The Physically effective fiber of total mixed rations and its effects on dairy cow performances

The aim of the present study was to test the differences between the Penn State Particle Size Separator (PSPS) and the Z-Box from the laboratory one (Endecotts Octagon 200). To compare faecal particles fraction retained by using three different sieve methods, a monolayer screen (1.18 mm), multilayer (Cargill Digestion Analyzer) and the laboratory one. To estimate whether NIRS could be used to predict particle size distribution of TMR samples, using as reference the PSPS.Another objective was to test the mathematical models proposed by Mertens (1997) where peNDF of TMR was regressed vs. the percentage of milk fat.The results showed that the Z-box was able to accurately predict the physically effective NDF of the TMR, while the PSPS grossly over predicted it. The mean differences between PeNDF(PSPS) vs. laboratory peNDF was 4.63% (P< 0.01), while peNDF(Z-Box) was 0.1% (0.11NS). NIRS was demonstrated to be good to predict particle size and to determine peNDF in TMR samples. The regression of measured peNDF(PSPS) and predicted peNDF(NIRS) was not different from the Y = X line (P < 0.1), R2= 0.95. The faecal fractions retained using monolayer and multilayer sieve were poorly correlated to the laboratory fraction, R²= 0.12 and R²= 0.27 respectively.Neither the fiber content of the diets nor their peNDF affected milk fat content. The most important variable was identified in the DM content of TMR, which in turn was inversely associated to its silage content.