Prediction of aspects of neutral detergent fibre digestion of forages by chemical composition and near infrared reflectance spectroscopy

7 pages, 5 tables.

Sixty-two herbage samples, harvested in natural meadows located in the mountains of Leon (north-west Spain), and characterised by a diverse botanical composition and different stages of maturity of the plants, were used to evaluate the ability of chemical composition and near infrared (NIR) spectroscopy to predict in vitro digestibility and in sacco degradability of the neutral detergent fibre (NDF) fraction. In vitro digestibility was performed as described by the Goering and Van Soest procedure. Three dry Holstein-Friesian cows fitted with a rumen cannula were used to incubate the herbage samples. A Bran+ Luebbe InfraAlyzer 500 spectrophotometer was used to obtain the NIR spectra corresponding to the 62 original herbage samples. Prediction equations for the estimation of in vitro digestibility and in sacco degradability parameters of the NDF fraction were generated using NIR spectra or chemical data as independent variables. The results showed that the in vitro digestibility and kinetic parameters of degradation of the NDF fraction could not be predicted accurately, probably as a consequence of the errors corresponding to the reference methods. In contrast, these errors did not greatly affect the extent of disappearance of the NDF fraction at later times, so the accuracy of prediction of these parameters was higher, especially when NIR spectra were used as independent variables. This is probably due to the close relationship that the parameters showed with the chemical data, since this kind of information, together with some physical characteristics of the samples, is included in the NIR spectra.

S. Andrés is grateful to the ‘Junta de Castilla y León’, Spain, for financial assistance with a pre-doctorate grant. We also thank the Animal Production and Feed Science Department of the University of Zaragoza (Spain) and the Animal Production Department of the University of Extremadura (Spain) for their essential contribution to this work. The project was funded by the Spanish Ministry of Science and Technology (Project 1FD97-0776).

Peer reviewed