A teca (Tectona grandis) é uma das principais espécies madeireiras do mundo, com alto valor econômico, muito famosa por sua beleza, resistência e durabilidade. A espécie ocorre naturalmente na Índia, Mianmar, Tailândia, Laos e Indonésia, onde estudos de diversidade têm sido realizados no que tange à conservação de recursos genéticos. Entretanto, existe a necessidade de estudos de diversidade genética de teca no Brasil que poderiam ser utilizados, principalmente, para a proteção de cultivares e para o melhoramento genético. Visto isso, o objetivo deste trabalho foi caracterizar a diversidade genética de genótipos de teca utilizados nos plantios brasileiros. Para tanto foram testados 10 primers de microssatélites, obtidos na literatura, para a avaliação de 60 genótipos, 33 provenientes de sementes de plantios de teca em Cáceres, 14 correspondentes a clones obtidos em Cáceres e 13 genótipos referentes a clones de procedências fora do Brasil, sendo estas, Honduras, Malásia, Índia, Indonésia, Costa do Marfim e Ilhas Salomão. Os genótipos foram divididos em oito grupos, de acordo com sua procedência, para a análise da diversidade genética. Foram realizadas análises multivariadas pelo método bayesiano no programa STRUCTURE, análises de agrupamento e coordenadas principais. Dos 10 primers testados, nove se mostraram polimórficos, sendo então utilizados para as análises estatísticas. Elevada variabilidade genética para os genótipos de teca foi detectada, sendo o número médio de alelos por loco igual a 5,22. Os genótipos de Cáceres apresentaram 100% de polimorfismo, seguido pelos clones da Índia com 90% de polimorfismo. A heterozigosidade média observada ( o= 0,352) foi menor que a heterozigosidade média esperada ( e =0,443). Coerentemente com outros estudos em teca, a maior parte da variabilidade genética concentrou-se dentro dos grupos (Hs = 0,436). Com as análises do programa STRUCTURE foi possível definir a divisão dos genótipos em três grupos, sendo 73,4% dispostos em um único grupo (vermelho) representado pela maioria dos genótipos de Cáceres, 13,3% alocados no grupo verde compostos por alguns clones da Índia, Ilhas Salomão, um clone da Malásia, um de Honduras e os clones da Costa do Marfim e os 13,3% dos genótipos restantes possuíram uma mistura dos dois grupos (vermelho e verde). A análise de agrupamento, utilizando índice de Jaccard, indicou a separação dos genótipos em seis grupos distintos: grupo I pertencente ao clone da Indonésia, grupo II possuindo dois clones da Índia, grupo III com os genótipos de Cáceres e dois clones de fora (um da Índia e outro da Malásia), grupo IV possuindo os genótipos de Honduras e Malásia, grupo V com clones da Índia e grupo VI pertencente aos clones da Costa do Marfim e das Ilhas Salomão, sendo coerente com a análise de coordenadas principais. Através do agrupamento utilizando distância de Nei, foi possível inferir duas possíveis origens da teca implantada no Brasil: Malásia e Índia. Após a avaliação das divergências genéticas, sugestões são feitas no que tange a utilização de genótipos contrastantes para o uso como parentais em programas de melhoramento genético. / Teak (Tectona grandis) is one of the main timber species in the world with high economic value, famous for its beauty, strength and durability. The species occurs naturally in India, Myanmar, Thailand, Laos and Indonesia, where diversity studies have been conducted with regard to the conservation of genetic resources. However, there is a need for studies of genetic diversity of teak in Brazil that could be used mainly for the protection of plant varieties and for breeding. Therefore, the objective of this study was to characterize the genetic diversity of teak genotypes used in Brazilian plantations. We tested 10 microsatellite primers, obtained in the literature, to assess 60 teak genotypes, 33 genotypes from seeds of plantations in Caceres, 14 clones obtained in Caceres and 13 clones originated from Honduras, Malaysia, India, Indonesia, Ivory Coast and Solomon Islands. The genotypes were divided in eight groups, in accordance to its origin, for the genetic diversity analysis. Multivariate analysis were conducted using the Bayesian method implemented in the program STRUCTURE, as well as cluster and principal coordinates analysis. Of the 10 primers tested, 9 showed polymorphism, and were then used for statistical analysis. High genetic variability for the teak genotypes was detected, with the average number of alleles per locus equal to 5.22. Caceres genotypes showed 100% polymorphism, followed by the clones from India with 90% polymorphism. The average observed heterozygosity ( o = 0.352) was lower than the average expected heterozygosity ( e = 0.443). Consistent with other studies in teak, most of the genetic variability was concentrated within groups (Hs = 0.436). With the analysis of the STRUCTURE software it was possible to define the division of the genotypes into three groups, 73.4% placed in one group (red) represented the majority of the genotypes of Caceres, and 13.3% allocated in the green group composed of some clones from India, a clone from Solomon Islands, Malaysia and Honduras and the clones of the Ivory Coast. The 13.3% of the remaining genotypes possessed a mixture of the two groups (red and green). Cluster analysis using Jaccard index indicated the separation of the genotypes into six distinct groups: group I belonging to the clone from Indonesia, group II having two clones from India, group III with genotypes from Caceres and two clones from India and Malaysia, group IV having the Honduras and Malaysia genotypes, group V with clones from India and group VI with clones belonging to the Ivory Coast and the Solomon Islands. This result was consistent with the principal coordinate analysis. From the results described above, together with the cluster analysis using Neis distance, it was possible to infer two probable origins of teak implemented in Brazil: India and Malaysia. After assessing the genetic divergences, suggestions were made concerning the use of contrasting genotypes as parents in breeding programs.

O gene DMRT3 tem sido descrito como o principal gene a atuar na determinação da marcha em diversas raças equinas. O alelo A do SNP 23:g.22999655C>A do DMRT3 foi apontado como responsável por essa característica. Na raça brasileira Mangalarga Marchador, a qual apresenta dois padrões de marcha com características bem definidas, os genótipos AA e CA vem sendo associados à marcha picada e o genótipo CC à marcha batida. O objetivo geral do presente prospectar regiões genômicas associadas às marchas batida e picada em equinos Mangalarga Marchador. Foram utilizados 1.230 dados fenotípicos sobre o tipo de andamento (marcha batida N = 1.006; marcha picada N = 227) e, considerando a totalidade da genealogia conhecida para cada animal, 3172 animais no pedigree. Primeiramente foram testadas estratégias de modelagem para esta característica a fim de determinar os efeitos a serem considerados no modelo, bem como a melhor forma de inclusão (efeito fixo ou aleatório). Posteriormente, foi estudada a relação entre as frequências alélicas e genotípicas do gene DMRT3 com os padrões de parentesco e endogamia de acordo com cada tipo de marcha. Um estudo de associação genômica ampla em passo único (considerando informações de animais genotipados e não-genotipados simultaneamente) foi conduzido para verificar regiões genômicas, polimorfismos de nucleotídeo único e genes relacionados com a determinação do tipo de marcha em cavalos Mangalarga Marchador. Vinte e dois polimorfismos de nucleotídeo único localizados nos cromossomos 4(N = 5), 6 (2), 16 (1), 23 (11), 26 (1) e 29 (2), foram responsáveis por 42,43% da variância genética aditiva. Foram associados ao tipo de marcha 69 genes, mas cerca de 39 não estavam anotados em equinos. Foi conduzido um blast a fim de recuperar a função mais provável destes genes. Foram encontradas oito vias metabólicas associadas ao tipo de marcha. Os principais genes envolvidos estavam relacionados à percepção de estímulos externos, metabolismo energético-oxidativo, sistema imune e aprendizado e ritmo da locomoção. Não foi possível identificar a(s) variante(s) causal(ais) do tipo de marcha, contudo este estudo foi o primeiro e verificar que a possível determinação genética do tipo de marcha em cavalos Mangalarga Marchador passa por diferenças em níveis metabólicos que garantem a adaptação dos animais ao tipo de andamento. / The DMRT3 gene has been described as the main gene to act in gait determination in several equine breeds. The allele A of the SNP 23:g.22999655C>A of DMRT3 gene was identified as responsible for this trait. In the Brazilian Mangalarga Marchador breed, which presents two gait patterns with characteristics well defined, the AA and CA genotypes have been associated with marcha picada gait and CC genotype with marcha batida gait. The general aim of this study was to prospect genomic regions associated with marcha batida and marcha picada gaits in Mangalarga Marchador equines. 