The application of near infrared reflectance spectroscopy in the nutritional assessment of tree leaves as potential protein supplements for ruminants
Mguvane, Mnisi Caven
MetadataShow full item record
This study was executed to calibrate and validate near infrared reflectance spectroscopy (NIRS) for use in predicting chemical composition, buffer nitrogen (N) solubility, in vitro ruminal dry matter (DM) and nitrogen (N) degradation and the in vitro ruminal fermentation of leaves from Acacia erioloba, A. nilotica and Ziziphus mucronata tree species harvested from two different growth environments. Section One defines the purpose of the study, while Section Two reviews the nutritional importance of browse leaves as potential sources of nutrients, particularly protein, for optimal animal production. In Section Three, the chemical composition of A. erioloba, A. nilotica and Z. mucronata leaves harvested from Molelwane and Masuthle, which are 40 km apart, was determined. Results from the study showed that growth environment influenced some chemical components but not others. Section Four presents an assessment of buffer solubility of N and in vitro ruminal DM and N degradability in tree leaves. Results showed that leaves with high buffer N solubility had high in vitro ruminal degradability. However, the presence of secondary plant compounds in the leaves was shown to affect their rumen degradability. Section Five presents an investigation into the in vitro ruminal biological activity of tannins present in the tree leaves with the aid of tannin-binding polyethylene glycol (PEG). An automated in vitro ruminal gas production technique was used as the tannin bioassay. The PEG inclusion, for all tree species, increased gas production and in vitro organic matter degradability; however, it reduced the partitioning factors. In Section Six, the NIRS was calibrated and validated as a rapid technique for the prediction of chemical composition and in vitro ruminal degradability of browse leaves. Results showed that NIRS can be a reliable tool to predict total N content because the NlRS model explained more than 80% of variation in total N in an independent sample when externally validated. It was concluded that a larger number of samples with accurate wet chemistry is required to increase the accuracy of prediction of other chemical components as well as in vitro rumen fermentation by NIRS.