In Situ Prediction of Fresh and Dry Weight of Peach Fruit

Abstract

To estimate the effect of treatments or cultural practices over fruit dimensions on peach, such as fresh and dry weight and equatorial and polar diameters (FW, DW, ED and PD, respectively), the use of destructive samples are frequent. These variables are generally not included, due to the time-consuming and research costs involved. With the purpose to determine in situ the FW and DW of fruit of seedling peach trees, two cubic regression models (CRM) were fitted with 1241 and 1119 field observations of FW, DW, and ED. To determine DW, fruits were cut off immediately and dried at 70C for 24-h. These measurements were taken during growing season of 1984 and 1985. At 2-week intervals, 12 samples were collected each year. Each sample consisted in harvesting randomly five fruits and around the middle part of trees. The CRM were fitted taking the mean of five fruits. FW and DW were used as dependent variables, while ED as independent variable. To validate both models, during the growing season of 1985, 11 samples (five fruits per sample) were taken again from other trees. The real and predicted values of FW and DW were analyzed by a linear regression model (IRM), to know the grade of adjustment between them. The CRM of both variables had significant fit (r2 = 0.975 and 0.941 for FW and DW, respectively). In contrast, the highest variation coefficient was observed in DW (29.14%), compared with FW (13.4%). In both cubic models, error mean square was the lowest compared to other models. The linear relation between real and predicted values ha values of r2 = 0.983 and 0.941 for FW and DW, respectively; while the variation coefficients were 9.59% (FW) and 17.32% (DW). The CRM's can be used in future seedling peach experimental works, to predict fruit weight after full bloom until harvest.