PATTERN OF ELEVATIONAL DISTRIBUTION AND RICHNESS OF NON VOLANT MAMMALS IN ITATIAIA NATIONAL PARK AND ITS SURROUNDINGS , IN SOUTHEASTERN BRAZIL

Itatiaia National Park (PNI) and its surroundings present a unique fauna due to different forest formations with well-defined climatic and vegetation bands. The Itatiaia massif has four vegetation types that follow an altitudinal gradient: lower montane forest, montane forest, upper montane forest, and Campos de Altitude. Hence, this region is ideal for studying geographical variation in biological diversity. The main objectives of this study were to report on nonvolant mammal species known to occur in Itatiaia National Park and its surroundings and to determine if their distributional pattern is related to elevation. A review of the literature and a complete survey of specimens deposited in museums, as well as small-mammal trapping were carried out in order to obtain a complete record of the species from the region. Precise locality data were obtained for all specimens recorded, allowing the inclusion of each collected or observed individual in an altitude and vegetational class. We made a direct ordination gradient of marsupial, primate, and rodent species abundance with the altitude. Sixty-nine mammal species were collected or reported for the Itatiaia massif, belonging to seven orders and 20 families. Of these, 33 species (47.8%) are included in the official list of threatened or believed-tobe threatened species in Rio de Janeiro State. The orders Rodentia, Carnivora, and Didelphimorphia had the highest species richness, with 25, 14, and 13 species respectively. When species were grouped according to the vegetation, 16 species occured in the lower montane, 56 in the montane forest, five in the upper montane, and 21 in the high-altitude fields (Campos de Altitude). The communities of marsupials, primates, and rodents have an ordination pattern related to the altitude. Species richness was higher between 500 m and 1,500 m above sea level in montane forest, which is in agreement with recent studies showing that species richness can reach its maximum at mid-elevations.


INTRODUCTION
Rio de Janeiro State has been largely ignored as far as the study of its mammals is concerned.In the only extensive study of this type (Tribe, 1987), 77 nonvolant mammal species (excluding Cetacea and considering only native species) were reported.Pereira et al. (2001), also studying nonvolant terrestrial mammals, listed 43 species from the southern coastal region of Rio de Janeiro State while Bueno (1998) found 21 species in Tijuca National Park.Since the beginning of last century, Itatiaia National Park (PNI) has attracted special attention because of its fauna (Ávila-Pires & Gouvêa, 1977;Holt, 1928;Pinto, 1951Pinto, , 1954) ) and the range of different forest formations found there, identified for the first time by Ule (1896).In 1965, Segadas-Viana & Dau showed that the Itatiaia massif presents zones and well-defined climatic and vegetation bands.According to Ururahy et al. (1983), the Itatiaia massif has four vegetation types that follow an altitudinal gradient: lower montane forest (from 400 to 499 m in the study area); montane forest (from 500 to 1,499 m); upper montane forest (from 1,500 to 1,999 m), and high-altitude fields (more than 2,000 m).Recently, Safford (1999) compared the high-altitude fields of some localities in sou-theastern Brazil (Itatiaia, Serra da Mantiqueira, Serra do Caparaó, Serra dos Órgãos, and Serra do Mar) with the páramos of the Andes, and showed a floristic similarity that extends to climate and soil.
The geographic localization of Itatiaia National Park, between two large urban centers, and the fact that the high-altitude fields are islands in a varied and extensive landscape, characterizes this region as a nucleus of the Atlantic Forest Biosphere Reserve, one of the biggest conservation units in the world (Conservation International of Brasil, 2000).
The PNI and its surrounding provide an ideal opportunity for studying geographical variation in biological diversity.Studies of animal distribution along an altitudinal gradient provide useful information on their ecology, habitat preferences, and species replacement (Goodman et al., 1996).The four vegetation types described for the Itatiaia massif present distinct soil and floristic composition (Ururahy et al., 1983).Hence, each species of the mammal fauna is expected to be more strongly associated with one or another type of vegetation.
Besides testing the hypothesis that species from some orders (Didelphimorphia, Primates, and Rodentia) follow a pattern of distribution associated with altitude, we also have made an up-to-date list of nonvolant mammal species known to occur in PNI and its surroundings, along with some karyological information as well as conservation status.

