RESEARCH ARTICLE Facial display and blushing: Means of visual communication in blue-and-yellow macaws (Ara Ararauna)? Aline Bertin1., Arielle Beraud12, Léa Lansade’, Marie-Claire Blache , Amandine Diot, Baptiste Mulot?, Cécile Arnould 1 PRC, CNRS, IFCE, INRA, Universihé de Tours, Nouzilly, France, 2 ZooParc de Beauval & Beauval Nature, Saint-Aignan, France aline bertin @inra.fr Abstract Mainly recognized for their cognitive performance, the visual communication system and, particularly, the potential function of facial displays in parrots remain thus far unexplored Here, we provide the first descriptive study of facial display use in captive blue-and-yellow macaws. We observed the feather position (sleeked or ruffled) on the crown, nape and cheek at the group level during the macaws’ daily routine and individually while interacting with a familiar animal caretaker. In the latter context, blushing was also assessed on the bare skin of the cheek. Group level observations showed that crown, nape and cheek feath- ers ruffling was more frequent in activities requiring no locomotion than in activities requiring locomotion. With the animal caretaker, crown ruffing was significantly more frequent when the caretaker was actively engaging with the parrot than during a control phase with no mutual interaction. In addition, a significantly higher proportion of naive observers judged blushing as being present on photographs taken during the mutual interaction phase than during the control phase. We thus showed significant variations in facial displays and bare skin colour based on the birds’ social context and activity. Our results broaden the scope for further studies to determine whether parrots’ faces provide visual social signals. determined. SUeH Es hilcators of positive emotions, remain to be Materials and methods Birds and housing conditions We observed parc de Beauval Saint Aignan (41110, France). All birds are part of a free-flying show. The birds had been trained daily (ie, handled daily) since weaning and were thus in close contact with humans, especially their caretakers. All the birds had fully adult-like plumage but were not yet sexually mature and did not express sexual or defensive behaviours. The five birds were housed in two adjacent aviaries of cm) freely connected to an outdoor area (250 cm x 850 cm x 260 cm)、Three of the macaws were housed in aviary 1 (Table 1). The other two were housed in aviary 2 with a red macaw (Ara macao) and a yellow-crested cockatoo (Cacatua sulphurea citrinocristata). The aviaries were equipped with several tree branches, perches and ropes. Enrichment was provided daily (cardboard and journal paper). Parrots were fed daily with fresh fruits and vegetables, germi- nated seeds (wheat, corn, sunflower, rice, and oat), millet seeds, oyster shells, and a commercial mix for exotic birds five hand-reared blue-and-yellow macaws, not exposed to the public, at the Zoo- Behaviour and feather postures Video-recording and image analysis. To determine during which activity feathers were rufled on the crown, cheek and nape, we used a focal sampling method with a hand held cam- era recordersony HDRP P410) capturing 24 images per second. We followed one focal bird’s behaviour for twenty minutes, and then successively followed another of the five birds Each twenty-minute session was repeated (on different days) in a random order until we obtained two hours of recording per bird. As birds were trained for free flight daily in the morning, we observed the birds in the afternoon between 2and S PM. The experimenter was familiar to them and moved around the aviary when necessary with the minimum of distur bance possible. We used a scan sampling method to analyse the videos Every 5 s, the experimenter recorded the bird’s hehaviour following the behavioural repertoire presented in Tablc2 and simulta neously, the feather position following the repertoire presented in (3i ). From preliminary observations, we determined three areas where feathers can move independently from one another: the crown (composed of green feathers), the nape (composed of blue yellow and black feathers) and the cheek (Wy .D. For each area. we determined two feather positions; sleeked or sleeked, continuous black lines can be observed, and whern were equipped with several tree branches, perches and ropes Enrichment was provided daily (cardboard and journal paper). Parrots were fed daily with fresh fruits and vegetables, germi- nated seeds (wheat, corn, sunflower, rice, and oat), millet seeds, oyster shells, and a commercial mix for exotic birds. Behaviour and feather postures Video-recording and image analysis. To determine during which activity feathers were ruffled on the crown, cheek and era recorder (Sony HDRP P1410) capturing 24 images per second. We followed one focal bird’s behaviour for twenty minutes, and then successively followed another of the five birds. Each twenty-minute session was repeated (on different days) in a random order until we obtained two hours of recording per bird. As birds were trained for free flight daily in the morning, we observed the birds in the afternoon between 2 and 5 PM. The experimenter was familiar to them and moved around the aviary when necessary with the minimum of distur- bance possible. nape, we used a focal sampling method with a hand held cam- We used a scan sampling method to analyse the videos. Every 5 s, the experimenter recorded the bird’s behaviour following the behavioural repertoire presented in Table 2 and simulta- neously, the feather position following the