1,230 phenotypic data were used on the type of gait (marcha batida N = 1.006; marcha picada N = 227) and, considering the totality of known genealogy for each animal, 3,172 animals in the pedigree. Firstly, modelling strategies were tested for this trait in order to determine the effects to be considered in the model, as well as the best form of inclusion (fixed or random effect). Based on the best modelling strategy to be used, the relationship between the allelic and genotypic frequencies of the DMRT3 gene with kinship and inbreeding patterns was studied according to each type of gait. A single-step wide genomic association study (considering information from both genotyped and non-genotyped animals simultaneously) was conducted to verify genomic regions, single nucleotide polymorphisms and genes related to determination of gait type in Mangalarga Marchador horses. Twenty two single nucleotide polymorphisms located on chromosomes 4 (N = 5), 6 (2), 16 (1), 23 (11), 26 (1) and 29 (2) were responsible for 42.43% of the additive genetic variance. 69 genes were associated with gait type, but about 39 were not annotated in horses. A blast was conducted in order to recover the most likely function of these genes. Eight metabolic pathways were found associated with gait type and the main genes involved were related to the perception of external stimuli, energy-oxidative metabolism, immune system and learning and rhythm of locomotion. It was not possible to identify the causal variant(s) of the type of gait; however, this study was the first and to verify that the possible genetic determination of gait type in Mangalarga Marchador horses goes through differences in the metabolic levels that guarantee the adaptation of the animals to the type of gait.

Amostras populacionais de Oc. scapularis foram coletadas nos municípios de Tremembé-SP (TRE), São Paulo-SP (SPA) e Itaboraí-RJ (ITA). Foram empregados como marcadores biológicos: o gene mitocondrial Citocromo Oxidase Subunidade-1 (COI), geometria alar e análise da genitália masculina. Tais marcadores são tradicionalmente reconhecidos pelo poder discriminante em estudos desta natureza. As populações ITA, TRE e SPA mostraram-se distintas quanto à forma alar, sugerindo baixo fluxo gênico entre elas. Foi verificado dimorfismo sexual em relação ao tamanho isométrico, à forma alar e ao grau de diferenciação populacional. A população de ITA apresentou menor tamanho dos centróides que as demais populações estudadas. Foi verificado amplo polimorfismo genético, tendo sido detectados 51 haplótipos de COI e apenas 11 compartilhados entre as populações ITA, TRE e SPA. Quando comparados com as populações de estudo anteriormente realizado por Devicari, 2010 encontramos um padrão parecido. Analisando conjuntamente os presentes dados com aqueles obtidos por Devicari, 2010, computamos 52 haplótipos sendo apenas alguns compartilhados. Os valores do índice de diferenciação genética <font face=\"Symbol\">Fst observados foram moderados somente entre SPA e ITA, nas demais populações estudadas a diferenciação foi baixa, a diferenciação observada foi compatível com a hipótese de distanciamento geográfico das populações coletadas. Analisando cada marcador biológico, concluímos que as populações estudadas não se tratam de complexo de espécies. Ainda não descartamos a existência de complexo dentro de Oc. scapularis, porém, para definitiva resposta a essa questão serão necessários mais estudos envolvendo populações de outras regiões. / Samples of Oc. scapularis were collected in the municipalities of Tremembé-SP (TRE), São Paulo-SP (SPA) and Itaboraí-RJ (ITA). We used the following biological markers: mitochondrial cytochrome oxidase subunit-1 gene (COI), wing geometry and shape of male genitalia. These markers are traditionally known by its discriminating power in studies of this nature. ITA, SPA and TRE populations, showed distinct wing shape, suggesting low gene flow. We observed sexual dimorphism concerning the isometric size, wing shape and the degree of populational differentiation. ITA sample exhibited the lowest centroid sizes. We found high genetic polymorphism in all populational samples, being 51 COI haplotypes. Out of them, only 11 haplotypes were noted to be shared between at least two or three populations. When comparing our results with those of a previous survey conducted by Devicari (2010), we found a quite similar pattern of hign polymorfism. In total, both studies comprised 52 haplotypes. The <font face=\"Symbol\">Fst index of genetic differentiation values were considered \"moderate\" between ITA and SPA and \"low\" in the other comparisons. Present results are consistent with the hypothesis that populations are subjected to isolation by geographical distance. Analyzing together each biologcal marker, we conclude that populations studied do not consist a species complex. We do not rule out the possible occurence of a complex in Oc. scapularis, however, a definitive answer to this question will require further studies and sampling of populations from elsewhere.

