This work follows the direction set by Sewall Wright in applying path analysis, and other multivariate statistical techniques, to the study of anthropometric variation. Data on anthropometric variation of six Solomon Island populations is analyzed using statistical models which can distinguish between within-group and between-group genetic variation. The correlation structure of 27 anthropometric measurements is examined by cluster analysis and principal components analysis. The six populations show a common pattern (in both males and females) which echoes earlier studies. The correlation matrix of measurements is then partitioned into genetic and environmental components and the genetic correlation matrix is examined, once again by cluster analysis and principal components analysis. There is a fairly close agreement between the genetic correlation structure and the phenotypic correlation structure. The environmental correlation matrix is not examined further because it is very poorly estimated. The partitioning of phenotypic correlations into genetic and environmental components is based on a multivariate generalization of a path model for the heritability of a continuous trait proposed by C. C. Li. The parameters estimated in the single trait model include additive genetic heritability, common home environment, and genetic correlation between spouses. In order to fit this model observations are required on parent-offspring, spouse-spouse, and sib-sib correlations. Heritability values for the Solomons are markedly lower than those reported elsewhere. However, when total heritability (ignoring subpopulation structure) is estimated for the six Solomons populations, the values are higher and form a more familiar pattern. The striking differences between the two kinds of heritability in the Solomon Islands emphasizes the danger of using total heritability estimates obtained from several subpopulations or a national sample. The between-group component of heritability for each measurement is compared to the within-group heritability and to levels of between-group phenotypic variation. The results demonstrate that high within-group heritability for a given trait does not imply that between-group variation in that trait is genetic in origin.

The Pacific is a world of islands, so human migrations are necessarily part of the story of this world. It is somewhat surprising therefore to find that there are so many points of contention among archaeologists, linguists and biologists about the role that migrations have played in producing culture change in Pacific prehistory -just how many migrations are we talking about, on what scale, by whom, and when, and which discipline provides more reliable evidence? The specifics of the debate may be local, but the substantive issues have broader applicability. This thesis attempts to contribute to the debate by considering the archaeological evidence for a hypothesised migration during the post-Lapita period of people speaking non-Austronesian languages into the Main Reef and Santa Cruz Islands in the southeast Solomon Islands. It describes the theoretical and methodological difficulties in trying to assess whether cladistic or rhizotic processes best account for culture change. Previously unpublished archaeological material from two Main Reef Islands sites is presented, and is compared with the linguistic and human biological evidence pertinent to the case study. The narratives of cross-cultural encounters during the historic period are examined to investigate how such interactions produce change in cultural traditions, and to assess the archaeological visibility of these contacts. While the archaeological record testifies to cultural borrowings and lendings in the Reef-Santa Cruz case, the argument for either large-scale population intrusion or replacement is not well supported. The archaeological record appears to be in conflict with the prevailing interpretation of history reconstructed from historical linguistics and human genetics.

This work follows the direction set by Sewall Wright in applying path analysis, and other multivariate statistical techniques, to the study of anthropometric variation. Data on anthropometric variation of six Solomon Island populations is analyzed using statistical models which can distinguish between within-group and between-group genetic variation. The correlation structure of 27 anthropometric measurements is examined by cluster analysis and principal components analysis. The six populations show a common pattern (in both males and females) which echoes earlier studies. The correlation matrix of measurements is then partitioned into genetic and environmental components and the genetic correlation matrix is examined, once again by cluster analysis and principal components analysis. There is a fairly close agreement between the genetic correlation structure and the phenotypic correlation structure. The environmental correlation matrix is not examined further because it is very poorly estimated. The partitioning of phenotypic correlations into genetic and environmental components is based on a multivariate generalization of a path model for the heritability of a continuous trait proposed by C. C. Li. The parameters estimated in the single trait model include additive genetic heritability, common home environment, and genetic correlation between spouses. In order to fit this model observations are required on parent-offspring, spouse-spouse, and sib-sib correlations. Heritability values for the Solomons are markedly lower than those reported elsewhere. However, when total heritability (ignoring subpopulation structure) is estimated for the six Solomons populations, the values are higher and form a more familiar pattern. The striking differences between the two kinds of heritability in the Solomon Islands emphasizes the danger of using total heritability estimates obtained from several subpopulations or a national sample. The between-group component of heritability for each measurement is compared to the within-group heritability and to levels of between-group phenotypic variation. The results demonstrate that high within-group heritability for a given trait does not imply that between-group variation in that trait is genetic in origin.

