Because of problems inherent in determining the digestible amino acid
content of batches of feedstuffs used for the formulation of feeds, fe
ed manufactures routinely overformulate for digestible amino acids. A
producer wishing to assure that no more than 20% of the batches of fee
d produced are nutritionally inadequate must overformulate by 7.5% for
a feed containing 65% maize, 25% soybean meal, and 5% poultry by-prod
ucts when using randomly-selected batches of these feedstuffs. Selecti
ng batches based on protein content allows for a reduction of this saf
ety margin to 6.3%, since protein is a predictor, albeit a poor one, f
or digestible amino acid content. Near-Infrared Reflectance Spectrosco
py (NIRS) was evaluated as a tool to directly predict the digestible a
mino acid content of animal meals with the objective to develop a nutr
itionally relevant quality control tool. Using a database containing 7
8 animal meals, we developed calibrations that explain 70-90% of the v
ariation in digestible amino acid content. Achieving such a degree of
accuracy for all feedstuffs of interest would permit overformulation t
o be reduced to 3%. The precision of NIRS technology and its short sam
ple analysis time would give feed mills an efficient, economical quali
ty control tool for producing feeds that more accurately match the nut
ritional profile calculated.