Examining feed preference of different pellet formulations for application to automated milking systems

Feed is often offered to a cow in the milking unit of an automated milking system. This offering provides nutrients but also acts as a reward to the cow for entering the unit. To complement the partial total mixed ration and to enable handling, flow, and delivery within this mechanized system, this offering is usually a mix of feeds that are combined and manufactured into a feed pellet. The objective of this experiment was to compare 4 different pelleting formulation strategies and measure the effects of feed preference in lactating Jersey cattle. To test the objective, a taste preference experiment was conducted with 8 multiparous lactating Jersey cattle (289 ± 25.3 d in milk, 26.0 ± 2.45 kg of milk yield, 19.36 ± 1.29 kg of dry matter intake). Four formulation strategies were tested including (1) a pellet containing feeds commonly included in the concentrate mixture of a total mixed ration, including 43.1% corn grain, 26.3% dried distillers grains, 3.18% soybean meal, and 5.6% vitamin and mineral premix (CMIX), (2) a pellet of dry corn gluten feed (CGF), (3) a pellet including feedstuffs that are considered to be highly palatable (53.2% wheat middling, 15.7% dried corn distillers grains and solubles, 15.2% cane molasses, and 1.81% oregano (FLVR), and (4) a high-energy pellet (ENG) consisting of 61% corn grain and 26.2% wheat middlings. Cows were offered 0.50 kg of each in a randomized arrangement within the feed bunk for 1 h or until the feed was fully consumed. According to the procedure, cows were offered all 4 treatments for the first 4 d, then the most preferred feed for each cow was removed, and the remaining 3 feeds were offered for 3 d. The process was repeated for the last 2 d. Feed preference was ranked from 1 to 4 with 1 being the most preferred and 4 the least. The resulting preference ranking was CGF (1.25 ± 0.463), FLVR (2.5 ± 0.926), CMIX (2.88 ± 0.835), and ENG (3.13 ± 0.991). These results were subsequently examined utilizing the Plackett-Luce analysis to examine the probability animals would choose a given pellet first based on the current data set. The analysis determined probabilities of first choice as 78.6 ± 0.601% CGF, 9.38 ± 0.438% FLVR, 4.94 ± 0.453% ENG, and 7.11 ± 0.439% CMIX. A Z-test was also conducted to determine if the percentage a treatment will be chosen first differed from the mean value of no preference at 25%. Corn gluten feed and ENG differed from the mean value while no difference was observed for FLVR and CMIX. Results suggest that animals exhibit a high degree of preference for CGF pellets and that this preference is greater than pellets containing other feed ingredients. Alternatively, cows appeared to exhibit the lowest preference for a high-energy pellet containing mostly corn and wheat middlings.