the other phytophysiognomies of the Cerrado listed only 39 such common species. (^Appendix Appendix ).

DISCUSSION

In relation to species composition, the mammal communities of the respective phytophysiognomies of the Cerrado clearly differ. The gallery forests harbor the majority of species common to the Cerrado (sensu latu) and the Brazilian humid forests (Fig. 1).

When comparing mammal communities from the various phytophysiognomies of the Cerrado, all values obtained through the Jaccard similarity index were less than 0.50. This suggests that the mammal communities of these phytophysiognomies must be considered as distinct from one another (Muller-Dombois & Ellenberg, 1925).

Our data agree with the conclusions of Redford & Fonseca (1986) in that the Atlantic rainforest has a greater influence over the mammalian species composition of the Cerrado’s gallery forests than does the Amazonian rainforest. Excluding those species which occur in both the Atlantic and Amazonian rainforests, 19 species were found to be common to the Cerrado gallery forests and the Atlantic rainforest, whereas only 8 gallery forest mammals were common to the Cerrado gallery forests and the Amazonian rainforest. A similar pattern of disproportionately high influence from the Atlantic rainforest may be observed in the other phythophysiognomies of the Cerrado (Figure 1). These results were not expected, as the area of the Amazonian forest (6,000,000 km²: Pires & Prance, 1988) is much greater than the area of the Atlantic rainforest (800,000 km²: IBGE 1982).

The the mammal community of the gallery forests appeared to be more similar to that of the cerrado (sensu stritu) than to those of any other phytophysiognomy of the Cerrado (sensu stritu), as was found by Alho et al. (1986), considering 25 mammal species. The lack of mammal studies in the cerradão may be considered responsible for the low similarity value between this phytophysiognomy and the gallery forests, considering that cerradão is structurally most similar to gallery forest.

Mares et al. (1986) found the greatest similarity, in relation to rodent species composition to be between campos and cerrado (sensu stritu), and that the gallery forests contained the most distinct rodent community. These results may differ from our conclusions due to the more restricted analysis used by Mares et al. (1986), or by particularities of the study site sampled by Mares et al. (1986).

Fry (1980) found that the avifauna of the cerradão is more similar to the cerrado (sensu stritu) than to that of the gallery forest. Our data considering mammals show values of 0.194 between the gallery forest and cerradão, and 0.281 between cerrado and cerradão. These values suggest that the same pattern found by Fry (1980) for the distribution of avifauna within the Cerrado may also apply to mammals.

The gallery forests considerably boost the diversity of the Cerrado non-volant mammal fauna, providing habitat for species which do not occur in other phytophysiognomies of the Cerrado (Redford & Fonseca, 1986). Additionally, 85% of the bat species found in the Cerrado depend upon gallery forests in some manner (Marinho-Filho & Reis, 1989).

In conclusion, the gallery forests provide habitat for a distinct mammal community within the Cerrado. These riverine forests appear to be the only favorable habitat for many species common to the Amazonian and Atlantic rainforest. For these species, the gallery forests likely serve as the primary, if not only, dispersion corridors within the Cerrado.

Further investigations are needed in order to reveal the extent to which the Cerrado gallery forests have historically served to link the Amazonian and Atlantic rainforests.

APPENDIX

Appendix

Publication Dates

History