RESUMO
BACKGROUND: Dry matter (DM) is a routine test for all animal feeds, facilitating feed comparisons and diet formulation. It is the most frequent test, yet the most challenging with respect to precision and accuracy. OBJECTIVE: Our objective was to evaluate the accuracy, repeatability, and physicochemical impacts of oven-drying times on LDM test results in animal feeds obtained by loss on drying (LoD) at 105°C. METHODS: Eighteen primary samples collected from different feed sources were grouped into high-moisture (HM) and low-moisture (LM) content materials. The tested methods were based on LoD at 105°C and Karl Fischer titration was adopted as the reference method. Test portions were oven dried at 105°C for 3, 6, 12, 16, and 24 h, and test results were compared to the reference method. Test portions were also subjected to a color evaluation using a colorimetric technique. RESULTS: The method based on 3 h of drying provided the closest estimates to those obtained by Karl Fischer titration. Extending heating time (i.e., above 3 h) increased the bias, especially for HM feeds, which was attributed to a higher occurrence of non-enzymatic reactions. This was corroborated by the color of the residues, which became darker with increased heating time. The repeatability of LoD methods was considered adequate, ranging from 0.32 to 0.73%. CONCLUSION: The LoD method based on the binomial 105°C × 3 h minimizes the bias in the water recovery and causes less non-enzymatic browning in the test portions. HIGHLIGHTS: The loss-on-drying method recommended for laboratory DM in animal feeds is drying the test portions at 105°C for 3 h.