**4. Discussion**

Using Qualitative Behavioural Assessment, observers from a range of backgrounds and with different levels of sheep experience distinguished behavioural expression of sheep subjected to different physical conditions during commercial transport (e.g., trailer crate type and deck level), between sheep of different breeds and even between sheep transported from a livestock saleyard compared with those from a feedlot. Our previous studies under experimental transport conditions showed behavioural expression reflected changes in the underlying physiological state of the livestock that were typically associated with the stress response [15–18]. Such behavioural responses are likely to predict the wellbeing of the animals, and as indicators of welfare they may be associated with production traits [24,25]. Importantly, the early identification of animal responses to husbandry procedures (e.g., farm/facility/transport setting) allows the modification of procedures or facilities in a way that optimises the health and welfare of animals [26,27]. Even subtle differences in demeanour can therefore be informative.

#### *4.1. Study A1: Trailer Crate Design*

The observers detected differences in merino sheep travelling in the upper deck of two types of trailer crates over the same road journey. Sheep in the convertible crate were scored as more *agitated*/*anxious* than sheep in the standard crate, which were more *calm*/*relaxed*. Hence, this dimension corresponds to a distinction between emotional valence of sheep, with similar terms used to those in our previous studies using a small, single-deck trailer [15,16]. Prior to this experiment, sheep had been handled in a similar way; they were yarded 2–3 h before loading, had minimal food curfew and were moved through the same loading facility.

Although the two crate designs had similar loading density, the larger pen size in the convertible crate allowed the animals the freedom to pack tightly together in the corners, leaving others more room to move. This may have allowed more physical movement and less sheep-to-sheep support within the whole pen; increased space availability has been demonstrated to increase movement in lambs, but also increase slipping and carcass bruising [28,29]. Structural differences in the sides and the floor of the upper decks of the two crates may have influenced how sheep travelled. The slight reduction of head room caused by the side rails on the standard crate may decrease physical movement or create a more confined feeling for the sheep. The floor of the upper deck of the convertible crate is less stable as it allows greater suspension and vibration than that of the standard crate. One possible explanation is that this increased flexibility of the floor may lead to sheep experiencing less secure footing.

#### *4.2. Study A2: Sheep Breed*

Observers detected differences in merino and fat-tail sheep travelling in the same type of trailer crate (a standard crate) over the same road journey. Merino sheep were scored as more *calm*/*relaxed* and *curious*/*alert* compared with the Fat-tail sheep on the same vehicle and route, which were scored as more *agitated*/*distressed* and *tired*/*content*. These dimensions appear to reflect activity and/or arousal. However, given that we only had five replicate transport events for each sheep breed group, our data warrants following up with additional studies on how different sheep breeds deal with road transport.

Prior to this experiment, sheep had been handled in a similar way (yarded 2–3 h earlier, with no feed curfew and moved through the same loading facility). Thus, the behavioural response to transport appeared to be linked to breed. The behavioural responses may reflect the ability of the sheep to cope with the transport stressors and this may be linked to their state of fearfulness. Previous studies have demonstrated physiological differences (plasma cortisol, packed cell volume) in sheep genotypes, with some more reactive to land transport than others [30]. In a non-transport example, Miller et al. [31] demonstrated that QBA scores were different for feral goats exposed to varying levels of human interaction, suggesting that the scores reflected habituation and possibly their ability to adapt to the challenge of confinement. Grandin [32] found that animals with an excitable temperament may have greater difficulty adapting to a situation, while calmer animals may adapt more easily and become less stressed. Identification and selection of breeds of sheep that better cope with the transport environment represents one way to improve the welfare of livestock.

