Using ytterbium-labeled forage to investigate particle flow kinetics across sites in the bovine reticulorumen

To better understand particle flow kinetics across the reticulorumen (RR), we monitored concentrations of Yb-marked forage across several RR sites over time. We offered three forages (bromegrass, low-fiber alfalfa, high-fiber alfalfa) ad libitum to lactating Holstein dairy cows in an incomplete 3 × 3 Latin square design. After dosing Yb-marked forage in the mid-dorsal sac (DS), we determined Yb concentrations of particulate samples taken from several sites (DS, ventral sac [VS], ventral blind sac [VBS], reticulum [RT]) from 0 to 72 h. For each site we constructed marker concentration profiles and applied non-parametric (local quadratic) and parametric (two-compartmental, G2 → G1 → O) regression models to these profiles. We determined mean residence time (MRT) and peak marker concentration using the original profile observations and regressions. Across forage diets, intake of dry matter differed (P=0.022) but that of other chemical fractions did not (P=0.125). The local regression fit the data well visually, and MRT calculated using this model and original observations closely agreed. This suggests the local regression successfully smoothed the original observations, which had considerable stochastic variation, without changing the underlying shapes they implied. Similar considerations showed the G2 → G1 → O model fit the data poorly. Visually, marker profiles shapes varied greatly, but we found few differences attributable to diet and site. Mean residence and peak marker concentration times, which reflect marker profile shape, did not strongly differ by diet (P=0.434, 0.0824) or site (P=0.113, 0.0824), though they did by cow (P=0.036, 0.004). These results further show differences in profile shapes (and the particle flow kinetics they reflect) are explained poorly by diet and site but somewhat explained by unidentified animal factors. Based on studies that demonstrate particles flows through RR in the partial order DS → VBS ↔ VS → RT, we expected MRT values to follow the order DS < VS ≈ VBS < RT. Our finding that MRT values were similar across sites suggests that once a particle escapes from the DS for the final time, it must escape from ventral regions (VS, VBS, RT) soon after entry. Our attempt to model these particle flow kinetics more quantitatively failed. Still, the experiment suggests that feed products designed to initially enter the ventral rumen and resist regurgitation should pass quickly from the RR.