INTERPRETATION OF BEHAVIORAL REACTIONS OF SHEEP TOWARDS FEAR-ELICITING SITUATIONS

The relationships between 27 behavioural variables, displayed by 206 s heep in three tests used to measure fear in sheep, were assessed using Principal Component Analysis, These animals were individually subject ed to the same standardised fear-eliciting situations: isolation from conspecifics (I), a surprise effect (S) and the presence of a human (H ). The first component (which explained 30.1% (I), 29.4% (S) and 34.8% (H) of the total variance) was interpreted as anxiety or fear induced by social isolation which is common to the three tests. Effectively, eight variables related to feeding, locomotor activity and localisatio n in the test pen showed consistent and high loadings on component 1 f or the three tests. In order to simplify the methodology, a 'fear mark ' was computed using the three major items related to component 1 for the three tests: the time spent feeding, the time in square 5 (locatio n of the trough) and in squares 1, 2 and 3 (away from the stimulus and the trough). This method of scoring allows positioning of an animal o n a continuous scale and therefore direct comparisons with other group s of animals, using a simplified methodology. Other behavioural variab les showed high loadings on component 2 in S and in H: this component was interpreted in terms of 'close interaction' with the stimulus in S and in terms of 'panic' in H. An 'interaction mark' (S) and a 'panic mark' (H) were computed using the major behavioural items. These marks could improve the interpretation of behavioural reactions of sheep to wards a surprise effect and the presence of an immobile human. (C) 199 8 Elsevier Science B.V. All rights reserved.