Maria Nazareth F. da Silva

MAMMALS OF THE RIO JURUÁ AND THE EVOLUTIONARY AND ECOLOGICAL DIVERSIFICATION OF AMAZONIA

Abstract We describe the nonvolant mammal fauna of the Rio Juruá of the western Amazon of Brazil, based on collections made during a year-long survey of the river. We, along with our colleagues Drs. Claude Gascon and Carlos Peres, designed the field project to examine the effects of the river on the differentiation among terrestrial vertebrates (mammals, birds, and amphibians and reptiles) at both the community and population levels. This monograph examines only the patterns of geographic variation and community structure of the small-bodied mammals. Species inventories were made at 16 primary trapping localities divided into eight pairs of cross-river sites, with two pairs in each of four regions from near the mouth to the headwaters of the Rio Juruá. A total of 81 species of nonvolant mammals were obtained, including nine new to science. Four of these are described herein; the others have been described elsewhere. We used a standardized trapping protocol to assess community structure at each of the 16 localities that included terrestrial and canopy trap stations in floodplain (várzea) and upland (terra firme) forest formations. Supplemental trapping was done in secondary habitats at all sites. We describe these sites, the trap effort expended, and the placement of trap stations relative to local habitats. We also describe each species of marsupial, sciurid rodent, murid rodent, and echimyid rodent encountered; comment on their systematics; and summarize aspects of habitat use, life history, geographic distribution, and geographic differentiation based on morphological and molecular traits. We examine patterns of differentiation in the mitochondrial cytochrome-b gene for samples of 41 of the 45 species of marsupials and rodents obtained within the Rio Juruá Basin, and discuss these patterns from the perspective of the entire Amazon and, in some cases, the Mata Atlântica of coastal Brazil. We also examine patterns of community organization within the Rio Juruá basin and throughout Amazonia, drawing attention to the geographic distribution of what appear to be major faunal units that are independent of habitat differences. Finally, we use principles of phylogeography to analyze patterns of geographic differentiation among the nonvolant mammals with regard to the Riverine Barrier Hypothesis. We show that, while there are few examples of taxa for which the Rio Juruá is apparently a barrier, most taxa either are largely undifferentiated throughout the basin or are sharply divided into reciprocally monophyletic mtDNA haplotype clades separable into upriver and downriver units. We argue that the concordance in the geographic placement of clade boundaries suggests a common history; moreover, both the age of these clades and their geographic position in relation to underlying geological features suggest that landform evolution has been an important, but underappreciated component of diversification within western Amazonia.

GENE GENEALOGY AND DIFFERENTIATION AMONG ARBOREAL SPINY RATS (RODENTIA: ECHIMYIDAE) OF THE AMAZON BASIN: A TEST OF THE RIVERINE BARRIER HYPOTHESIS

Sequence variation in the mitochondrial cytochrome b gene was examined in the arboreal spiny rat, Mesomys hispidus, collected at 15 sites along the Rio Juruá in western Amazonia, Brazil, to determine the importance of riverine barriers in the diversification of this taxon. Twenty individual haplotypes were uncovered, most of which were unique to single localities but some of which were shared among adjacent sites either along or across the river. Genealogical analyses suggest that gene flow is limited and, in combination with the unique distribution of most haplotypes, suggest that populations of this species are strongly substructured along the river. Thus, most sharing of haplotypes between adjacent localities is probably caused by historical association rather than to ongoing gene flow. Two haplotype clades were uncovered, but these correspond to headwaters versus mouth areas, not to opposite sides of the river, as would be expected by the Riverine Barrier Hypothesis. Moreover, haplotype sharing across the river was greater at its mouth than in the headwaters, a pattern opposite that expected if the river were a substantive barrier. Broader scale phylogeographic patterns of this species show that both clades have relationships to areas well outside the Rio Juruá basin. This suggests that the basin represents a relatively recent point of invasion between two more broadly distributed and differentiated geographic units of the species.