MATERIAL AND METHODS
The field work aimed to increase the number of localities studied in the area and to provide material for correct specific identification (Rodentia, Sigmodontinae).We used banana, manioc, and peanut butter as bait in Sherman live traps.These were set differently depending on development of the research, topography, and the presence of trees, in which they were set at up to 7 m from the ground.The traps were generally placed in lines ca. 5 to 20 m apart.Each line had from 10 to 25 traps, with none or a minimum of five arboreal traps, for a total of 100 traps open at least during three nights.Field work was carried out during the following periods: 12/29/95-1/2/96, 2/9/97-2/12/97, and 6/13/97-6/17/ 97 in Brejo da Lapa; 7/17/97-7/19/97 and 9/29/00-10/2/00 in Serrinha do Alambari for trapping small non-volant mammals; and 8/20/99-8/23/99, 10/31/ 99-11/3/99, and 2/1/00-2/4/00 for direct observations of large mammals and for trapping small mammals at Piscina Maromba.
A review of the literature and a complete survey of specimens deposited in the Museu Nacional, Rio de Janeiro (MN-UFRJ), Museu de Zoologia da Universidade de São Paulo (MZUSP), and Museu do Parque Nacional de Itatiaia were carried out.
Field work was done with a permit issued by the Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis (IBAMA), # 025/99, to H. G. Bergallo and L. Geise, and # 02001.005474/98-96 to D. E. P. Bossi.After the complete analysis, specimens (skull, partial skeleton, and skin) will be deposited in the MN-UFRJ.
Precise locality data were obtained from the United States Board on Geographical Names -Brasil (1963), maps (Instituto Brasileiro de Geografia e Estatística -IBGE, scale 1:25,000 and 1:50,000), and during field work with a GPS and an altimeter.Vegetation types follow Ururahy et al. (1983).
When species identification was not possible by morphological comparison only, karyological data were used.Chromosomes were obtained from bone marrow cultures with cell cultures in MEM Dulbecco's medium (Ford & Hamerton, 1956) with some modifications.Karyological data (diploid number = 2n, fundamental number = FN) for rodents of the subfamily Sigmodontinae (Muridae) are given according to Geise (1995) and Geise et al. (1998).
We made a direct ordination gradient of marsupial, primate, and rodent species abundance with altitude, since these orders were more widely reported in this study, by visualization, trapping, and voucher specimens.Direct gradient analysis focuses on response curves of species with respect to a complex of environment variables that gradually change in geographic space (Ter Braak & Looman, 1995).Each species has its own response to the environmental gradient, and most of them have one optimum along the gradient and decline to either side (Gauch, 1982).
The record of each collected or observed individual was placed in an altitude class, each of which comprised 100 m.The first class comprised altitudes between 400 and 499 m above sea level (a.s.l.); the second one, 500 and 599 m a.s.l.; and so on, up to the last, from 2,200 to 2,499 m a.s.l., representing two classes joined together, as few specimens were collected above 2,399 m a.s.l.To find the position of the species in the altitude gradient, the mean altitude reported for each species was calculated and the species ranked from the lowest to the highest mean.As capture results differed between altitudes, capture frequencies of each species in the altitude classes were transformed into proportions.Bergallo) and LG (L.Geise).To avoid repetition of coordinates and altitude data, only the locality name is given after the first reference to a locality.The symbol for males is ; for females, ; and when sex was indeterminate, I.The conservation status (CS) is given when species are on the official list of threatened or believed-to-be threatened species in Rio de Janeiro State (Bergallo et al., 2000).

