Sources of variation in corn silage and total mixed rations of commercial dairy farms

ABSTRACT Information on sources of variation in feed and diet characteristics is needed to develop appropriate strategies to reduce uncertainty and to separate true variation from that associated with measurements. The objectives were to determine sources of variation in DM content and particle size distribution in corn silage (CS) and TMR. Ten dairy farms in Argentina were visited on 3 consecutive days, samples of CS and TMR were taken, and an audit of feed management was conducted. Corn silage and TMR were sampled in duplicate each day. Variance components were calculated with the Mixed Linear Models of InfoStat for CS and Generalized Linear Mixed Models for TMR. For CS, the model included the effects of farm and day within farm, and for TMR, the model included farm, pen within farm, day within pen, and feed bunk site within pen. Residual effects accounted for sampling and analytical variation. Farm was the greatest source of variation for DM and particle size distribution of CS and TMR, explaining 40 to 92% of total variation. For CS, day within farm variation was greater compared with residual variation in DM (7 and 0.6%, respectively), meaning real changes occurred from one day to the other. For TMR, daily variation in DM content was high and possibly associated with feed management errors. For particle size distribution in TMR, sampling and assaying variation was greater than feed bunk site variation, indicating increased replication and averaging is needed to increase precision.