Department of Animal and Poultry Science, University of Guelph, Kemptville Campus, 830 Prescott Street, Kemptville, ON K0G 1J0, Canada.
J Dairy Sci. 2011 Dec;94(12):6103-10. doi: 10.3168/jds.2011-4589.
Time sampling techniques are useful in collecting feeding behavior data because they minimize the time required for observation. Instantaneous recording is often used in the collection of feeding behavior data for dairy calves; however, the recording intervals used vary widely and it is unclear what minimum interval is necessary to yield accurate data. The objective of this study was to validate data obtained using instantaneous recording of feeding behavior of dairy calves across a range of time intervals with data obtained from continuous recording. Ten Holstein bull calves were observed continuously using time-lapse video for 3 d during the milk-feeding period while they were fed milk ad libitum and for 3 d post-milk-weaning while they consumed solid feed. Feeding behavior data obtained from continuous recording were compared with data obtained from instantaneous recording at intervals ranging from 15 s to 10 min. As expected, the strength of linear association between behavior measures obtained from continuous recording and instantaneous recording decreased with increasing recording interval. The relationship varied between feeding behavior measures; feeding time was represented well (R² >0.76) by instantaneous recording at up to 5-min intervals, but a strong linear association of meal frequency and meal time (R² >0.8) required intervals no greater than 1 min and 30 s, respectively. The relationship between feeding behavior measures obtained from continuous recording and recording at different intervals was similar in both periods; however, sensitivity of time sampling data across recording intervals was greater during the milk-feeding period. Sensitivity was low in both periods (<0.7 with scanning intervals as short as 1 min), indicating that instantaneous recording may not represent within-meal feeding behavior well. Instantaneous recording can provide accurate calf feeding behavior data if the recording interval is sufficiently short.
时间采样技术在收集喂养行为数据时非常有用，因为它可以最大限度地减少观察所需的时间。瞬时记录通常用于收集奶牛犊牛的喂养行为数据；然而，所使用的记录间隔差异很大，不清楚获得准确数据所需的最小间隔是多少。本研究的目的是验证使用瞬时记录在一系列时间间隔内获得的数据与使用连续记录获得的数据在奶牛犊牛喂养行为上的一致性。在牛奶喂养期，10 头荷斯坦公牛犊牛在连续 3 天内通过延时视频进行连续观察，在此期间，它们自由采食牛奶，在断奶后 3 天内，它们食用固体饲料。连续记录获得的喂养行为数据与瞬时记录在 15 秒至 10 分钟的间隔内获得的数据进行了比较。正如预期的那样，连续记录和瞬时记录之间行为测量值的线性关联强度随着记录间隔的增加而降低。这种关系在不同的喂养行为测量值之间有所不同；在长达 5 分钟的间隔内，瞬时记录可以很好地表示进食时间（R²>0.76），但餐频和餐时的强线性关联（R²>0.8）分别需要不超过 1 分钟和 30 秒的间隔。连续记录和不同间隔记录获得的喂养行为测量值之间的关系在两个时期相似；然而，在牛奶喂养期，时间采样数据的灵敏度在不同的记录间隔下更高。在两个时期，灵敏度都很低（扫描间隔短至 1 分钟时，灵敏度<0.7），这表明瞬时记录可能不能很好地代表进食过程中的喂养行为。如果记录间隔足够短，瞬时记录可以提供准确的犊牛喂养行为数据。
