Esta dissertação fez parte do projeto temático “Conservação da Biodiversidade em Paisagens Fragmentadas no Planalto Atlântico de São Paulo", que teve como objetivo geral estudar os efeitos da fragmentação da Mata Atlântica sobre diversas comunidades e processos ecológicos na região de Caucaia do Alto, SP. A dissertação está dividida em três capítulos centrais e duas abordagens principais. A primeira delas utilizou dados coletados por mim e outros pesquisadores com o objetivo de testar o uso de uma metodologia de captura de pequenos mamíferos relativamente nova e pouco utilizada nos trópicos, comparando a sua eficiência com a da metodologia mais tradicional de coleta de dados sobre estes animais (capítulo 2). Estudos recentes demonstram que a matriz, conjunto de ambientes alterados que circundam os remanescentes em paisagens fragmentadas, considerada inicialmente inóspita e homogênea, funciona como um mosaico de unidades com diferentes permeabilidades ao deslocamento ou à ocorrência das espécies. Dentro deste contexto, a segunda abordagem desta dissertação corresponde à descrição e comparação do uso, pelos pequenos mamíferos, dos ambientes alterados que envolvem os remanescentes de floresta (capítulo 3) e à utilização deste tipo de informação para o estudo da influência da estrutura da paisagem sobre a distribuição de espécies (capítulo 4). No capítulo 2, para avaliar a eficiência e a congruência entre tipos de armadilhas, amostramos a comunidade de pequenos mamíferos de forma padronizada usando armadilhas Sherman e de queda em 26 sítios, 20 localizados em fragmentos florestais e seis em floresta contínua. Os resultados sugerem que armadilhas de queda, quando constituídas por baldes grandes e profundos e utilizadas nas épocas chuvosas e quentes, não só complementam a diversidade encontrada com o uso de armadilhas tradicionais, como podem triplicar a riqueza por sítio de amostragem, incluindo espécies raras, de hábitos semi-fossoriais e espécies escansoriais/arborícolas geralmente não atraídas pela isca das armadilhas tradicionais. Apesar das diferenças de eficiência, os resultados sugerem que os dados obtidos com os dois tipos de armadilhas são congruentes e que variações de diversidade e abundância entre sítios são semelhantes entre os métodos. As armadilhas de queda também parecem ser essenciais em estudos demográficos, já que capturam indivíduos em uma amplitude maior de peso, incluindo os jovens, raramente capturados nas armadilhas tradicionais. No capítulo 3, para avaliar a qualidade da matriz para os pequenos mamíferos na região de Caucaia do Alto, utilizamos os dados de amostragens padronizadas realizadas em 23 sítios, sete localizados em remanescentes florestais (floresta contínua e fragmentos) e 16 distribuídos nos quatro tipos predominantes de habitats da matriz (vegetação nativa em estádios iniciais de regeneração, plantações de eucalipto, áreas de agricultura e áreas rurais com construções). Houve forte segregação na comunidade de pequenos mamíferos entre a vegetação nativa e demais habitats, indicando a capacidade extremamente reduzida das espécies florestais de ocupar habitats antropogênicos. A vegetação nativa em estádios iniciais apresentou uma razoável permeabilidade para as espécies florestais, sugerindo o potencial da regeneração natural para a restauração de paisagens fragmentadas de Mata Atlântica. Os habitats antropogênicos foram dominados por espécies generalistas, exóticas ou típicas de biomas abertos, sugerindo que paisagens muito alteradas possam ser dominadas por estas espécies, com conseqüências para a regeneração das florestas e para a saúde humana. No capítulo 4, a partir dos resultados da distribuição das espécies de pequenos mamíferos nos habitats do mosaico, informações sobre a qualidade destes habitats foram utilizadas na avaliação da influência da estrutura da paisagem sobre a ocorrência das espécies em 20 fragmentos florestais, com o objetivo de comparar o poder explanatório de métricas que consideram ou não a qualidade da matriz e investigar a importância da escala espacial. Os resultados indicam que a influência da estrutura da paisagem é maior para as espécies que não ocorrem na matriz, ou entre as que ocorrem, para aquelas que não ocupam todos os ambientes da paisagem. Em geral, os modelos com as métricas que consideram a qualidade da matriz apresentaram maior poder explanatório sobre a distribuição das espécies nos fragmentos. Dentre as métricas que consideram a qualidade da matriz, aquelas que incorporam a distância entre as diferentes manchas de habitat mostraram vantagens em relação às que consideram apenas a área das manchas, pois apresentaram maior poder explanatório sobre a ocorrência de um maior número de espécies e menor variação do poder explanatório em relação a mudanças na escala espacial considerada. Esta ultima característica é importante dada a falta de uniformidade observada na resposta das espécies à variação da escala espacial