Comparação de métodos para estimar o tamanho ótimo de parcela em bananeira
Arquivos
Data
2010-02-12
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Viçosa
Resumo
O objetivo deste trabalho foi avaliar os métodos da Máxima Curvatura, Máxima Curvatura Modificado e Modelo Linear de Resposta Platô na determinação de tamanho ótimo de parcela, uma vez que os métodos tradicionais da Máxima Curvatura e Máxima Curvatura Modificado superestimam e subestimam, respectivamente, o tamanho da unidade básica. Por essa razão, foi proposta neste trabalho a comparação entre esses métodos e o método do Modelo Linear de Resposta Platô, considerando-se dados de genótipos de bananeira. Avaliaram-se os caracteres vegetativos altura da planta, perímetro do pseudocaule, número de folhas vivas no florescimento, número de filhos emitidos até o florescimento e número de folhas vivas na colheita. Nessa avaliação, cada planta foi julgada como uma unidade básica (ub) com área de 6 m2, perfazendo, assim, 360 unidades, cujas adjacentes foram combinadas de modo a formar 23 tamanhos de parcelas preestabelecidas, com formatos retangulares dispostos em fileiras. Os resultados indicaram que os tamanhos de parcela variaram com o método utilizado e a variável avaliada. No método da Máxima Curvatura, o tamanho da unidade básica variou entre 17,5 para Número de Filhos Emitidos no Florescimento (NFF) e 20,5 para Número de Folhas Vivas após a Colheita (NFVC), enquanto o coeficiente de variação variou entre 3,2 para Altura da Planta (AP) e 5,3 para Número de Folhas Vivas após a Colheita (NFVC). No método da Máxima Curvatura Modificado, o tamanho da unidade básica variou entre 0,74 para Altura da Planta (AP) e 6,88 para Número de Filhos Emitidos no Florescimento (NFF), enquanto o coeficiente de variação variou entre 5,2 para Altura da Planta (AP) e 13,84 para Número de Filhos Emitidos no Florescimento (NFF). Pelo método do Modelo Linear de Resposta Platô, o tamanho da unidade básica variou entre 10,50 para Número de Folhas Vivas (NFV) e 13,22 para Número de Filhos Emitidos no Florescimento (NFF), enquanto o coeficiente de determinação variou de 71,77% para Número de Filhos Emitidos no Florescimento (NFF) a 80,33% para Número de Folhas Vivas (NFV). Dessa forma, recomenda-se a utilização simultânea de mais de um método para determinação do tamanho ótimo da parcela, a fim de que o tamanho realmente adotado atenda, na medida do possível, aos diversos fatores considerados em cada método.
The objective of this study was to evaluate the methods of maximum curvature, modified maximum curvature and linear model Response Plateau, in determining the optimum plot size, since the traditional methods of maximum curvature and Modified maximum curvature respectively underestimate and overestimate the size of the basic unit. For this reason a comparison between these methods and the method of Linear Model Response Plateau and the data are genotypes was proposed in this paper. We evaluated the vegetative characteristics, plant height, pseudostem circumference, number of green leaves at flowering, number of offspring produced at flowering and the number of green leaves at harvest. In the evaluation, each plant was treated as a basic unit (ub) with an area of 6 m², thus making up 360 units, which were combined adjacently to form 23 pre-set plot sizes, in rectangular shaped molds. The results showed that the plot sizes vary with the method used and the variable evaluated. In the method of maximum curvature the basic unit size ranged from 17.5 for the number of offspring at flowering (NFF) to 20.5 for the number of green leaves after harvest (NFVC), while the coefficient of variation ranged from 3.2 for plant height (AP) to 5.3 for the number of green leaves after harvest (NFVC). For the method of modified maximum curvature, the size of the basic unit ranged from 0.74 for plant height (AP) to 6.88 for the number of offspring at flowering (NFF), while the coefficient of variation ranged from 5.2 for plant height (AP) to 13.84 for the number of offspring at flowering (NFF). In the linear model response plateau the size of the basic unit ranged from 10.50 for the number of green leaves (NFV) to 13.22 for the number of offspring at flowering (NFF), while the coefficient of determination ranged from 71.77% for the number of offspring at flowering (NFF) to 80.33% for the number of green leaves (NFV). Thus, it is recommended to simultaneously use more than one method to determine the optimum plot size so that the size adopted attends to the needs, as far as possible, of the various factors considered in each method.
The objective of this study was to evaluate the methods of maximum curvature, modified maximum curvature and linear model Response Plateau, in determining the optimum plot size, since the traditional methods of maximum curvature and Modified maximum curvature respectively underestimate and overestimate the size of the basic unit. For this reason a comparison between these methods and the method of Linear Model Response Plateau and the data are genotypes was proposed in this paper. We evaluated the vegetative characteristics, plant height, pseudostem circumference, number of green leaves at flowering, number of offspring produced at flowering and the number of green leaves at harvest. In the evaluation, each plant was treated as a basic unit (ub) with an area of 6 m², thus making up 360 units, which were combined adjacently to form 23 pre-set plot sizes, in rectangular shaped molds. The results showed that the plot sizes vary with the method used and the variable evaluated. In the method of maximum curvature the basic unit size ranged from 17.5 for the number of offspring at flowering (NFF) to 20.5 for the number of green leaves after harvest (NFVC), while the coefficient of variation ranged from 3.2 for plant height (AP) to 5.3 for the number of green leaves after harvest (NFVC). For the method of modified maximum curvature, the size of the basic unit ranged from 0.74 for plant height (AP) to 6.88 for the number of offspring at flowering (NFF), while the coefficient of variation ranged from 5.2 for plant height (AP) to 13.84 for the number of offspring at flowering (NFF). In the linear model response plateau the size of the basic unit ranged from 10.50 for the number of green leaves (NFV) to 13.22 for the number of offspring at flowering (NFF), while the coefficient of determination ranged from 71.77% for the number of offspring at flowering (NFF) to 80.33% for the number of green leaves (NFV). Thus, it is recommended to simultaneously use more than one method to determine the optimum plot size so that the size adopted attends to the needs, as far as possible, of the various factors considered in each method.
Descrição
Palavras-chave
Modelos lineares, Estatística, Curvatura, Bananeira, Linear models, Statistics, Curvature, Banana
Citação
SILVA, Willerson Custódio da. Comparison of methods to estimate the optimum plot size in musa. 2010. 49 f. Dissertação (Mestrado em Estatística Aplicada e Biometria) - Universidade Federal de Viçosa, Viçosa, 2010.