The correlation between physical parameters of guava like axial dimensions, projected area, volume, and mass is essential for developing postharvest machineries especially grading systems. The present study focused on measuring physical characteristics (dimensions, projected area, and volume) of guava (cv. Allahabad safeda), and the development of predictive linear and nonlinear (linear, quadratic, power, and S-curve) models to determine the mass of guava. The fruits were graded based on the maximum equatorial diameter in three grades that is, large (phi = 66-75 mm), medium (phi = 54-65 mm), small (phi = 43-53 mm), and mass modeling was performed. The model equations were also fitted on ungraded fruits samples for comparison purpose. The major, intermediate, minor intercepts, geometric mean diameter, weight, volume, and criteria projected area of the ungraded lot were 63.76 +/- 6.03 mm, 59.90 +/- 4.71 mm, 59.66 +/- 4.43 mm, 61.05 +/- 4.76 mm, 126.80 +/- 30.88 g, 132.6 +/- 35.0 cm(3), and 33.10 +/- 8.17 cm(2), respectively. It was observed that predictions of mass models fitted on ungraded fruit lots were found superior to fitted on individual grades. The higher coefficient of determination (R-2) and low mean relative deviation (MRD) indicated that quadratic models based on geometric mean diameter (R-2 >=.984, MRD = 2.32) and ellipsoidal volume (R-2 >=.986, MRD = 2.28) can effectively predict the mass of guava fruits. The possible applications of established mass models for developing an integrated and effective grading system and the prospective utilization of graded fruits for processing into a variety of value-added products are also discussed. Practical Application Fruits with uniform grades usually have higher demand and consumer preference. Grading is the essential unit operation in postharvest management to achieve dimensional uniformity. The grading process becomes complex when fruits are graded with a similar appearance but difference in mass; therefore, mass-based grading of fruit plays a vital role in the design of advanced machineries. Recent advancements and automation employ mass as a parameter that enhances the overall efficacy of grading operations. The developed mass models and outcome of this study will be beneficial for developing advanced grading machineries. The possible integration of machine vision systems with developed mass models will simultaneously enable the grading of guava with both dimension and mass. Additionally, information on some grade-specific physical properties and their potential utilization will be necessary for the design of process equipment and the development of various value-added products.