e a escassez de informações que embasem a escolha de escalas espaciais adequadas para espécies tropicais. / This dissertation was developed as part of the thematic project “Biodiversity Conservation in Fragmented Landscapes at the Atlantic Plateau of São Paulo", which aimed at studying the effects of the fragmentation of the Atlantic forest on several communities and ecological processes in Caucaia do Alto, SP. This dissertation is divided in two main approaches that comprise three chapters. In the first approach, I used data collected by myself and by other researchers with the objective of testing the use of a relatively new and little explored methodology to capture small mammals in the tropics, comparing its efficiency to the methodology that is tradicionally ised to collect data about these animals (chapter 2). Recent studies have shown that the matrix of altered habitats surrounding remnants in fragmented landscapes, considered initially as inhospitable and homogeneous, act as a mosaic of units presenting different degrees of permeability to the movement or to the occurrence of species. Within this context, the second approach of this dissertation concerns the description and comparison of use of the altered habitats surrounding forest remnants by small mammals (chapter 3) and to the use of this type of information for the study of the influence of landscape structure on species distribution (chapter 4). In chapter 2, aiming at the evaluation of the efficiency and congruence of different types of traps, we sampled the small mammal community using a standardized procedure, with Sherman and pitfall traps in 26 study sites, 20 located in forest fragments and six in the continuous forest. Results suggest that large and deep enough pitfall traps used during wet and warm periods, not only complement the diversity found using traditional traps, but also leads to an considerable increase in the number of species found at each study site, including rare species, semi-fossorial species and scansorial/arboreal species that are usually not attracted to the bait used in traditional traps. Despite differences in efficiency, results suggest that data gathered with the two types of traps are congruent and that the variation in diversity and abundance among study sites is similar between the two methods. Pitfall traps also seem to be essential in demographic studies because they are able to capture individuals within a larger range of weights, including young individuals, which are rarely captured using traditional traps. In the chapter 3, to evaluate matrix quality for the small mammals in Caucaia do Alto, we used data from standardized sampling in 23 study sites, seven of which were located in forest remnants (continuous and fragmented forest) and 16 were distributed in the four main types of matrix habitats (native vegetation in initial stages of regeneration, eucalyptus plantation, areas of agriculture and rural areas with buildings). There was a strong segregation in the small mammal community between native vegetation and the other habitats, indicating the extremely low capacity of forest species to occupy anthropogenic habitats. Native vegetation in initial stages presented a reasonable permeability to forest species, suggesting the potential of natural regeneration to restore fragmented Atlantic forest landscapes. Anthropogenic habitats were dominated by generalist species, exotic species or species that are typical from open biomes, suggesting that highly altered landscapes may be dominated by these species, with consequences to forest regeneration and to human health. In chapter 4, using the results of the distribution of the small mammal species in the habitats of the mosaic, information on the quality of these habitats was applied in the evaluation of the influence of landscape structure on the occurrence of species in 20 forest fragments, aiming to compare the explanatory power of metrics that consider or not matrix quality and to investigate the importance of spatial scale. Results indicate that the influence of landscape structure is stronger for species that do not occur in the matrix, or for species that do not occur in all habitats in the mosaic. In general, the models using metrics that consider matrix quality presented higher explanatory power on the distribution of species. Among the metrics that consider matrix quality, those that incorporate distance among different habitat patches presented advantages in relation to those that consider only patch area, because they presented a higher explanatory power to the occurrence of a larger number of species and lower variation in the explanatory power in relation to changes in spatial scale. The later characteristic is important given the lack of uniformity observed in species response to variations in spatial scale, and the scarcity of information that could support the choice of adequate spatial scales for tropical species.

