Feed particle size reduction in dairy cattle

Abstract

This research was carried out at CoRFiLaC in collaboration with the CoRFiLaC research group and technical staff. The understanding of how cows chew their feed might be one fundamental previous step necessary to study if, and how feed particle size could alter efficient feeding, milk production and composition and animal health. Our main objective was to learn some rules, of how feed particle size is reduced during the ingestive mastication in dairy cattle, in order to get an idea about rumen mat consistency from diet particle size and intake. We measured lengths distributions of feed and respective bolus particles which are potentially contributing to rumen mat formation, estimated an approximate dry matter proportion of this sample fraction, and evaluated differences in chewing behaviour between dry and lactating cows. Poppi et al. (1985) reported that particles retained on a 1.18 mm screen, using a vertical sieving technique, are highly resistant to passage from the rumen in cattle. As vertical sieving techniques divide particles by their widths rather than their lengths and Mertens (1984) reported constant lengths to widths ratios varying from 3.4:1 for corn silage to 10:1 for alfalfa and bermudagrass hay, we concluded that particles resistant to passage out of the rumen might have lengths probably not shorter than 5 mm. We focussed on the analysis of particle lengths > = 5 mm. Analyses were performed using an image analysis technique. A mean length (ML) was calculated considering only particle lengths > = 5 mm. In order to determine the mass of particles > = 5 mm, we found that a screen with 1.6 mm openings, using a horizontal wet sieving technique, separated best the particles > = 5 mm from those of smaller dimensions. The dry proportion of samples on that screen was called PROP_1.6. We performed two experiments. In the first experiment, 6 rye grass hay treatments differing in particle lengths, one corn silage, one grass silage and one TMR sample were fed to four dry and four lactating, rumen fistulated, dairy cows after rumen were emptied, and boli, rumen mat and feces were sampled. Rye grass hay treatments were as follows: rye grass cut at 50 mm lengths and dried to hay, long rye grass hay, chopped rye grass hay retained on the Penn State Particle Separator (PSPS) screens of 19 mm, 8 mm, 1.18 mm and on the bottom pan. In the second experiment we selected 10 total mixed ration (TMR) samples from Sicilian dairy farms. We divided TMRs into fractions by sieving them through four sequential sieves of 19 mm, 8 mm, 2.5 mm and 1.18 mm. All fractions and the unprocessed TMRs were fed to three dry dairy, rumen fistulated cows, after rumen were emptied, and boli were sampled. Feed (with exception of the long rye grass hay, rye grass cut at 50 mm lengths and dried, chopped rye grass hay retained on the 19 mm and the 8 mm screen), bolus, rumen mat and feces were sieved through a 1.6 mm screen to obtain PROP_1.6, particles retained were separated, imaged and ML was calculated. Dry matter (DM) and contents of crude protein (CP) and neutral detergent insoluble fiber (aNDF) were determined in the feed samples. The longest ML of a TMR particle the cows were able to swallow was about 18 mm. This was nearly twice the longest ML of rye grass hay bolus particles, which could be swallowed at ML < = 10 to 11 mm. As a difference to TMR particles, all rye grass hay fractions probably stimulated chewing to some extent. Even the smallest rye grass hay particles retained on the PSPS bottom pan were apparently reduced in PROP_1.6 although this fraction is not defined physical effective (Kononoff and Heinrichs, 2003). Ingestive chewing reduced PROP_1.6 of the unprocessed TMRs. Eating reduced also PROP_1.6 of TMR fractions retained on screens with openings of at least 2.5 mm, but not PROP_1.6 of the TMR fractions passing that screen. Eating reduced ML of TMR fractions retained on screens with openings of at least 8 mm, but neither ML of particles which had passed the 8 mm screen, nor ML of the unprocessed TMRs were reduced significantly (p > 0.05). The sum of TMR residues on the two upper PSPS screens with 19 and 8 mm openings might underestimate TMR physical effective fiber (pef), while the sum of residues from all three PSPS screens might overestimate TMR pef. We suggest the use of a sequential sieve set containing a 19 mm, a 8 mm sieve and an additional screen with 2.5 mm openings, instead of the additional screen with 1.18 mm openings, for diet evaluation on the farm. Feed and bolus PROP_1.6 were highly correlated (R2 = 0.94, 65 observations), when unprocessed TMRs and TMR fractions were fed, with y = 0.79x + 0.03 and y being bolus and x feed PROP_1.6. Rye grass hay particles and most silage samples were apparently chewed more intensely and having lower bolus PROP_1.6. Feed and bolus ML were highly correlated (R2 = 0.86, 47 observations), when unprocessed TMRs and TMR fractions were fed and feed ML < = 20 mm, with y = 0.76x + 2.08 and y being bolus and x feed ML. Most silages apparently fitted this regression, too. When rye grass hay was fed and feed ML < = 20 mm, feed and bolus ML were correlated with R2 = 0.43 at 21 observations, with y = 0.24x + 6.13 and y being bolus and x feed ML. Rye grass hay and TMR particles with ML > 20 mm were apparently chewed to constant lengths, with feed particle size not being related to bolus particle size. Chopped rye grass hay particles with ML > 20 mm were particles retained on the 19 and the 8 mm PSPS screen, whereas TMR with ML > 20 mm were retained only on the 19 mm screen. Only particle size of feeds with ML under this threshold might be able to influence parameters such as rumen retention time, intake and rumen degradation of feed, if these parameters were related to bolus particle size. Rye grass hay was chewed more intensely compared to TMR particles. Rye grass hay particles were dry, while TMR treatments contained 36 to 48% water. Rye grass hay particles contained 12 to 14 (% DM) CP and 54 to 59 (% DM) aNDF, whereas CP and aNDF of TMRs were 14 to 24 and 20 to 48 (% DM), respectively. During eating, chemical parameters influenced more reduction of ML of longer compared to shorter particles, but reduction of PROP_1.6 was more affected in the shorter particles. For each % decrease in sample DM bolus ML increased approximately 0.2 mm, under the particular condition where feed particles ML ranged between 14.7 and 43.7 mm, CP content ranged between 12.8 and 13 (% DM) and aNDF content ranged between 47.5 and 51 (% DM). Chewing behaviour of dairy cows was not altered by physiological stage nor by the interaction of treatment feed by physiological stage. Bolus PROP_1.6 and ML from dry cows were not different from lactating cows. Even though dry and lactating cows received different diets, PROP_1.6 and ML of rumen mat and feces were alike. Approximately 26% and 36% dry matter of rumen mat was retained on a 1.6 mm screen in dry and lactating cows, respectively. Rumen mat particles retained on that screen and > = 5 mm, had ML of 8.6 mm in dry and 10.3 mm in lactating cows. There was a trend for higher PROP_1.6 of fecal particles from dry cows compared to lactating cows (p = 0.105), but ML of 7.4 and 7.9 in dry and lactating cows, respectively, were not statistically different. The lengths of the longer particle fraction from TMR, which was retained on a 1.6 mm screen, were reduced to a higher extent during rumination compared to the longer fraction of ingested particles from rye grass hay. Only approximately 38% or less of TMR DM is constituted by particles potentially retained in the rumen. However, these particles might need a longer time for rumination compared to chewed hay particles. In both, dry and lactating cows, proportions of rumen mat particles > 4 and > 6 mm, respectively, at individual lengths, retained on a 1.6 mm screen, were highly correlated to proportions of bolus particles. Rumen mat ML can be estimated from lengths distribution of bolus particles retained on a 1.6 mm screen.