Evaluation of markers, sampling sites and models for estimating rates of passage of silage or hay in dairy cows

Abstract

A study was conducted using seven rumen-cannulated lactating Holstein cows (mean body weight, 619 kg; mean milk production, 20.9 kg per day) to compare markers [ytterbium (Yb) and chromium (Cr)], sampling sites (rumen and rectum) and models (multi-compartmental and double-compartmental) for estimating passage of silage and hay. Cell walls from alfalfa silage and alfalfa—grass hay were prepared and subsequently mordanted with 5% Cr or labelled with 2.5% Yb by soaking. Cows were fed diets containing 60% concentrate and 40% forage (dry matter basis), with ingredients allocated separately twice daily. Dry matter intake averaged 16.6 kg per day. Faecal samples were obtained every 3 h until 24 h followed by every 4 h until 72 h, every 6 h until 108 h, and then every 12 h until 144 h post-dosing. Rumen contents were sampled from the ventral sac below the fibre raft every 4 h until 72 h post-dosing. Parameters estimated by the double-compartmental model were dependent upon the initial parameter estimates. Consequently, rumen turnover rate ( k 1) was first estimated using only the terminal exponential. Good initial parameter estimates were not necessary using the multi-compartmental model. Both models provided similar estimates of k 1 and mean retention time in the digestive tract ( P > 0.05); however, apparent turnover rate of the second compartment ( k 2) was double ( P < 0.01) and transit time was 35% longer ( P < 0.01) when estimated using the multi-compartmental model compared with the double-compartmental model. The apparent rumen turnover rate of Yb-labelled silage was 40% faster ( P < 0.01) than Cr-mordanted silage, whereas transit time was 25% longer and k 2 estimates were similar for both markers. Estimates of rumen turnover rate obtained by rumen sampling were 16% greater than rates determined using both faecal models, and estimates between sites were only moderately correlated ( r = 0.7).