An investigation of the effects of commercial mechanised selective logging on rain forest vegetation and mammals, was undertaken in the Lope Reserve, central Gabon, between January 1989 to July 1991. Vegetation in Lope is mostly semi-evergreen lowland tropical rain forest, but there are some localised patches of savanna., which are thought to be natural in origin, but which are maintained today by regular fires started by humans. Study sites were established in areas of forest logged 20-25, 10-15, and 3-5 years previously, a fourth was logged during the study, and a fifth remained unexploited. None of the study sites had been subject to hunting in the recent past. A line-transect five kilometres in length was cut across the drainage in each site. Forest composition and structure was assessed along each transect, by identifying and measuring trees and lianes in botanical plots, counting stem density of herbaceous vegetation in the families Marantaceae and Zingiberaceae, and by measuring canopy cover at three heights. At total of 4885 trees and lianes of 327 species occurred in five 2.5 ha samples of plants greater than 10 cm dbh, whilst 1832 individuals of 137 species were found in five 25ha samples of trees greater than 70cm dbh. There were marked differences in structure and species composition both between and within sites. A model was developed to attempt to explain this variation, based upon the theory that much of Lope had been covered by savanna vegetation during a previous cool, dry climatic phase, and that forest structure and composition reflected recolonisation of the savanna by forest. Physical features such as swamps, rocky outcrops and altitude were also considered. Two types of multivariate analysis were applied to botanical data and supported the model. The effects of logging on forest vegetation were assessed by returning after logging to botanical plots established before exploitation. Damage levels were low, compared to other parts of the World, resulting in about a 10% reduction in canopy cover. Patterns of fruit production were studied by counting fallen ripe and unripe fruit on transects. Fruits encountered were classified on the basis of their morphology and dispersal mechanism. There was a period of low fruit production during the major dry season, when frugivores are likely to suffer dietary stress. A number of plant species which did produce fruit at this time were identified as species which might represent 'keystone' resources. Over 70% of fruit species were animal-dispersed, demonstrating the the important role animals play in the ecology of tropical rain forests. Forest elephant (Loxodonta africana cyclotis) diet, ecological role and group structure were analysed. The bulk of the diet consisted of the bark and leaves of trees, and some monocotyledons in the families Marantaceae and Zingiberaceae, but fruit was also important. Large-scale seasonal movements in response to fruit availability were detected. Elephants were important seed dispersers f o r many plant species, and were responsible for less than 1% of natural tree mortality. The social structure of these forest elephants differed from that of populations that have been studied i n east and southern Africa. Average group size was 2.8, and no groups of more than 10 individuals were encountered. Densi ties of primates, ungulates and squirrels were assessed using standard line-transect censuses. Resolution was poor, but statistical differences were detected between sites for some species. Chimpanzee, (Pan t. troglodytes) densities declined in logged forest, but no other species could be shown to decline after logging. Some other differences between sites were related to vegetation composition. Biomass was high, estimated a t up to 4692.6 kg km-2, but was dominated by elephants, which made up 25-82%. Conservation implications of this study are discussed, and recommendations made.

Os roedores Cricetidae (Myomorpha, Muroidea) são representados por cinco subfamílias: Arvicolinae, Cricetinae, Neotominae, Tylomyinae e Sigmodontinae. Os Sigmodontinae, por sua vez, são os representantes sul-americanos com a quase totalidade das espécies pertencentes a esta subfamília. Representa o mais diversificado grupo de mamíferos da América do Sul, o que significa aproximadamente 20% da mastofauna do continente. Algumas contribuições marcaram as pesquisas paleontológicas dos Sigmodontinae sul-americanos, como os registros do naturalista dinamarquês Peter Lund, que coletou grande número de material nas grutas e cavernas de Lagoa Santa, Minas Gerais. Todavia, os dados paleontológicos são escassos quando comparados à diversidade atual dos roedores sigmodontíneos. Com relação à região do Parque Nacional Serra da Capivara, no estado do Piauí, conhecida pelas descobertas arqueológicas e paleontológicas, numerosos restos de mamíferos de pequeno porte foram recuperados de grutas e abrigos calcários, entre eles os roedores Sigmodontinae, objeto do presente trabalho Os materiais estudados provenientes de três localidades da região (Toca do Gordo do Garrincho, Toca do Barrigudo e Toca do Serrote das Moendas) foram identificados e descritos. São registrados aqui: Bibimys labiosus, Necromys lasiurus, Cerradomys sp., Holochilus sciureus, Pseudoryzomys simplex, Calomys sp. e Wiedomys sp. Aspectos ecológicos e biogeográficos de cada táxon também foram considerados para discutir aspectos paleoambientais da região, destacando-se o primeiro registro de B. labiosus para o Quaternário do nordeste brasileiro, sugerindo uma distribuição paleobiogeográfica diferente da atualmente conhecida para