Order Didelphimorphia
Family
The orders Rodentia, Carnivora, and Didelphimorphia had the highest species richness, with 25 (36.2%), 14 (20.3%), and 13 species (18.9%), respectively.Those orders with less species were Artiodactyla, with three (4.3%) and Perissodactyla and Lagomorpha, both with one species (1.4%), the last of which is a monospecific order in South America.
When species were grouped according to vegetation, 16 species occurred in the Floresta Sub-Montana; 56, in the Floresta Montana; five, in the Floresta Alto-Montana; and 21, in the Campos de Altitude.Some species occurred in more than one vegetation type, and others occurred in only one.Of the 56 species recorded for the Floresta Montana, 31 were registered only in this vegetation; from the two species listed for the Floresta Alto-Montana, only one species occurred only there; and eight species were specific to the Campos de Altitude (Table 1), of which six species belongs to the sub-Family Sigmodontinae.
Figs. 1, 2, and 3 show that the communities of marsupials, primates, and rodents formed an altituderelated continuum.Marsupials were mainly recorded between 700 and 900 m a.s.l.while only five species were observed above 1,000 m (Fig. 1).Primates were recorded in greater number between the lower level of the park and 1,300 m a.s.l., but none were observed above 1,700 m a.s.l.(Fig. 2).Rodents showed a clear distributional pattern along the altitude gradient, varying from those restricted to lower altitudes to those restricted to higher ones (Fig. 3).Species richness was higher between 800 m and 1,100 m a.s.l.(Fig. 3).However, as species data in this study were not collected with identical methodologies along the altitude gradient, the results must be analyzed with care.

