Traditional analytical methods of assessing forage nutritional value are time-consuming and expensive. The near infrared reflectance method appears as an alternative that does not utilize chemicals, is not destructive and is extremely fast. However, in order to be used in a routinely basis, calibration curves have to be developed from laboratory results and mathematical equations. The objective of this paper was to develop calibration curves for corn silage, using a total of 246 samples collected in different regions of Rio Grande do Sul. The samples were analyzed using the reference methods to determine crude protein (CP), dry matter (DM), acid detergent fiber (ADF), neutral detergent fiber (NDF) and minerals (Ca, K, Mg and P). The same samples were also analyzed by spectroscopy, using a wave length region that ranges from 1100 to 2500 nm. Data was mathematically treated by the Match program and then a regression model was fitted, having the results from the chemical analysis as independent variables and the results from the near infrared reflectance as dependent variables. The evaluation of the adjustment level was made through the determination coefficients (R²). The accuracy was elevated (R² = 0.99) for CP, ADF, NDF and DM, indicating the validity of this method to predict such parameters in corn silage. Even though the R² for minerals have ranged from 0.92 to 0.94, further research about the behavior of the spectrum is needed in order to evaluate the type of mineral complex formed and its relation with the chemical method. The averages for CP, NDF, ADF and DM were 7.86, 60.66, 29.87 and 94.17%, respectively. The averages for Ca, P, K and Mg were, respectively, 0.27, 0.21, 1.14 and 0.22%.
forage analysis methods; protein; fiber; dry matter; minerals
