Anderson Rodrigo da Silva

Tamanho de amostra para caracterização morfológica de frutos de pimenteira

O objetivo deste estudo foi determinar o tamanho apropriado de amostra por meio da tecnica de simulacao de subamostras para a caracterizacao de variaveis morfologicas de frutos de oito acessos (variedades) de quatro especies de pimenteira (Capsicum spp.), que foram cultivadas em area experimental da UFPB. Foram analisados tamanhos reduzidos de amostras, variando de 3 a 29 frutos, com 100 amostras para cada tamanho simulado em um processo de amostragem com reposicao de dados. Realizou-se analise de variância para os numeros minimos de frutos por amostra que representasse a amostra de referencia (30 frutos) em cada variavel estudada, constituindo um delineamento experimental inteiramente casualizado com duas repeticoes, onde cada dado representou o primeiro numero de frutos na amostra simulada que nao apresentou nenhum valor fora do intervalo de confianca da amostra de referencia e que assim manteve-se ate a ultima subamostra da simulacao. A tecnica de simulacao utilizada permitiu obter, com a mesma precisao da amostra de 30 frutos, reducoes do tamanho amostral em torno de 50%, dependendo da variavel morfologica, nao havendo diferencas entre os acessos.

Métodos de agrupamento em estudo de divergência genética de pimentas

The objective of this study was to compare three methods to determine number of groups in studies with hierarquical cluster analysis, based at data from characterization of Capsicum accessions, in order to identify those with the greatest power of discrimination. Mojena method, Tocher method and RMSSTD method were applied to determine the optimal number of groups formed in final stage of the UPGMA method. Forty nine Capsicum chinense accessions from the Vegetable Germplasm Bank of the Universidade Federal de Vicosa were analyzed, in relation to ten morphological characters for identifying the most similar accessions group, making possible the selection of superior genotypes, i.e., of commercial interest. The RMSSTD method allowed to conclude on the existence of seven groups, with a greater power of discrimination for this method, compared to the Tocher optimization method and the Mojena method, which formed, respectively, four and three groups.

Morphological phenotypic dispersion of garlic cultivars by cluster analysis and multidimensional scaling

Multivariate techniques have become a useful tool for studying the phenotypic diversity of Germplasm Bank accessions, since they make it possible to combine a variety of different information from these accessions. This study aimed to characterize the phenotypic dispersion of garlic (Allium sativum L.) using two multivariate techniques with different objective functions. Twenty accessions were morphologically characterized for bulb diameter, length, and weight; number of cloves per bulb; number of leaves per plant; and leaf area. Techniques based on generalized quadratic distance of Mahalanobis, UPGMA (Unweighted Pair Group Method with Arithmetic Mean) clustering, and nMDS (nonmetrric MultiDimensional Scaling) were applied and the relative importance of variables quantified. The two multivariate techniques were capable of identifying cultivars with different characteristics, mainly regarding their classification in subgroups of common garlic or noble garlic, according to the number of cloves per bulb. The representation of the phenotypic distance of cultivars by multidimensional scaling was slightly more effective than that with UPGMA clustering.

Soil chemical management drives structural degradation of Oxisols under a no-till cropping system

Physical degradation of the subsurface layer of soils reduces the effectiveness of no-till (NT) as a sustainable soil management approach in crop production. Chemical factors may reduce the structural stability of Oxisols and thereby exacerbate compaction from machinery traffic. We studied the relationship between chemical management and structural degradation in Oxisols cultivated under NT at three sites in southern Brazil. The surface and subsurface layers of the soils were characterised chemically and mineralogically and three physical attributes related to soil structural stability (readily dispersible clay in water, mechanically dispersible clay in water, and water percolation) were quantified for each layer. The same characterisations were performed on Oxisols collected from adjacent non-cultivated areas, to provide reference data for non-degraded soil. The levels of dispersed clay in the cultivated soil from the surface layer matched those of the non-cultivated soil, but for the subsurface layer higher dispersed clay levels in the cultivated soil showed that it was physically degraded relative to the non-cultivated soil. Water percolation was found to be slower through the Oxisols cultivated under NT, irrespective of the soil layer. The relationships between the three indicators of soil structural stability and the measured chemical and mineralogical variables of the soils were explored through an analysis of canonical correlation. The principal variables associated with the lower stability of the cultivated vs non-cultivated Oxisols were the lower concentrations of organic carbon and exchangeable aluminium and, for the surface layer, the higher pH. It is argued that structural degradation of Oxisols cultivated under NT, observed predominantly in the subsurface layer, has been aggravated by the accumulation of amendments and fertilisers in the surface soil and reduced levels of organic matter, especially in the subsurface layer.

An alternative procedure for performing a power analysis of Mantel's test

This study proposes a simple way to perform a power analysis of Mantels test applied to squared Euclidean distance matrices. The general statistical aspects of the simple Mantels test are reviewed. The Monte Carlo method is used to generate bivariate Gaussian variables in order to create squared Euclidean distance matrices. The power of the parametric correlation t-test applied to raw data is also evaluated and compared with that of Mantels test. The standard procedure for calculating punctual power levels is used for validation. The proposed procedure allows one to draw the power curve by running the test only once, dispensing with the time demanding standard procedure of Monte Carlo simulations. Unlike the standard procedure, it does not depend on a knowledge of the distribution of the raw data. The simulated power function has all the properties of the power analysis theory and is in agreement with the results of the standard procedure.