espécie. Assim, contribuímos com novos registros de sigmodontíneos para a Serra da Capivara, ampliando o conhecimento sobre o grupo no Quaternário do nordeste do Brasil e mostrando que os materiais recuperados na região apresentam grande potencial para fornecer esclarecimentos sobre a dinâmica biogeográfica dos Sigmodoninae ao longo do Quaternário. / The Cricetidae rodents (Myomorpha/Muridae) are composed by five groups: Arvicolinae, Cricetinae, Neotominae, Tylomyinae and Sigmodontinae. The Sigmodontinae are the South American representatives with almost all Cricetidae species in this subfamily. They are the most diverse mammalian taxon in South America, which means about 20% of South American mammalian fauna. Some important contributions marked the Sigmodontinae paleontological researches in south American, as the records of the Danish naturalist Peter Lund, who collected numerous materials in Lagoa Santa caves, Minas Gerais. However, the paleontological data are scarce when compared to current sigmodontinae diversity. In relation to the region of Serra da Capivara National Park, State of Piauí, is known by archaeological and paleontological researches, and there many remains of small mammals were recovered from caves and rock shelter, between them the Sigmodontinae rodents are the object of this study. Remains of the Sigmodontinae from three shelters in the region (Toca do Gordo do Garrincho, Toca do Barrigudo e Toca do Serrote das Moendas) that were identified and described. The taxa identified is: Bibimys labiosus, Necromys lasiurus, Cerradomys sp., Holochilus sciureus, Pseudoryzomys simplex, Calomys sp. e Wiedomys sp. Ecological and biogeographical feature from each taxon were evaluated to discuss the feature paleoenvironment from region, with the first quaternary record from B. labiosus in Brazilian northeast, that suggest a different paleobiogeographic range from currently know. Therefore, we contributed through new Sigmodontinae records from Serra da Capivara, expanding the quaternary sigmodontinae knowledge from Brazilian northeast and showing that there are a large potential for clarify the Sigmodontinae biogeographic dynamic.

This work concerns the use of engineered genetic tools to build maps of the mammalian nervous system. Within the practice of circuit neuroscience, one of the most effective tools to emerge in recent years are the neurotropic viruses. Among these are modified strains of rabies virus which are made safe for laboratory use. We introduce here a novel form of engineered rabies virus with substantially improved utility for exploring the structure and function of neural circuits. Additionally, using this new tool, an investigation of an important motor circuit, the cortico-striatal circuit, is presented.

Orientador: Nivar Gobbi / Banca: Katia Cury Roselli / Banca: Eliana Ferraz Santos / Banca: Ricardo Tadeu Santori / Banca: Sônia Aparecida Talamoni / Resumo: O estudo das características microscópicas dos pêlos de mamíferos tem sido desenvolvido com o propósito de utilizá-las como critério taxonômico, especialmente para a identificação de restos presentes em fezes, conteúdos estomacais e bolotas de regurgitação de predadores. Os pêlos são menos danificados no processo digestivo, e são freqüentemente a única prova para identificação da presa em estudos de hábito alimentar, já que ossos e dentes sofrem intensa fragmentação durante a mastigação apresentando baixo valor diagnóstico, portanto um conhecimento maior da estrutura do pêlo facilita esses estudos e oferece uma base para estudos mais completos de predação e seus efeitos sobre as espécies de presa. Assim, os objetivos do presente projeto foram o desenvolvimento de técnicas de preparação de lâminas de pêlos para coleções e observações diretas; o desenvolvimento de uma chave de identificação de alguns roedores de mata atlântica baseada em pêlos; e a análise morfométrica dos pêlos de algumas espécies de roedores para a comparação com dados biométricos. Foram desenvolvidas 3 novas técnicas para a preparação de lâminas para a observação da medula de pêlos e outras 3 para a observação do padrão cuticular. Essas técnicas proporcionam um maior índice de sucesso na preparação de lâminas, aumentando também a visibilidade dos padrões e consequentemente as possibilidades de pesquisas na área. / Abstract: The mammal hair identification has been used across the world for studies in forensics, in the field of public health, in textile fibers research, and in mammals feeding ecology, being useful on the identification of leftovers on feces and stomach contents. Hairs are the less damaged structures by digestive processes and nonetheless are the only clue for the identification of the prey in studies of feeding ecology of predators, since bones and tooth suffer heavy damage during the chewing process, being of a low diagnostic value. These factors show the importance of a better knowledge of the hair structure for prey or predator ecology studies. Also the hair trapping and identification has a important value in animal research bioethics, since the mammal surveys can be developed with hair-traps or scats/pellets without any harm or interference for the animals which have no risk to get addicted to the bait, which usually bias surveys, and minimizing the stress for being caught. The aim of the present thesis were the development of better techniques for the preparation of hair plates for collections and direct observations; the development of a identification key for Rain Forest rodents; and the morphometrical analysis of those mammals hairs for further comparison and analysis with biometrical data. There were developed 6 new techniques, being 3 for medulla pattern observation, and another 3 for cuticular impressions. Those techniques showed a higher success index on plate preparation, enabling the development of high level research on this field. / Doutor