DISCUSSION
The number of species recorded for the Itatiaia massif is similar to the first list for the whole of Rio de Janeiro State (Tribe, 1987).An important study about the mammal fauna of the PNI was carried out by Ávila-Pires & Gouvêa (1977) who listed 62 mammals, of which 47 are nonvolant species.Faunal survey scarcity, taxonomic problems, and difficulties in collecting some orders, explain present lack of knowledge of most Atlantic Forest mammal fauna.However, increasing numbers of specialists and more collecting in the area of PNI and its surroundings have increased knowledge of the local mammal fauna.Consequently, the species listed by Ávila-Pires & Gouvêa (1977) increased by 22, which may also be explained by the increase of taxonomic characters currently used (e.g., karyotypes and DNA sequences), as well as rectification of misidentifications of some species.This difference might have been greater if genera with morphologically similar species (e.g., Brucepattersonius, Smith & Patton, 1999) had been systematically reviewed, allowing a formal species description through genetic techniques, as has been already done with Delomys, now recognized through karyological data to have three different species (Bonvicino & Geise, 1995;Voss, 1993).Some species not listed by us probably occur in the area, such as Mazana nana (Artiodactyla) and Potos flavus (Carnivora), as the PNI is included in the area of their geographic distribution (Eisenberg & Redford, 1999).In the public exhibition of the PNI Museum some specimens of both species are found.While lacking collection data, local people say that these were captured in the park.
A fauna was observed that was peculiar to the Campos de Altitude.Included species (mainly Sigmodontinae rodents) are also observed in other localities with similar habitats, e.g., "Akodon" serrensis, Akodon reigi, and Delomys collinus, found in the National Park of Bocaina, Caparaó, and Serra dos Órgãos (Bonvicino et al., 1997;L. Geise, pers. obs.).Such allopatric distributions can probably be explained by isolation in montane rainforest remnants during climatic dry periods (Montane Isolate Hypothesis, Moreau (1966)) and, if we consider the low divergence rates observed in samples of Akodon montensis between such montane isolates (Geise et al., 2001), they have not resulted in significant phenotypic evolution or speciation, maintaining until now the same species.The Montane Isolates Hypothesis, first introduced by Moureau (1966) and postulating that isolation in montane rainforest remnants during climatic dry periods led to divergence and speciation, has not been explicitly examined in the Atlantic Forest, even though the mountainous topography of the region suggests that montane isolation is a plausible mechanism for geographic isolation.For these taxa, montane isolation almost certainly holds, but it is not clear whether it will continue to drive future speciation.Again, future analyses of DNA variation in small mammals will be useful.Species richness in PNI and its surroundings was higher in Floresta Montana between 500 and 1,500 m a.s.l. with a mid-elevation richness peak.Although it is postulated that species richness is higher in lowland tropical forest (MacArthur, 1972;Patterson et al., 1996Patterson et al., , 1998)), recent studies have showed that species richness can reach its maximum in mid-elevations, yielding a curvilinear pattern along elevational gradients (e.g., for birds, Goodman et al., 1996; for small mammals, Heaney, 2001).Hypotheses explaining elevational changes in species richness vary (Brown, 2001;Rahbek, 1995).Heaney (2001), studying small mammal elevational gradients in the Philippines, suggested that annual rainfall, total abundance of individuals in the community, food resource diversity, areas of reduced competition from other organisms, areas characterized by high rates of speciation, and areas of community overlap were correlated with diversity.Although the terms Floresta Submontana, Floresta Montana, Floresta Alto-Montana and Campos de Altitude are useful for descriptive purposes, what actually exists is a continuum between them.The three orders studied here showed a continuum pattern related to the altitude gradient.The majority of marsupials presented an optimum between 500 and 1,000 m, forming a distribution with few intermediate species.
On the other hand, the rodent community showed all kinds of species distributions, including those with wide distribution and those restricted to a single altitude band.The Order Primates formed an altitude-related continuum pattern, but none of the species were observed above 1,720 m a.s.l.Grelle (2000) observed a 1,500 m maximum limit for endemic primate species from the Atlantic Forest.According to Grelle (2000), the absence of primates in high altitudes there may be due to the tree-line lowering during the glaciations (Safford, 1999).This also explains why only cursorial mammals were found in Campos de Altitude.The low diversity of Didelphimorphia in this area of the PNI shows the necessity of a major sampling effort, requiring especially the use of pitfall traps, since such fauna have already been collected in other Campos de Altitude (L.Geise, pers.obs.; Bonvicino et al., 1997).
A higher richness was observed in the Floresta Montana and the Campos de Altitude.Both Submontana and Alto-Montana areas showed a lower richness, probably due to the fact that the former has a narrow altitude band (400 to 500 m) in Itatiaia and fewer samples were collected in both bands (authors, pers.obs.).This could be observed due to the occurrence of some species in all bands except in Alto Montana (e.g., Cerdocyon thous, Leopardus tigrinus, Galictis cuja, Conepatus semistriatus and Lontra longicaudis).Although specimens of Oligoryzomys nigripes were only collected in Campos de Altitude, and specimens of Philander frenatus were not collected in the submontana region, it is possible that they are present in all vegetational bands since these species can be found at sea level in other regions (e.g., Pereira et al., 2001).
An Akodontini domain was observed in the Campos de Altitude, as half of the Sigmodontinae rodents are from this Tribe.The same pattern was mentioned by Bonvicino et al. (1997) for Caparaó National Park.In the Akodon genus, a species segregation occurs, since A. cursor is found at low altitudes, "Akodon" montensis and A. serrensis at mid-and high altitudes, and Akodon reigi only in high altitudes.Forest fragments in Rio de Janeiro State have been drastically reduced in recent decades (Tanizaki-Fonseca & Moulton, 2000), with no records of original diversity.The peculiarities of the PNI region and constant loss of its vegetation cover by anthropic pressures, such as fire, indicate the need to increase understanding of its biota, through such measures as long-term research programs with mammals, including species of Chiroptera.Such programs, including sporadic captures as well as correct species identification through morphological analyses and using genetic techniques, are necessary to adequately assign correct diversity status in this important area.It is also essential to conserve and protect the pristine Atlantic Forest vegetation as well as the Campos de Altitude ecosystem, where most cases of endemism in the PNI are found.Research on other groups of animals and plants must be promoted as it will improve identification of evolutionarily distinct regions and important ecological attributes so as to develop more informed conservation efforts.

Fig. 1 -Fig. 2 -
Fig. 1 -Distribution of the relative frequency of capture of marsupial species in Itatiaia massif, along an altitudinal axis.

Fig. 3 -
Fig. 3 -Distribution of the relative frequency of capture of rodent species in Itatiaia massif, along an altitudinal axis.Table 1 -Order and species reported for the Itatiaia National Park and its surroundings in different vegetation types.Filled cells indicate the species presence in that vegetation type.

Table 1 -
Order and species reported for the Itatiaia National Park and its surroundings in different vegetation types.Filled cells indicate the species presence in that vegetation type.