Five trials were conducted to develop a system to estimate rumen-undegradab
le protein (RUP) of legume and grass silages using near infrared reflectanc
e spectroscopy. In situ procedures were the reference method used to determ
ine silage RUP content. Trials 1, 2, and 3 were devoted to improving in sit
u procedures. In trial 1, alfalfa silage with and without heat treatment wa
s incubated ruminally in 30 cows. The standard deviation of in situ RUP att
ributable to cow and diet was 0.82 and 3.80 g/10(-1) kg of crude protein 1(
CP) for the unheated and heated alfalfa, respectively. Based on trial 1, it
was determined that 8 cows would be required to establish RUP standards. I
n trial 2, low (13.3 g/10(-1) kg of CP) and high (44.5 g/10(-1) kg of CP) R
UP standards were developed using eight ruminally cannulated cows.
In trial 3, 11 new RUP standards were developed by mixing trial 2 RUP stand
ards together. The RUP standards were used to employ a calibration curve te
chnique in ruminally cannulated cows. The technique was employed in four ru
minally cannulated cows to estimate RUP contents of 121 silages, and RUP va
lues were used for near-infrared reflectance spectroscopic analysis in tria
l 4. Trial 4 procedures yielded a calibration for RUP content of silages wi
th an R-2 of 0.84 and a standard error of calibration of 1.55 g/10(-1) kg o
f CP. In trial 5, the equation was tested on 300 silage samples. The mean p
redicted RUP content was 21.8 g/10(-1) kg of CP. Data suggest near-infrared
reflectance spectroscopy can predict RUP content of silages.