Análises biométricas de palmeiras juçara de fragmentos florestais no sul do Espírito Santo

Nenhuma Miniatura disponível
Data
2015-02-19
Autores
Senra, João Felipe de Brites
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal do Espírito Santo
Resumo
The species Euterpe edulis Martius is native to the Atlantic Forest and popularly known as jussara palm. It is a key species of the Atlantic Forest, producing fruits that are consumed by about 30 bird species and 13 mammal species, and also presenting great economic potential for exploitation of its fruit in the form of pulp, sweets and ice creams. Currently the jussara palm is on the official list of endangered species in Brazil. The objectives of this study were to: estimate the coefficient of repeatability and determination of the characteristics of fruits and seeds, analyze the number of measurements for a 95% confidence level and test different methods for estimation of repeatability, study the genetic diversity among accessions and evaluate the pattern of seed germination via the Logistic, Gompertz, von Bertalanffy and Will Bill nonlinear models, quantify the inter-population genetic diversity of the species based on the correlation between plant traits and related to fruit production. Collection of the jussara palm fruits was performed in 20 forest fragments in the following municipalities in the southern region of the state of Espírito Santo: Alegre (eight fragments), Guaçuí (four fragments), Ibitirama (three fragments), Jerônimo Monteiro (one fragment), Mimoso (three fragments) and Muqui (one fragment). For the study of repeatability 198 accessions were used, 25 fruits per plant, evaluating: longitudinal and equatorial diameter of the fruit and seed (DLF, DLS, DEF and DES) in millimeters (mm), and fruit and seed mass (MF and MS) in grams (g). Deviance was estimated by the restricted maximum likelihood (REML). The coefficient of repeatability and determination and the required number of measurements was determined using the REML methods, principal components via the correlation matrix, and covariances and structural analysis based on correlation and covariance matrices. Estimates of the repeatability coefficients and determination of all variables analyzed were greater than 0.78 and 98% respectively, for all tested methodologies, performing 25 measurements. For 95% reliability, 5 measurements are required for the variables DLF, DEF, DLS, DES and MS and 4 measurements for the variable MF. The methods used do not differ with respect to estimate of the repeatability coefficient for the variables DLF, DES, MF and MS, while for the other variables the differences were minimal. To analyze the identity of the models, seeds from 40 accessions were used to evaluate: germination, germination speed index, mean germination time, first germination count and the germination percentage. Based on the quantitative data the generalized Mahalanobis distance matrix (d²) was calculate and later accessions were clustered by the Tocher optimization method divided into four groups. In each group germination was analyzed via non-linear regression using the Logistic, Gompertz, von Bertalanffy and Will Bill models. Fit of the models was evaluated by the determination coefficient (R²), residual mean error (QMR), mean absolute deviation (MAD), Akaike information criterion (AIC) and Bayesian information criterion (BIC). The identity test of the models was performed by the F-test. The Gompertz model showed intermediary values of ß1, ß2 and ß3 compared to other models which avoid under or over estimation of these regression parameters; in addition, this model presents the highest R 2 values and the lowest values of QMR, DMA, AIC and BIC, and is the most indicated to describe the pattern of germination. The identity test of the models was not significant and therefore a single regression is not efficient to describe the germination process. Groups three and four have the highest values of TG, indicating greater vigor, and for seedling production the ideal time for transplanting would be 30 days for groups one, two and 4, and 22 days for group three. For the study of genetic diversity, data from fruit of the 198 accessions utilized in the analysis of repeatability were used. In this case we evaluated the traits of: mass of one hundred fruits (MCF) and mass of one hundred seeds (MCS) in grams, pulp yield (RP) in percentage, stem diameter at 1.5 m above the ground (D1,5), stem diameter at 1.0 m above the ground (D1,0) and stem diameter at 0.5 m above the ground (D0,5) in millimeters, number of clusters (CN), cluster length (CC) in centimeters, number of rachilles (NR), and height of the cluster (AC) in meters. Based on these data the genetic correlation was estimated, and a principal components analysis and correspondence analysis were performed. Genetic correlation between the traits MCF x NR (0.44) is positive. The characteristic NR is the main determinant in variation of MCF and indirect selection can be effective, depending on estimates of heritability. The fragments GU3, AL4, JM1, MI1 and MQ1 from different municipalities are those that presented the greatest inter-population genetic divergence among the 20 fragments studied. The fragments GU2, GU3 and AL4 have exceptionally large individual producers regarding the mass of one hundred fruits (211.54 to 352.24 g) which make them important sources of selection for use in jussara palm breeding programs focused on the production of fruit pulp.
Descrição
Palavras-chave
Euterpe edulis , Biometrics , Germination , Pre breeding , Repeatability , Diversity , Pré-melhoramento , Repetibilidade (Genética)
Citação