Fachbereich Veterinärmedizin



    Technical note:
    evaluation of data loggers for measuring lying behavior in dairy calves (2013)

    Zeitschriftenartikel / wissenschaftlicher Beitrag
    Bonk, S (WE 19)
    Burfeind, O (WE 19)
    Suthar, V S (WE 19)
    Heuwieser, W (WE 19)
    Journal of Dairy Science; 96(5) — S. 3265–3271
    ISSN: 0022-0302
    URL (Volltext): http://edocs.fu-berlin.de/docs/receive/FUDOCS_document_000000019642
    DOI: 10.3168/jds.2012-6003
    Pubmed: 23498014
    Tierklinik für Fortpflanzung

    Königsweg 65
    Haus 27
    14163 Berlin
    +49 30 838 62618

    Abstract / Zusammenfassung

    Lying behavior might indicate how the animal interacts with its environment and is an important indicator of cow and calf comfort. Measuring behavior can be time consuming; therefore, behavioral recording with the help of loggers has become common. Recently, the Hobo Pendant G data logger (Onset Computer Corp., Bourne, MA) was validated for measuring lying behavior in cows but no work to date has validated this logger for measuring lying behavior in calves. The objective of this study was to test the accuracy of the Hobo Pendant G data logger for measuring total lying time and frequency of lying bouts in dairy calves. In 2 experiments (experiment 1: thirty-seven 2-h observation periods; experiment 2: nineteen 24-h observation periods), we tested the effect of 2 different recording intervals, the effect of attachment to different legs, and the effect of removing short, potentially erroneous readings. We found an excellent relationship when comparing the 30-s and 60-s recording intervals. For total lying time and bout frequency, the highest correlation was found when the logger was attached to the hind legs and recording was conducted with a 60-s sampling interval. In experiment 2, average total lying time was 1,077 ± 54 min/24 h (18.0 ± 0.9h/24h), with an average frequency of 19.4 ± 4.5 bouts per day. Predictability, sensitivity, and specificity for experiment 2 were >97% using the 60-s recording interval and removing single readings of lying or standing from the data set compared with direct observation as reference. The data logger accurately measured total lying time and bout frequency when the sampling interval was ≤ 60 s and short readings of lying and standing up to 1 min were converted into the preceding behavior. The best results were achieved by attaching the logger to the right hind leg.