Embora já existam estudos anteriores a respeito da estrutura de população em cana-deaçúcar, até o momento nenhum estudo foi feito usando marcadores SNPs gerados a partir de plataformas de genotipagem de larga escala, como por exemplo, Sequenom iPLEX MassARRAY. No presente trabalho, foi investigada a estrutura populacional no painel brasileiro de variedades de cana-de-açúcar. Esse painel é formado por materiais elites, ancestrais importantes e cultivares utilizados em programa de melhoramento. Um total de 1033 marcadores SNPs foram utilizados para genotipar os acessos do painel. A classificação dos dados feita usando o software SuperMASSA. A estrutura de população foi analisada por meio de análise de componentes principais (ACP), análise de agrupamentos e usando o software STRUCTURE. Devido ao fato que no software STRUCTURE não é possível dados de marcadores moleculares provenientes de espécies poliploides com aneuplodia frequente, o conjunto de dados foi separado e analisado de acordo com nível de ploidias dos SNPs. Com a finalidade de comparar os resultados, foi feita uma análise de coordenadas principais na matriz de distância, com os elementos definidos por 1 - coeficiente de parentesco. A análise de componentes principais revelou presença de estrutura de população. O primeiro componente separou o acesso IN84-58 (S. spontaneum) dos outros acessos que por sua vez estão separados em três grupos: o primeiro grupo formado pelos acessos que são S. sinense, o segundo grupo formado pelos cultivares modernos de cana-de-açúcar e o terceiro grupo formado por acessos que são espécies S. officinarum. Resultados da análise de agrupamento usando distância de alelos compartilhados são condizentes com resultados da ACP. Por outro lado, análise de coordenadas principais e método de agrupamento UPGMA usando o coeficiente de parentesco mostraram uma maior dissimilaridade genética entre os acessos separando as progênies do cultivar RB72454 do grupo formado pelos genitores e ou progenies do cultivar NA56-79. A diferença entre os resultados da análise de componentes principais e de coordenadas principais é devido principalmente a pressuposições nas estimativas do coeficiente de parentesco que são irrealísticas. Com relação a análise feita com o software STRUCTURE, o número de subpopulações e a matriz Q estimada variou de acordo com nível de ploidia dos marcadores. De um modo geral, as análises de estrutura de população mostrou que há evidências de estreitamento da base genética dos acessos devido a cruzamentos recorrentes de indivíduos aparentados. Espera-se que estas informações sejam importantes para o mapeamento associativo e melhoramento genético da espécie. / Although there are several studies inferring population structure in sugarcane, none of them have yet used SNP markers generated from high-throughput platforms, such as, Sequenom iPLEX MassARRAY platform. In this study, it was investigated the population structure in a Brazilian panel of sugarcane varieties. This panel is comprised by elite breeding materials, important ancestors, and cultivars mostly used by breeding programs. 1,033 SNP markers were scored. SNP genotype calling was made using software SuperMASSA. The population structure was analyzed via principal components analysis (PCA), cluster analysis and using the software STRUCTURE. Due to the fact that STRUCTURE is not possible to analyze molecular markers data scored on species withmixed ploidy level, the dataset was separated and analyzed according to SNPs level ploidy estimates. With purpose of comparing the results, it was made a principal coordinate analysis (PCO) of distance matrix, with elements defined by 1 - kinship coefficient. The principal components analysis revealed some structure. The first component separated out IN84-58 (Saccharum spontaneum) from the others acessions, whereby these others acessions were allocated in three groups, comprised by S. sinense species, sugarcane modern cultivars and S. officinarum species. Results from cluster analysis using allele shared distance are in agreement with PCA results. On the other hand, principal coordinates analysis and UPGMA hierarchical clustering method based on kinship coefficient showed a broader genetic dissimilarity between acessions, allocating RB72454 progenies apart from its parents and/or progenies of NA56-79. The difference of results between PCA and PCO are mainly due to irrealistics assumptions in the calculation of kinship. Regarding analysis from the STRUCTURE, the number of subpopulations and Q matrix estimated varied with level ploidy. In general, the study of population structure showed some evidence of narrow genetic distances between accessions, due to recurrent crosses between related individuals. The information presented hereby could be important for association mapping and sugarcane breeding program.