The Pacific is a world of islands, so human migrations are necessarily part of the story of this world. It is somewhat surprising therefore to find that there are so many points of contention among archaeologists, linguists and biologists about the role that migrations have played in producing culture change in Pacific prehistory -just how many migrations are we talking about, on what scale, by whom, and when, and which discipline provides more reliable evidence? The specifics of the debate may be local, but the substantive issues have broader applicability. This thesis attempts to contribute to the debate by considering the archaeological evidence for a hypothesised migration during the post-Lapita period of people speaking non-Austronesian languages into the Main Reef and Santa Cruz Islands in the southeast Solomon Islands. It describes the theoretical and methodological difficulties in trying to assess whether cladistic or rhizotic processes best account for culture change. Previously unpublished archaeological material from two Main Reef Islands sites is presented, and is compared with the linguistic and human biological evidence pertinent to the case study. The narratives of cross-cultural encounters during the historic period are examined to investigate how such interactions produce change in cultural traditions, and to assess the archaeological visibility of these contacts. While the archaeological record testifies to cultural borrowings and lendings in the Reef-Santa Cruz case, the argument for either large-scale population intrusion or replacement is not well supported. The archaeological record appears to be in conflict with the prevailing interpretation of history reconstructed from historical linguistics and human genetics.

This thesis examines unsewn cloth pieces which once belonged to Victoria Wood and places them into their social and historical context. It uses the biography of Victoria Wood and her fabrics to argue for the importance of fabric collecting and dressmaking for New Zealand women from 1935 to 1955. It questions why ubiquitous fabrics bought for dress making are not represented in historical accounts of women, or in more general accounts of historical clothing and dress. Aspects of material culture theories are employed to analyse the material properties of the fabric pieces. These are situated within a wider domestic context to demonstrate that there were intrinsic qualities of fabric that influenced and were imagined by many women in this period. Oral histories and other documentary research add to the wider account and provide evidence of the way that dressmaking fabrics reflected the shifting notions of domesticity. The thesis suggests that fabrics bought for the creation of clothing can represent the past or a person. It also demonstrates how dressmaking fabrics simultaneously embody personal and social narratives which reflect the emotional and cultural values of a particular period. In this thesis I construct narratives which are based on the social and historical findings to highlight the importance of fabric collecting and dressmaking as an everyday domestic practice.

This work follows the direction set by Sewall Wright in applying path analysis, and other multivariate statistical techniques, to the study of anthropometric variation. Data on anthropometric variation of six Solomon Island populations is analyzed using statistical models which can distinguish between within-group and between-group genetic variation. The correlation structure of 27 anthropometric measurements is examined by cluster analysis and principal components analysis. The six populations show a common pattern (in both males and females) which echoes earlier studies. The correlation matrix of measurements is then partitioned into genetic and environmental components and the genetic correlation matrix is examined, once again by cluster analysis and principal components analysis. There is a fairly close agreement between the genetic correlation structure and the phenotypic correlation structure. The environmental correlation matrix is not examined further because it is very poorly estimated. The partitioning of phenotypic correlations into genetic and environmental components is based on a multivariate generalization of a path model for the heritability of a continuous trait proposed by C. C. Li. The parameters estimated in the single trait model include additive genetic heritability, common home environment, and genetic correlation between spouses. In order to fit this model observations are required on parent-offspring, spouse-spouse, and sib-sib correlations. Heritability values for the Solomons are markedly lower than those reported elsewhere. However, when total heritability (ignoring subpopulation structure) is estimated for the six Solomons populations, the values are higher and form a more familiar pattern. The striking differences between the two kinds of heritability in the Solomon Islands emphasizes the danger of using total heritability estimates obtained from several subpopulations or a national sample. The between-group component of heritability for each measurement is compared to the within-group heritability and to levels of between-group phenotypic variation. The results demonstrate that high within-group heritability for a given trait does not imply that between-group variation in that trait is genetic in origin.

The relevance of the sustainable development approach for land conservation in tropical hill regions is often assumed, but is seldom evaluated against local realities. This study analyses the causes of land degradation and unsustainability in the seasonally fire-prone region of Navosa, Viti Levu, Fiji Islands. A complex and multi-layered set of connections exists between livelihood strategies and the environment in this region where fire is used as a tool. Traditional institutions governing land management have been undermined, however, and the sustainability of the vanua (or land-people nexus) is threatened. Despite their reliance on fire, local participants reported uncontrolled burning (caused by a minority of villagers) and deforestation leading to land degradation, lower productivity and damage to fisheries. Degradation was increasing parallel with escalating human and animal populations, despite relatively equitable, resilient and livelihood-enhancing cultural institutions. Non-indigenous models of resource use, imposed by colonial and neocolonial authorities, have exacerbated land degradation and compromised indigenous resource management. Traditional tenure and leadership are impeded, leading to inequities in access to land as populations increase, and leaving local leaders unable to enact conservation. An intrusive market-based economic system encourages increased resource exploitation with little regard for environmental sustainability. A traditional model of agroarboriculture and indigenous development in the form of still-observable (but largely disused) irrigated terraces suggests the relevance of a sustainable alternative based on indigenous knowledge. To achieve sustainable development in Navosa, the emphasis should be on socio-environmental amelioration rather than on economic development (which exacerbates degradation), and to this end, local participants expressed a desire for particular conservation strategies.

