Nursing vocalizations, signalling lactation, were recorded for ten Large White x Landrace sows, parity 1-8 held in the minimal disease piggery at The University of Queensland, Veterinary Science Farm (Australia). The aim was to characterize the individual nursing sounds using Canary 1.2 Software to produce a spectrogram from a recording, to discern auditory features of the nursing grunts that could form the basis of the piglets' ability to discriminate among sows. Sows were significantly different (P < 0.001) from one another for both slow and rapid grunts for five variables MaxF (maximum frequency), MinF (minimum frequency), MainF (the section of the spectrogram with greatest power in dB), Length (duration of the call in s) and Gap (duration of interval between successive grunts). A sixth variable, Volume (maximum volume), was less reliable. Individual sows also showed significant difference in three variables (MaxF, MainF, Volume) measured early (up to 14 days post parturition) and late (more than 14 days post parturition) in the suckling phase for both slow and rapid grunts. It was also found that there were distinct changes over time (P < 0.0001) in the six variables, MaxF, MainF, Volume, MinF (minimum frequency of grunt), Length in seconds (duration of grunt) and Gap in seconds (duration of interval between successive grunts) for the slow vocalizations, when all sows were combined. Also, changes in rapid grunts over time for the combined sows were significantly different in MaxF, MainF and Volume (P < 0.0001). There was strong evidence (using Discriminant Function Analyses) that the variables measured do enable sows to be distinguished on the basis of their vocalizations.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below