#### *4.3. Study A3: Point of Origin*

We only found subtle differences in the demeanour of sheep that originated from either a saleyard or feedlot—the only differences were on GPA dimension 3, which only accounted for 5.8% of the total variation in observer scores. Merino sheep being transported from a saleyard were more *nervous*/*curious* than those transported from a feedlot, which were more *comfortable*/*relaxed*. We had expected that there would be differences in demeanour for these animals because sheep had been exposed to very different experiences over the preceding 24–48 h. Prior to yarding for 2–3 h before loading, sheep from the feedlot had been fed in a pre-export feedlot for approximately 5 days. These sheep were penned in large flocks and most likely habituated to the feed, shed and daily routine; they also had minimal time off food and water prior to transport. By contrast, sheep loaded at the saleyard had experienced a novel, noisy and high arousal environment over the previous 1–3 days. Potential stressors for saleyard sheep include a transport journey from their farm of origin, feed curfew of 24–72 h, separation and mixing of social groups, handling through several small yards by sheep dogs, as well being confronted by the presence of humans at close proximity. The subtle differences in demeanour could indicate that the transport environment is sufficiently novel and challenging to mostly overshadow prior affective condition. Further studies comparing more diverse backgrounds, such as from farm of origin and comparing varied lengths of time in confinement or curfews could provide further insight into the effect of point of origin.

#### *4.4. Study B: Deck Level*

We recorded a significant difference in the demeanour of sheep positioned in the upper and lower decks on for both vehicle crate types, with sheep on the upper deck scored as more *alert*/*curious* and *agitated*/*cramped* than those on the lower deck, which were scored as more *miserable*/*unsure*, *dejected*/*weary*, *nervous*/*worried* and *distressed*/*stressed*. The final study that allowed observers to compare sheep expression directly from journeys from all four positions, that is standard upper and lower and convertible upper and lower, confirmed the significant difference was detected between deck levels but not crate type, indicating that effect of deck level overshadowed any difference between crate design. Interestingly, the term *nervous* was highly loaded on the low arm of the GPA3 for all but one analyses but the term *cramped* was only loaded on GPA3 analysis of CU-CL. Although the third dimension accounts for a small percentage of variation, it seems sheep in the upper deck were viewed as more cramped than the lower.

How important this deck position is to sheep is unknown and may be influenced by other factors such as weather and climate. Sheep positioned on the upper deck are likely to have more visual and tactile stimulation as they experience wind, sun or rain, which may affect how they experience the journey whereas those in lower decks are largely screened from such exposure. It has been demonstrated that many novel aspects associated with transportation, apart from the vehicle movement itself, can trigger behavioural responses in sheep [33]. The behavioural dimensions identified in this study concur with key qualitative descriptors for sheep demeanor that were identified in previous studies, such as *agitated*, *distressed*, *alert* and *curious* [13,15,16,34]. These dimensions used terms semantically consistent with a valence of 'mood' (GPA 1) and 'arousal' (GPA2) that have been similarly reported elsewhere [13,27,31,34].

#### *4.5. Limitations of This Study*

Because we were working under commercial conditions and made every attempt to not interfere with normal transport processes, we had relatively little control over the conditions that animals were transported under. For example, we had only few transport events for fat-tail sheep that were transported under matched conditions as merino sheep, which restricted our sample size for this treatment comparison. Furthermore, in order to restrict the number of variables between transport events, clips were taken from routine transport events that matched each other in as many ways as possible.

The convertible crate design is relatively new, and only a few livestock trucks of such design in Western Australia. Multiple transport events on specific vehicles (driven by its owner/driver) were therefore not strictly independent. We tested for such an effect in our data (driver included as a random factor) and identified significant effects of vehicle driver on the demeanour of sheep, which could reflect either their driving style or the specific design of their vehicle.

Finally, although we attempted to obscure evidence of the animals being on a truck by editing our video clips, we cannot guarantee that observers were unaware of the treatment difference for our deck level experiment. Obtaining footage from the lower decks with adequate lighting and adequate visualisation of groups of sheep with limited head room was challenging. The contrast with sheep in sunshine on the upper deck could not be avoided. Such cues are likely to influence how observers score footage because QBA scoring is sensitive to environmental clues which can introduce undesirable bias due to the observers' judgment of that context [35,36]. We used different cameras to account for different lighting levels on the different decks and therefore the quality of the footage was equitable, but animals on the lower decks are more obviously in a confined place than those on the top deck.