Molecular phylogeography and the evolution and conservation of Amazonian mammals

The phylogeographic structure of 15 genera of Amazonian marsupials and rodents is summarized based on comparative sequence of the mitochondrial cytochrome b gene. The data are limited in geographical coverage, with samples widely scattered throughout Amazonia from the base of the Andes in Peru to the Guianan coast and eastern Brazil. We use this approach to define species boundaries, based minimally on the principle of reciprocal monophyly, in conjunction with morphological or other genetic discontinuities. The taxa so defined are older than previously appreciated, with many lineages dating from 1 to more than 3 Myr, and thus apparently predating the early Pleistocene. We relate patterns of concordant geographical shifts with underlying tectonic history and to current positions of major rivers. Finally, we provide comments on the utility of these data and patterns to conservation, articulating a need to incorporate phylogeographic information as part of the rationale in establishing conservation priorities at the organismal and geographical area levels.

Riverine Barriers and Gene Flow in Amazonian Saddle-Back Tamarins

We describe patterns of genotypic and phenotypic variation in saddle-back tamarin (Saguinus fuscicollis) populations along the central and upper Rio Juruá, western Brazilian Amazonia. The genetic data are sequence haplotypes of the mitochondrial cytochrome b gene; phenotypic data are pelage colour variants that define sharply demarcated subspecies of this extremely variable tamarin species. We show that gene flow occurs between adjacent subspecies, but that this phenomenon is restricted to the headwater section of the river, which is consistent with expectations from the riverine barrier hypothesis. In this model, the major first-order tributaries of the Amazon form effective barriers to dispersal, with between-bank gene flow limited to the narrowed sections of headwater streams and parallel divergence increasing along both banks from the headwaters to the mouth of a given river. In meandering rivers such as the Rio Juruá, we suggest passive transfer through river channel dynamics as the main mechanism permitting genetic contact between populations on opposite banks of the river. Finally, we argue that in the case of plant and animal species that are largely restricted to unflooded (terra firme) forests, such as tamarins, seasonally flooded (várzea) forest can operate as a critical additional barrier to between-bank gene flow.

Relationships among didelphid marsupials based on sequence variation in the mitochondrial cytochrome B gene

Variation in the mitochondrial cytochrome b gene (nucleotide and amino acid sequences) is evaluated for 9 genera and 15 species of American opossums in the family Didelphidae, using the American caenolestid rat opossumLestoros and the New Guinean peroryctid bandicootEchimypera as outgroups. Phylogenetic analyses (parsimony and distance) strongly support the monophyly of the Didelphidae and delineate two major clades; (1)Didelphis andPhilander are strongly aligned sister taxa, withMetachirus weakly but consistently associated with them, and (2)Marmosa plusMicoureus, withMonodelphis falling outside that pair. The generaMarmosops, Caluromys, andGlironia exhibit varied relationships, depending upon the method of analysis and data (DNA or amino acid sequences) used, but generally are placed individually or in combinations near or at the base of the didelphid radiation. Some aspects of these relationships are consistent with current taxonomic views, but others are in marked contrast. Specifically, a clade comprised of the mouse opossumsMarmosa, Micoureus, andMarmosops is strongly rejected by log-likelihood analysis, contrary to expectations from some current classifications. Also, the woolly opossumsCaluromys andGlironia also do not form a sister-taxon relationship, as suggested by their placement in a subfamily separate from the remaining didelphids examined. However, such a relationship cannot be rejected from log-likelihood analyses. The relationships suggested fromcyt-b sequences are strongly concordant with those based on DNA-DNA hybridization analyses. In addition to systematic and phylogenetic properties, molecular evolution of the didelphid cytochrome b gene sequence is characterized according to nucleotide bias and rate differentials at each codon position and across the entire sequence.

POPULATION GENETIC STRUCTURE OF TWO ECOLOGICALLY DISTINCT AMAZONIAN SPINY RATS: SEPARATING HISTORY AND CURRENT ECOLOGY

Abstract. Population history and current demographic and ecological factors determine the amount of genetic variation within and the degree of differentiation among populations. Differences in the life history and ecology of codistributed species may lead to differences in hierarchical population genetic structure. Here, we compare patterns of genetic diversity and structure of two species of spiny rats in the genus Proechimys from the Rio Jurua of western Amazonian Brazil. Based on the ecological and life‐history differences between the two species, we make predictions as to how they might differ in patterns of genetic diversity and structure. We use mitochondrial sequence data from the cytochrome b gene to test these predictions. Although both species maintain nearly the same number of mitochondrial haplotypes across the sampled range, they differ in levels of genetic diversity and geographic structure. Patterns of gene flow are also different between the two species with average M‐values of nearly three in P. steerei and less than one in P. simonsi. Our initial predictions are largely upheld by the genetic data and where conflicting hypotheses arise, we suggest further studies that may allow us to distinguish among evolutionary scenarios. Separating the effects of history and ongoing demography on patterns of genetic diversity is challenging. Combining genetic analyses with field studies remains essential to disentangling these complex processes.

