Prediction of grain content of compounded feeds

Ten sets of compounded feeds (T₁–T₁₀) having 2 to 20% grain of maize (MC), rice kani (RC) or wheat (WC) and six more sets of compounded feeds containing 25 to 50 per cent maize (T₁₁–T₁₆) were analyzed to develop equations for predicting the grain level. At the same level of grain, the NDF value differed (P<0.05) for MC, RC and WC but the starch content remained similar. Every 2% increase in grain level increased (P<0.05) the starch content of the compounded feeds but the change in trend of NDF values was less consistent. For predicting the grain content, 14 simple linear regression equations based upon the starch and NDF values of the compounded feeds were developed and the predictability of grain content had more than 85 per cent coefficient of correlation (r²). To test the accuracy of the derived prediction equations, several compounded feeds with known grain levels of 10 – 48% were formulated. The starch and NDF values were fitted in the derived equations and the r² values were determined. The equation Y = −14.31 + 1.323 X; where, Y = the grain content and X = the starch content of the compounded feed, had high accuracy of prediction with grain content up to 30 per cent. From the above study, it may be concluded that the prediction equations based on the starch content are most reliable for predicting the grain content of the compounded feeds containing maize, rice kani or wheat up to 30 per cent level.