A study to investigate whether plants adjacent to failed plants are able to partially or completely compensate the absent plants, and whether this compensation depends on clone and edaphic-climatic conditions was carried out in two northwestern Minas Gerais counties. The experiments were set up in December 1998. A randomized complete block design with split plot treatments and three replications was used. The plots consisted of seven different clones and the split plot was the variable number of plants in the stand, as a function of the failure percentage. The split plot that did not present any failure had 10 plants in a row. To constitute the other plots, five different failure levels were adopted (10, 20, 30, 40 and 50%). The assessed trait was wood volume (solid m³/ha) at 37 months of age. Individual analyses of variance were performed for each location. Joint analyses of location were also performed. Both types of analyses involved data at the plant mean and plot total levels. Sources of variation that involved the variable stand were partitioned into linear and quadratic effects and deviation from the regression. The regression b values were estimated for stand, individual clone and location regardless of the clone. The following conclusions were obtained from the results: a) most of the assessed clones did not totally compensate the absence of the neighboring plants; b) the compensation capacity varies with clones and edaphic-climatic conditions; c) regardless of the clone or location, stand correction by regression, and especially by covariance, improved the clone selection efficiency, in particular with the total plot data.
Eucalyptus urophylla; clone; stand; failure