A Taxonomic Reassessment of Cacajao melanocephalus Humboldt (1811), with the Description of Two New Species

The author of the last published systematic review of Cacajao recognized 2 subspecies of black-headed uakaris (black uakaris): Cacajao melanocephalus melanocephalus and C. m. ouakary. As a result of a series of black uakari surveys and collecting expeditions to several tributaries of the Rio Negro and of morphological and molecular analyses of museum specimens and specimens we collected during field expeditions, we reassess their taxonomy. We describe a newly discovered species of black uakari from the Rio Aracá, a left bank tributary of the Rio Negro, Amazonas, Brazil. We also show that ouakary is a junior synonym of melanocephalus and provide a new name and a new description for Cacajao melanocephalus melanocephalus in the Pico da Neblina region of Brazil and Venezuela. Based on genetic, morphological, and ecological evidence, we propose that there are 3 species of black uakaris. We named the Rio Aracá species Cacajao ayresi sp. nov. (Ayres uakari) in honor of the late José Márcio Ayres, a pioneer in uakari research and conservation. We named the Neblina black uakari Cacajao hosomi, after the Yanomami word for uakaris. The new taxonomic arrangement provided here implies that the conservation status of black uakaris needs to be reassessed.

Spatial and temporal patterns of diversification on the Amazon: A test of the riverine hypothesis for all diurnal primates of Rio Negro and Rio Branco in Brazil.

The role of Amazonian rivers as drivers of speciation through vicariance remains controversial. Here we explore the riverine hypothesis by comparing spatial and temporal concordances in pattern of diversification for all diurnal primates of Rio Negro and its largest tributary, Rio Branco. We built a comprehensive comparative phylogenetic timetree to identify sister lineages of primates based on mitochondrial cytochrome b DNA sequences from 94 samples, including 19 of the 20 species of diurnal primates from our study region and 17 related taxa from elsewhere. Of the ten primate genera found in this region, three had populations on opposite banks of Rio Negro that formed reciprocally monophyletic clades, with roughly similar divergence times (Cebus: 1.85 Ma, HPD 95% 1.19-2.62; Callicebus: 0.83 Ma HPD 95% 0.36-1.32, Cacajao: 1.09 Ma, 95% HPD 0.58-1.77). This also coincided with time of divergence of several allopatric species of Amazonian birds separated by this river as reported by other authors. Our data offer support for the riverine hypothesis and for a Plio-Pleistocene time of origin for Amazonian drainage system. We showed that Rio Branco was an important geographical barrier, limiting the distribution of six primate genera: Cacajao, Callicebus, Cebus to the west and Pithecia, Saguinus, Sapajus to the east. The role of this river as a vicariant agent however, was less clear. For example, Chiropotes sagulata on the left bank of the Rio Branco formed a clade with C. chiropotes from the Amazonas Department of Venezuela, north of Rio Branco headwaters, with C. israelita on the right bank of the Rio Branco as the sister taxon to C. chiropotes+C. sagulata. Although we showed that the formation of the Rio Negro was important in driving diversification in some of our studied taxa, future studies including more extensive sampling of markers across the genome would help determine what processes contributed to the evolutionary history of the remaining primate genera.

New distribution limits of Bassaricyon alleni Thomas 1880 and insights on an overlooked species in the Western Brazilian Amazon

No abstract available

Reassessment of the occurrence of the kinkajou (Potos flavus Schreber, 1774) and olingo (Bassaricyon beddardi Pocock, 1921) in the northern Brazilian Amazon

Brazils only records of Pococks olingo (Bassaricyon beddardi) are based on sightings from the northernmost state of Roraima, where the similar kinkajou (Potos flavus) was reported as absent. Our recent field work in the region led to the collection of two specimens and several more observations of kinkajous and a complete lack of evidence of the presence of olingos. Furthermore, the name used locally to describe the nocturnal procyonids previously treated as olingos, gogó de sola, refers to the leathery bare throat patch that we believe to be a characteristic unique to kinkajous. Thus, we conclude that previous records of olingos in Roraima represent misidentifications of kinkajous and recommend that, until supported by a specimen, B. beddardi be treated as absent from Brazil.