This work follows the direction set by Sewall Wright in applying path analysis, and other multivariate statistical techniques, to the study of anthropometric variation. Data on anthropometric variation of six Solomon Island populations is analyzed using statistical models which can distinguish between within-group and between-group genetic variation. The correlation structure of 27 anthropometric measurements is examined by cluster analysis and principal components analysis. The six populations show a common pattern (in both males and females) which echoes earlier studies. The correlation matrix of measurements is then partitioned into genetic and environmental components and the genetic correlation matrix is examined, once again by cluster analysis and principal components analysis. There is a fairly close agreement between the genetic correlation structure and the phenotypic correlation structure. The environmental correlation matrix is not examined further because it is very poorly estimated. The partitioning of phenotypic correlations into genetic and environmental components is based on a multivariate generalization of a path model for the heritability of a continuous trait proposed by C. C. Li. The parameters estimated in the single trait model include additive genetic heritability, common home environment, and genetic correlation between spouses. In order to fit this model observations are required on parent-offspring, spouse-spouse, and sib-sib correlations. Heritability values for the Solomons are markedly lower than those reported elsewhere. However, when total heritability (ignoring subpopulation structure) is estimated for the six Solomons populations, the values are higher and form a more familiar pattern. The striking differences between the two kinds of heritability in the Solomon Islands emphasizes the danger of using total heritability estimates obtained from several subpopulations or a national sample. The between-group component of heritability for each measurement is compared to the within-group heritability and to levels of between-group phenotypic variation. The results demonstrate that high within-group heritability for a given trait does not imply that between-group variation in that trait is genetic in origin.

This work follows the direction set by Sewall Wright in applying path analysis, and other multivariate statistical techniques, to the study of anthropometric variation. Data on anthropometric variation of six Solomon Island populations is analyzed using statistical models which can distinguish between within-group and between-group genetic variation. The correlation structure of 27 anthropometric measurements is examined by cluster analysis and principal components analysis. The six populations show a common pattern (in both males and females) which echoes earlier studies. The correlation matrix of measurements is then partitioned into genetic and environmental components and the genetic correlation matrix is examined, once again by cluster analysis and principal components analysis. There is a fairly close agreement between the genetic correlation structure and the phenotypic correlation structure. The environmental correlation matrix is not examined further because it is very poorly estimated. The partitioning of phenotypic correlations into genetic and environmental components is based on a multivariate generalization of a path model for the heritability of a continuous trait proposed by C. C. Li. The parameters estimated in the single trait model include additive genetic heritability, common home environment, and genetic correlation between spouses. In order to fit this model observations are required on parent-offspring, spouse-spouse, and sib-sib correlations. Heritability values for the Solomons are markedly lower than those reported elsewhere. However, when total heritability (ignoring subpopulation structure) is estimated for the six Solomons populations, the values are higher and form a more familiar pattern. The striking differences between the two kinds of heritability in the Solomon Islands emphasizes the danger of using total heritability estimates obtained from several subpopulations or a national sample. The between-group component of heritability for each measurement is compared to the within-group heritability and to levels of between-group phenotypic variation. The results demonstrate that high within-group heritability for a given trait does not imply that between-group variation in that trait is genetic in origin.

This thesis is an ethnography of Anglo-Indians in Calcutta. All ethnographies are accounts arising out of the experience of a particular kind of encounter between the people being written about and the person doing the writing. This thesis, amongst other things, reflects my changing views of how that experience should be recounted. I begin by outlining briefly who Anglo-Indians are, a topic which in itself alerts one to complexities of trying to get an ethnographic grip on a shifting subject. I then look at some crucial elements that are necessary for an “understanding” of Anglo-Indians in Calcutta: the work that has already been done in relation to Anglo-Indians, the urban context of the lives of my research participants and I discuss the methodological issues that I had to deal with in constructing this account. In the second part of my thesis I explore some crucial elements of the lives of Anglo-Indians in Calcutta: the place of Christianity in their lives, education not just as an aspect of socialisation but as part of their very being and, finally, the public rituals that now give them another way of giving expression to new forms of Anglo-Indian becoming. In all of my work I was driven by a desire to keep close to the experience of the people themselves and I have tried to write a “peopled” ethnography. This ambition is most fully realised in the final part of my thesis where I recount the lives of three key participants.