It has been well documented that protein calorie malnutrition (PCM) gives rise to physiological and behavioral deficits. These deficits include changes in emotional, exploratory and social behaviors of the malnourished organism. In particular, previous research has demonstrated that Feci from infancy results in avoidance of and failure to initiate social interactions as well as decreased contact with the environment, which in turn, further disrupt emotional and social development. This study examined the effects of chronic protein malnutrition on the social behavior of adult rhesus macaques by experimentally testing the hypothesis that deficient monkeys, unlike normal well-fed ones, are more likely to avoid social encounters than to seek them out. In addition, the animals' social interactions were recorded and analyzed. Subjects consisted of eleven adult rhesus monkeys (Macaca mulatta). Five were fed a protein-deficient diet (3.0 % of total kilocalories) and six were fed adequate amounts of protein (14 % of total kilocalories) from birth. After habituation to the test apparatus, subjects were trained to perform an operant response which opened a sliding door. During training, the response allowed access to food. In the final testing, opening the door allowed access to a social partner. If the subject performed the response and released a social partner, the social behaviors of the pair was recorded for ten minutes. Each subject was given three opportunities, on three separate occasions, to release every other subject. Protein-deficient subjects habituated to the experimental apparatus and acquired a simple operant response at the same rate as the control subjects. The protein-deficient monkeys, however, failed to generalize this operant response as rapidly as the control monkeys. As predicted, protein-deficient monkeys performed an operant response allowing access to a social partner less frequently than did the control monkeys. For like-diet pairings both the control and deficient subjects released approximately 60% of their partners; however, controls were far more likely to release a dissimilar diet partner (84% probability) than were deficient subjects (39% probability). Diet condition of the releasor was a significant factor, whereas diet condition of the release was not. Variables which could confound these findings were examined. It was found that: 1. The difference between diet groups was not accounted for by proximity of home cages; 2. Sex of the animals was not a confounding factor; 3. Although body weight and diet condition were highly correlated, body weight alone did not exert an effect above and beyond that of diet condition; 4. Dominance status, although correlated with both diet condition and body weight, showed only a weak correlation with the likelihood of one subject releasing another when the effect of diet condition was partialed out. In summary, diet condition played the major determining role in the frequency of release rates. Social behavior data was collected throughout the final phase of the experiment. Both groups of animals exhibited minimal play and sexual behaviors. Protein-deficient monkeys were more submissive than their matched controls. Subjects deviated most dramatically from one another in two behavioral clusters: disturbed (defined as self-stimulatory, autistic-like behaviors) and exploratory behaviors. Deficient monkeys engaged in more disturbed behaviors, while control monkeys engaged in more exploratory behaviors. Results are discussed in terms of behavioral similarity to social isolate animals, and possible nutritional-environmental interaction leading to chronic or persistent deficits in social development.

Magister Scientiae (Biodiversity and Conservation Biology) / Desertification deteriorates the landscape functionality of rangelands, affecting the resilience of biome boundaries which have a cascade effect not only in vegetation composition and characteristics but also in animal communities. The balance between plants and small mammals are essential in maintaining the functionality (i.e. nutrient cycling, soil stability, and water infiltration) of rangeland which includes the arid Steinkopf communal rangeland. However, the landscape functionality of rangelands has not been extensively studied, especially in communal areas where desertification is a serious concern since many people farm with livestock to survive. The aim of this study was to assess landscape functionality along an arid biome boundary and its relationship to small mammal assemblages. This study further aimed to provide a local landscape perspective of the current desertification process and projected expansion of the Desert Biome into more mesic biomes in South Africa. Three replicates for three Desert Biome, ecotone and Succulent Karoo Biome sites (27 sites in total) were selected to assess landscape functionality and survey small mammal assemblages over a one-year period. The change in landscape functionality between the three areas was assessed to determine the resilience of the biome boundary to desertification. For each small mammal survey, 216 live Sherman traps were set up during the summer and winter seasons to account for breeding, mortality, and possible migration. With the data, the body condition index, population density, and diversity were quantified.

