Abstract #W14

# W14
Continuous recording versus scan-sampling in behavioral studies with growing heifers fed high-concentrate diets.
Ana Madruga*1, Alfred Ferret1, María Rodríguez1, Eva Mainau1, Jose Luis Ruiz de la Torre1, Luciano Gonzalez2, Xavier Manteca1, 1Animal Nutrition and Welfare Service (SNIBA), Universitat Autonòma de Barcelona, Bellaterra, Spain, 2Centre for Carbon, Water and Food, The University of Sydney, Camden, NSW, Australia.

Continuous observation is an accurate method for behavioral measurements in ruminants but is very time-consuming. An alternative is scan-sampling (SC), in which behavior is recorded at regular intervals. The aim of this study was to compare the loss of information when different SC intervals were chosen to record heifer behavior. Behavior of 8 growing heifers fed a high-concentrate diet was recorded on 6 non-consecutive days. Continuous sampling (CS) was considered the control treatment and was used to evaluate scan-sampling intervals of 2, (SC2), 5 (SC5), 10 (SC10), 15 (SC15), 20 (SC20) and 30 (SC30) min. Two additional sampling techniques were considered: a time sampling of the first 10 min of every hour (TS), and the continuous observations during 4 h post-feeding (4PF). Behaviors were expressed as a percentage of the total daily activity. A square root-arcsine transformation was applied to achieve normal distribution. Differences were analyzed by using the GLIMMIX procedure of SAS. The model contained the fixed effect of treatment, and the random effects of heifer and day. Pearson correlations were used to compare CS with SC treatments. Daily time in each behavioral activity using the SC and TS treatment did not differ from CS (P > 0.10). In contrast, 4PF affected behavioral measurements. Time spent ruminating, social behavior and self-grooming were 50, 41 and 38% lower (P < 0.001), while eating and resting times were 28 and 17% higher (P < 0.001) in 4PF than CS. For all behaviors, high correlations were observed between CS and SC2 (r > 0.93; P < 0.001) and between CS and SC5 (r > 0.75; P < 0.001). For eating, drinking, social, self-grooming, tongue-rolling, licking and biting fixtures, and rummaging in wood shavings, the longer the scan interval, the lower was the correlation. For ruminating and resting, correlations between CS and SC30 were greater than 0.90 (P < 0.001). In conclusion, the use of SC or the TS treatment did not result in any loss of information. However, 4PF is not an accurate technique for measuring behavioral activities in ruminants.

Key Words: beef cattle, behavior, scan-sampling method