GENETIC EVALUATION OF HALF-SIB Eucalyptus urophylla PROGENIES BY THE REML/BLUP AND MINIMUM SQUARES PROCEDURE

Realizou-se avalia\ue7\ue3o gen\ue9tica em cinco testes de prog\ueanies de meio-irm\ue3os de Eucalyptus urophylla S. T. Blake procedentes da Indon\ue9sia, com o uso dos procedimentos REML/BLUP (m\ue1xima verossimilha\ue7a restrita/melhor predi\ue7\ue3o linear n\ue3o-viesada) e pelo m\ue9todo dos quadrados m\uednimos E (QM). Os ensaios foram estabelecidos separadamente por proced\ueancia, sendo o delineamento experimental em blocos casualizados, com cinco repeti\ue7\uf5es e parcelas lineares de seis plantas, no espa\ue7amento 3,0 x 2,0 metros, em Guanh\ue3es, MG. Nos cinco testes avaliados aos 58 meses de idade, para a caracter\uedstica di\ue2metro \ue0 altura do peito (DAP), apresentaram valores dentro dos padr\uf5es da esp\ue9cie, al\ue9m de exibir variabilidade gen\ue9tica significativa, pelo teste F a 1% de probabilidade. Os coeficientes de varia\ue7\ue3o gen\ue9tica aditiva apresentaram valores maiores no teste TP-71 originado de Wetar e no TP-68 originado de Alor, e nos outros testes apresentaram valores similares para a caracter\uedstica di\ue2metro \ue0 altura do peito (DAP).Os ganhos de sele\ue7\ue3o foram estimados na ordem de 12,8; 22,9; 9,5; 21,0 e 25,3% e tamanho efetivo populacional (Ne) na ordem de 60,2; 58,0; 131,8; 167,9 e 224,1 para ao testes TP-67, TP-68, TP-69, TP-70 e TP-71 respectivamente. O ganho de sele\ue7\ue3o no Pomar de Sementes Clonal (PSC) foi de 26,8 % com a sele\ue7\ue3o dos 21 indiv\uedduos portadores dos maiores valores gen\ue9ticos aditivos. Nos dois procedimentos, m\ue1xima verossimilhan\ue7a restrita (REML) e esperan\ue7a de quadrados m\uednimos E(QM), os valores dos par\ue2metros gen\ue9ticos foram semelhantes, exceto entre familias, em que o procedimento REML proporcionou valores mais elevados com acur\ue1cia superior a 70% em todas as popula\ue7\uf5es, mostrando-se como ferramenta apropriada para esse fim.Five progeny tests with half-sibs Eucalyptus urophylla S. T. Blake were evaluated. Four 4 of them were from provenances Indonesia and one of them from Brazil. It was used the REML/ BLUP (restricted maximum likelihood / best linear unbiased prediction) procedures and the minimum square method for the estimation of genetic values. The tests were established separately according to the origin, in a complete randomized block design, with five replications and row plots of six plants in 3.0 x 2.0 meter spacing in Guanh\ue3es, Minas Gerais State. When the plants were 58 months-old for the characteristics of diameter at breast height (DAP), the five evaluated tests showed standard species values, besides presenting significant genetic variability by the F test at 1% of probability. The coefficients of additive genetic variation presented greater values in the TP-71 test from Wetar and TP-68 from Alor, and presented similar values in other tests for the diameter at breast height (DAP) characteristic. Selection gains were estimated at values of 12.8; 22.9; 9.5; 21.0 and 25.3%, and the effective population size (Ne) at values of 60.2; 58.0; 131.8; 167.9 and 224.1 for the TP-67, TP-68, TP-69, TP-70 and TP-71 tests, respectively. A selection gain of 26.8 % was obtained in the clonal seed orchard with selection of 21 plants that were carriers of the greatest additive genetic values. In both procedures, the values of the genetic parameters were similar, except among families, in which the REML procedure em obtained higher values with accuracy higher than 70% in all populations, proving to be appropriate for this purpose