[Truncated abstract] Narrow-leaf lupin (Lupinus angustifolius L.) was first recorded as having been introduced into Germany during the mid-19th century for use as green manuring and as fodder crops. However, it was not until post World-War I that there was any serious attempt to domesticate the species. Since that time several key domestication genes have been incorporated to enable the species to be grown as a crop over a range of climates, harvested as a bulk commodity and, the seed used for both animal and human consumption. However, the recent domestication of this species has seen a rather limited use of wild germplasm largely as a result of the difficulty in retaining these key domestication genes. To make the task of retaining these genes manageable, it was decided to resort to molecular technology. A mapping population of F8 derived recombinant inbred lines (RILs) has previously been established by the Department of Agriculture and Food, Western Australia, from a cross between a domesticated breeding line 83A:476 and a wild type P27255 in narrow-leaf lupin. The parents together with 89 RILs (of a population of 115) were subjected to DNA fingerprinting using microsatelliteanchored fragment length polymorphism (MFLP) to rapidly generate DNA markers for construction of a linkage map. Five hundred and twenty two unique markers of which 21% were co-dominant, were generated and mapped. Phenotypic data for the domestication traits: mollis (soft seeds), leucospermus (white flower and seed colour); Lentus (reduced pod-shattering), iucundis (low alkaloid), Ku (early flowering) and moustache pattern on seed coats; were included. Three to 7 molecular markers were identified within 5 cM of each of these domestication genes. The anthracnose resistance gene Lanr1 was also mapped. Linkage groups were constructed using MapManager version QTXb20, resulting in 21 linkage groups consisting of 8 or more markers. ... Five pairs of QTLs were found to be involved in epistasis, 2 of these having an effect on early vigour and another 3 influencing the time to opening of the first florets. Variation explained for each trait ranged from 28% for seed size, to 88% for days to flowering. We showed that it was possible to use this data to predict genotypes of superior progeny for these traits under Mediterranean conditions. QTL regions were compared on a second published linkage map and regions of conserved synteny with the model legume Medicago truncatula high-lighted. The work presented in this thesis demonstrates the importance of tight linkage between markers and genes of interest. It is especially important when dealing with genetically diverse material as found in the wild. One of the main problems faced by molecular scientists is the phenomenon known as linkage disequilibrium in marker populations caused by either small population size or 4 insufficient opportunity for recombination. This frequently results in the development of markers with little or no application outside of the population in which it was developed. Although the relatively small size of the population used in this study exposes it to such constraints, in this case excellent and valuable results were achieved in developing useful markers to at least 3 of the domestication traits within a relatively short time period of less then 4 years.

Agronomic and physiological traits associated with drought adaptation were assessed within the Seri/Babax recombinant inbred line population, derived from parents similar in height and maturity but divergent in their sensitivity to drought. Field trials under different water regimes were conducted over three years in Mexico and under rainfed conditions in Australia. Under drought, canopy temperature (CT) was the single-most drought-adaptive trait contributing to a higher performance (R2= 0.71, p<0.0001), highly heritable (h2= 0.65, p<0.0001) and consistently associated with yield phenotypically (r= -0.75, p<0.0001) and genetically [R(g)= -0.95, p<0.0001]. CT epitomises a mechanism of dehydration avoidance expressed throughout the growing season and across latitudes, which can be utilised as a selection criteria to identify high-yielding wheat genotypes or as an important predictor of yield performance under drought. Early response under drought, suggested by a high association of CT with estimates of biomass at booting (r= -0.44, p<0.0001), leaf chlorophyll (r= -0.22,p<0.0001) and plant height (r= -0.64, p<0.0001), contrast with the small relationships with anthesis and maturity (averaged, r= -0.10, p<0.0001), and with osmotic potential (r= -0.20, p<0.0001). Results suggest that the ability to extract water from the soil under increasing soil water deficit is a major attribute of drought adaptation. Ample genetic variation and significant transgressive segregation under drought suggested a polygenic governance feasible of dissection via molecular markers of CT and associated physiological and agronomic traits. Bulked segregant