Abstract: We compared the size of skeletal and ornamental traits, and asymmetries in bilateral skeletal traits, between victims of cannibalism, cannibals and bystanders within small groups of caged female White Leghorns at the time of cannibalistic attacks (i.e. injurious pecks resulting in bleeding). We hypothesised that victims of cannibalism have discernible morphological traits that predispose them to cannibalistic attack. We predicted that victims would have smaller skeletal traits (body length, ulna length, metatarsus length and width, toe length), lower body weight, poorer body condition, smaller combs and more asymmetrical bilateral skeletal traits than their flock mates. Contrary to our prediction, victims of cannibalistic attacks to the head/neck area (N=23) tended to have larger combs than their flock mates (Wilcoxon matched-pairs signed-ranks test, S=59, P=0.037, NS after sequential Bonferroni adjustment). Their cannibals were more asymmetrical than non-cannibalistic bystanders (metatarsus length, S=48, P=0.011 and composite asymmetry, S=62.5, P=0.002, significant after sequential Bonferroni adjustment). In agreement with our prediction, victims of cannibalistic attacks to other body parts (N=27), including the back, wings, rump, tail, cloaca, abdomen and toes, were more asymmetrical (composite asymmetry, S=78, P=0.022, significant after sequential Bonferroni adjustment) and tended to have lower body weights (S=79.5, P=0.029, NS after sequential Bonferroni adjustment) than their flock mates. Their cannibals did not differ in skeletal or ornamental traits from the non-participating bystanders. The results suggest that large combs either elicit attacks to the head and neck area or increase vulnerability to injury during such attacks. Attacks to other body parts appear to be directed towards birds with signs of weakness relative to their flock mates. In these attacks, there were no distinguishing features separating cannibals from bystanders, suggesting that the bystanders could all be potential cannibals.

Abstract: Abstract Considering the ever-growing demand of various breeding organizations for an objective, inexpensive, reliable, and easily conducted assessment of the behavior of horses, the aim of our study was to implement a novel-object test and a startling test into any kind of breeding performance testing to assess horses' temperament. Additionally, the influence of testing areas (familiar or unfamiliar), riders, and horse factors such as levels of training, breed, and age were of interest. Furthermore, recommendations for the practical implementation concerning the parameters should be given. Therefore, 1,028 horses over a period of 3 years participated in a temperament test consisting of 5 different stimuli. The horses were either ridden (61.8 %) or led by hand (38.2 %) by an unfamiliar professional rider (N = 43) or a familiar rider (N = 20). Live behavioral observations were taken by a trained observer. Overall, horses' scores for reactivity in the present temperament test were distributed over the whole scale, with lower means and higher standard deviations (6.7 ± 2.2-7.6 ± 2.1) than corresponding scores from the conventional personality evaluation in performance tests (7.7 ± 0.8-8.2 ± 0.5; P < 0.01). High correlations (r = 0.3-0.9; P < 0.001) between the scores for reactivity and the other behavioral parameters (emotional expression, activity, time to calm down, rider's aids) show a large influence of these parameters in assessing the horses' temperament. Factors like breed type, sex, and age had significant influences (P < 0.001) on different scores of the temperament test. In most cases, the rider or handler had no influence on the different scores assessed during the temperament test. The training level and the testing modus never had a significant influence on different scores. Only the testing station or location had a small influence on the scores for the stimulus “bridge” in some horses. Based on the results, it could be concluded that an implementation of a temperament tests into performance testing is possible during various types of testing procedure. Especially the assessment of reactivity, emotional expression, interest in the stimulus and rider's aids during and after passing the stimulus, as well as the time to calm down are important parameters for analyzing the horses' personality.