This thesis began as an investigation into evolution of the platypus family (Ornithorhynchidae, Monotremata), now known from both Australia and South America. The thesis broadened its scope with inclusion of non-ornithorhynchid Mesozoic monotremes from Lightning Ridge, NSW. This change in direction brought an unexpected result: a fossil mammal from Lightning Ridge investigated for this thesis (presumed to be monotreme: Flannery et al., 1995) appears to be a new and unique type of mammal. Specimens were procured through Queensland Museum (Riversleigh material); Australian Museum (Lightning Ridge material); and Museum of Victoria and the South Australian Museum (fossil ornithorhynchids). Specimens were examined under a light microscope and scanning electron microscope; specimens were photographed using light photography and a scanning electron microscope; and illustrations and reconstructions were done with a camera lucida microscope attachment and photographic references. Parsimony analysis utilised the computer programs PAUP and MacClade. Major conclusions: 1) analysis and reconstruction of the skull of the Miocene platypus Obdurodon dicksoni suggest this robust, large-billed platypus was a derived northern offshoot off the main line of ornithorhynchid evolution; 2) the well-preserved skull of Obdurodon dicksoni shows aspects of soft anatomy previously unknown for fossil ornithorhynchids; 3) two upper molars from Mammalon Hill (Etadunna Formation, late Oligocene, central Australia) represent a third species of Obdurodon; 4) the South American ornithorhynchid Monotrematum sudamericanum from the Paleocene of Argentina is very close in form to the Oligocene-Miocene Obdurodon species from Australia and should be considered congeneric; 5) a revised diagnosis of the lower jaw of the Early Cretaceous monotreme Steropodon galmani includes the presence of two previously undescribed archaic features: the probable presence of postdentary bones and a meckelian groove; 6) morphological evidence is presented supporting a separate family Steropodontidae; and 7) analysis of new fossil material for Kollikodon ritchiei suggests that this taxon is not a monotreme mammal as originally identified but is a basal mammal with close relationships to allotherian mammals (Morganucodonta; Haramiyida). Kollikodon is provisionally placed as basal allotherian mammal (Allotheria sensu Butler 2000) and is unique at the ordinal level, being neither haramiyid nor multituberculate. A new allotherian order ??? Kollikodonta ??? is proposed.

The widespread expression of polypeptide growth factors from the earliest stages of embryonic development through to mature issues in the adult organism suggests an involvement in a reiterated developmental process affecting the underlying cellular growth and differentiation of many tissues. The hair follicle has taken on increased significance with the observation that many genetic mutations in these peptide growth factor genes affect its development. The targeted disruption of genes encoding members of the EpidermalGrowth Factor (EGF) and Fibroblast Growth Factor (FGF) families in the mouse has revealed a functional role for these proteins in the regulation of hair follicle growth. Experimental data and other factors are examined and results given. A second experimental system was used to determine if a functional relationship between certain peptide growth factors was conserved in the Merino sheep. The induction of a catagen-like state in the wool follicle and other epidermal changes associated with EGF treatment may be related to the transciptional induction of these peptide growth factors / Doctor of Philosophy (PhD)

This project aimed to reanalyse two large historical data sets from two different locations in New Zealand (Fiordland in the South Island and Pureora Forest Park in the North Island). The data describe populations of mice (Mus musculus), rats (Rattus rattus and R. norvegicus), and stoats (Mustela erminea) collected using standard monitoring techniques from five distinct types of forest habitat. The new analysis methods selected were an index of patchiness and Site Occupancy analysis. The objectives of the analysis were (1) to evaluate whether the patchiness index and Site Occupancy analysis methods might contribute to improved protocols for monitoring small mammal populations in the future, and (2) to use formal tests of five hypotheses to evaluate two of the assumptions made by the conventional density index often used in small mammal studies. I describe the results of the analyses for each species, including any problems encountered (such as the inability of the Site Occupancy method to analyse very sparse data sets). I also describe the results pooled from each of the two study locations and potential consequences for small mammal monitoring and control. This analysis has suggested that in most cases the density index is not a rigorous measure of small mammal populations. However, both the index of patchiness and Site Occupancy analysis provided useful, new information about these populations of rodents and stoats, despite the fact that these historical data sets were not designed for use with modern methods of analysis. Please note: some figures and tables were printed separately and added to the thesis as unnumbered pages. These can be found in the file 03Plates_and_Tables.pdf.