analysis of selected secondary traits was utilised as an alternative to complete genome mapping, due to a low polymorphism (27%) within the cross and limited chromosomic linkage of loci. The assessment of the extremes of expression in a genotypic subset with a composite molecular database of 127 markers (PCR-based and AFLPs) allowed evaluation of the three hexaploid wheat genomes and coverage of all chromosomic groups, except 3D. One-way analysis of variance indicated significant associations of loci explaining phenotypic variance under drought and rainfed conditions, of 20-70% in Mexico and 20-45% in Australia (F>5.00, p<0.05). Significant loci were established in both latitudes for all physiological and agronomic traits assessed via BSA, with CT being the trait with the most numerous associations (in Mexico, 34 loci; in Australia, 24). Results demonstrate an efficient development of molecular markers associated to physiological traits under specific soil water conditions in Mexico and Australia, and suggest further genomic and transcriptomic studies be conducted for unravelling the complex relationship between drought adaptation and performance under drought. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1284279 / Thesis (Ph.D.) -- University of Adelaide, School of Agriculture, Food and Wine, 2007

Les tubercules andins oca (Oxalis tuberosa Molina), ulluco (Ullucus tuberosus Caldas) et mashua (Tropaeolum tuberosum Ruiz & Pav.) sont des espèces tubéreuses cultivées originaires des régions hautes des Andes, où elles revêtent une importance particulière aux niveaux alimentaire, agronomique, culturel et économique. La diversité génétique au sein de ces espèces est très grande, mais est menacée d'érosion génétique. Dans ce contexte, notre étude s'est basée sur des échantillons de oca, ulluco et mashua, maintenus dans des système de conservation in situ et ex situ au Pérou et en Bolivie, dans l'objectif de contribuer à la conservation efficace (in situ et ex situ) de ces espèces négligées. Cette étude a combiné les connaissances autochtones andines, ainsi que des données agronomiques, morphologiques et moléculaires. Nous avons montré que l'agriculture andine conserve une grande diversité au niveau inter-spécifique, mais aussi au niveau intra-spécifique, en terme de nombre de variétés locales. Nous avons également mis en évidence de la présence de variétés hétérogènes, la congruence entre les données moléculaires et morphologiques, et une structure génétique influencée par la provenance géographique. Enfin, nous avons compilé l'ensemble de nos résultats dans un modèle récapitulatif. Nous avons montré l'importance des caractéristiques intrinsèques de l'espèce (mode de reproduction), ainsi que les spécificités du système agricole andin (socioculturels, économiques et environnementales). Cette étude a contribué de manière significative à la compréhension de la diversité génétique et de la structure des tubercules andins.

Mr Clarke used molecular genetic techniques to determine the origins of the sweet potato (kumara) and the bottle gourd in Oceania. Both are thought to have been introduced into Oceania by Polynesian voyagers who collected them from South America about 1,000 years ago (the word kumara is used by indigenous South Americans). Using new and improved DNA techniques, Mr Clarke analysed several hundred varieties of sweet potato and about 40 varieties of bottle gourd. The results are consistent with a pre-European introduction of these crops into Polynesia from South America, and have revealed something about their dispersal patterns within Polynesia, including New Zealand.

Mr Clarke used molecular genetic techniques to determine the origins of the sweet potato (kumara) and the bottle gourd in Oceania. Both are thought to have been introduced into Oceania by Polynesian voyagers who collected them from South America about 1,000 years ago (the word kumara is used by indigenous South Americans). Using new and improved DNA techniques, Mr Clarke analysed several hundred varieties of sweet potato and about 40 varieties of bottle gourd. The results are consistent with a pre-European introduction of these crops into Polynesia from South America, and have revealed something about their dispersal patterns within Polynesia, including New Zealand.

Mr Clarke used molecular genetic techniques to determine the origins of the sweet potato (kumara) and the bottle gourd in Oceania. Both are thought to have been introduced into Oceania by Polynesian voyagers who collected them from South America about 1,000 years ago (the word kumara is used by indigenous South Americans). Using new and improved DNA techniques, Mr Clarke analysed several hundred varieties of sweet potato and about 40 varieties of bottle gourd. The results are consistent with a pre-European introduction of these crops into Polynesia from South America, and have revealed something about their dispersal patterns within Polynesia, including New Zealand.