The objectives of the first study were to evaluate the effects of direct-fed microbials (DFM) with or without monensin plus tylosin on the performance, feeding behavior patterns and feed efficiency of steers transitioned to a high-concentrate diet. Crossbred steers were randomly assigned to 1 of 4 treatments (n = 32) in a 2 x 2 factorial design, with Factor 1 being diets with DFM (25 g/d; Natur's Way) or without DFM (CON), and Factor 2 being diets with monensin (40 g/ton) and tylosin (10 g/ton; MON) or without MON (CON). During the grower/transition period, MON-fed steers had 9.5% higher (P < 0.05) ADG, and improved F:G (7.8 vs 8.94; P = 0.06) and RFI (-0.28 vs 0.27 kg/d; P < 0.01) vs those not supplemented MON. DFM-fed steers had lower (P < 0.01) ADG than non DFM-fed steers and tended (P < 0.09) to have a higher RFI. Daily variances of bunk-visit and meal event frequencies were reduced (P < 0.01) in MON-fed steers vs those fed diets without MON during the grower/transition period. There were MON x DFM interactions (P < 0.10) for ADG and F:G during the finisher period. Numerically, steers had a 4.7% lower F:G when the MON diet without DFM was fed, whereas, F:G was higher in steers fed the MON diet containing DFM. During the finisher period, steers fed MON diets had lower (P < 0.01) DMI and lower RFI (P < 0.01; -0.23 vs 0.23 kg/d). Conversely, DFM-fed steers had higher RFI (P < 0.01) compared to respective controls. The steers fed the MON diets consumed 6.7% less DMI, spent 9% more (P < 0.05) time consuming meals, and thus had 14% slower (P < 0.01) meal eating rates than steers fed diets without monensin. The objectives of the second study were to evaluate and validate the sensitivities, specificities and accuracies of an ultra-wideband radio frequency identification system (RFID) system to measure the frequency and duration of bunk visit (BV) events and brush usage (BU) events. For the first trial, 4 algorithms were developed to quantify the frequency and duration of BV events by recording the timestamps for the start and end of each BV event. A virtual line was established that ran parallel to the edge of the feed bunk, defined at the 0-cm line. Additional virtual lines were established that ran parallel with the 0-cm virtual line at 0, 5, 10, and 15 cm from the edge of the feed bunk to define the end of a BV event. The algorithms were compared to determine the most accurate algorithm for recording BV events. The 0-10 cm in-and-out algorithm had the highest r2 values of 0.81 and 0.91 for frequency and duration of BV events, respectively, with the overall accuracy of BV presence and absence being 85.9%. For the second trial, 4 algorithms were developed to quantify frequency and duration of BU events by recording the timestamps for the start and end of each BU event. The 4 algorithms evaluated had x-y dimensions of 90 x 90, 120 x 120, 150 x 150 and 180 x 180 cm surrounding the brush. The 150 x 150 cm algorithm had the highest r2 values of 0.73 and 0.79 of frequency and duration of BU events, respectively, with the overall accuracy of BU presence and absence being 76.1%. The ultra-wideband RFID geolocation system was able to accurately measure BV events; More research is warranted to determine if the accuracy of the system for BU usage can be improved.
