| Abstrakt: |
Sample moisture may impair accuracy in predicting forage quality based on near infrared‐reflectance (NIR), but the literature contains little specific information on the topic. The purpose of this study was to determine the extent to which moisture interferes with the ability of NIR to predict acid detergent fiber (ADF) in silage‐type sorghums [Sorghum bicolor(L.) Moench.]. Comparisons were made between laboratory and predicted ADF on samples that were oven‐dry or had been stored at 43% (low), 63% (medium), and 100% (high) relative humidities. Results showed no significant differences between laboratory and predicted ADF for samples that were dry or had been stored under conditions of low humidity. However, when samples were stored under medium or high relative humidities, a significant (P ≤ 0.05) over‐estimation of ADF resulted. Increased sample moisture lowered predictability, since the standard error of estimate increased from 1.27 and 1.18 for dry and low moisture samples, to 2.28 and 4.15 for medium, and high moisture samples, respectively. Consistently repeatable ADF values were obtained when calibrations and NIR analyses were conducted on oven‐dry material, which emphasizes that NIR analyses should be conducted only on recently oven‐dried material that has been stored in a desiccator. |
Plný text