The tomato plant yield of the first five clusters was used in the development of mathematical model to estimate yield obtained in nine trusses. Six experiments were carried out, being four in unheated greenhouse (A, B, C and D) and two in the field (E and F), in the Federal University of Viçosa. All the experiments consisted of ten treatments with differentiated N rates that ranged from 0 to 1,953.7 kg./ha-1. The data up to the 1st, 2nd, 3rd, 4th and 5th trusses accumulated total yield of the experiments C and E were submitted to the analysis of variance and regression models and were adjusted to 9th trusses accumulated yield. The proposed models validity was based on the correlation (r) among the observed and estimated data by the model and by the medium square deviations (MSD) analysis and its additive components. The proposed models, based on the yield accumulated until 5th trusses were precise in estimate the total fruits yield accumulated in the tomato plant with 9 trusses in unheated greenhouse and field. The models should be compared and validated with data of other studies. The medium square deviations (MSD) analysis and its components are useful to verify the regression models validity through the model deviations quantification and identication causes of those deviations.
Lycopersicon esculentum; truss; yield estimate