repertoire presented in (Fig 1). From preliminary observations, we determined three areas where feathers can move independently from one another: the crown (composed of green feathers), the nape (composed o blue, yellow and black feathers) and the cheek (Fig 1). For each area, we determined two feather positions: sleeked or ruffled. When the cheek feathers are sleeked, continuous black lines can be observed, and when they are not distinguish individual feathers or see their tips. When ruffled, individual feathers and their tips can be distinguished. When some parameters were unobservable, the scan was deleted. We the lines appear discontinuous. For the crown and nape, when sleeked, one can- Table 1. Age, sex and location of the 5 captive macaws used for the study. Age (years) Sex Aviary Individuals Aruba Antonio Imotep Gédeon Petry female male female ONE Table 2 Alimentation the bird is ground-foraging and eating, drinking the bird is Maint with the beak or the feet) the bird is immobile obtained a mean total number of 1327.8 37.2 scans per individual (minimum: 1198, maxi- mum: 1407).TThe same experimenter conducted all observations. We assessed the observer reli- ability by rescoring 20 min of video per bird. The percentage of intra-observer accordance was 90.5396. The observations were conducted in April and May 2017 with temperatures ranging from 15’C to 22 C Feather postures and blushing when interacting with the animal caretaker Protocol. We aimed to determine whether variations in feather position and skin colour could be observed when interacting with the most familiar caretaker. The caretaker has trained birds for free flight daily for months and was thus very familiar to the birds (during free-fight training the birds were taught to come to hand when called after unrestricted outdoor flight). For each session, each bird was taken by hand at random by the caretaker and transported a few meters away to an empty aviary (250 cm x 520 cm x 260 cm) equipped with a familiar Feather postures and blushing when interacting with the animal caretaker Protocol. We aimed to determine whether variations in feather position and skin colour could be observed when interacting with the most familiar caretaker. The caretaker has trained birds for free-flight daily for months and was thus very familiar to the birds (during free-flight training the birds were taught to come to hand when called after unrestricted outdoor flight) For each session, each bird was taken by hand at random by the caretaker and transported a few meters away to an em (250 cm x 520 cm x 260 cm) equipped with a familiar h and a camera recorder mounted on a tripod (Sony HDRP P4I0), All birds were used to being handled for daily training and weighing. Each bird was then placed on the perch in front of the animal caretaker. Each session was composed of two 2-minute phases Crown feathers Nape feathers Cheek feathers Sleeked Ruffled Fig i. Repertoire of head feather displays Photographic repeesentation of the poition of crown, nape and check teathers (Altgraphs taken by A. Beraud) Phase 1: Mutual interaction. During this phase, the animal caretaker actively interacted with the bird by looking at and talking to them (no scratch or massage). Phase 2 Control In a control context, we observed each bird in the presence of the animal caretaker but with no mutual interaction. During this phase, the animal caretaker remained at the same distance but turned her back to the bird. We considered this phase as a less positive valence for the birds than phase 1. This phase also allowed observing whether the birds expressed “seeking behaviour” towards the animal caretaker (the bird tried to grasp the care- taker’s clothes with its beak or claws, bent on its side and looked at the caretaker, or jumped on the caretaker). We hypothesized that if the mutual interaction was positive for the birds, they would more positively seek attention from the animal caretaker to restore the active inter- action during the control phase than during phase 1. At the end, the bird was transported back to its aviary We repeated the entire procedure (phases 1 and 2) ten times (10 sessions) per bird. All ses- sions were conducted in the afternoon on two successive weeks with only one session per day We did not counterbalance the order of the two phases because in prior trials, the birds were unwilling to stay on the perch for the first two minutes if not actively engaged. Image analysis. Each phase (phase 1: mutual interaction, phase 2 control) was video recorded with a Sony HDRR PI410 capturing 24 images per second. Positive seeking behav ours were assessed on videos for each bird and phase With the same protocol as described in the first part of the study, we used a scan sampling method every 5 s on the video to assess the feather position (ruffled or sleeked) based on the repertoire (Eig D. We obtained a mean t SE total number of 213.60 : 4.91 scans per individual (minimum: 205; maximum: 227) during the control period and a mean total number of 211.20 9.70 scans per individual (minimum: 175; maximum: 228) during the mutual interaction. We also took one close-up photograph of the bird’s left profile at the end of each phase (20 photographs per bird) with a Nikon D3100. To assess the presence or absence of blushing on the bare white cheek skin, 100 images were taken with 20 images per bird (1 per phase x 10 ses- sions). Three images were discarded due to blurring (all with the caretaker’s back to the bird) A panel of four naive observers visually assessed the presence or absence of blushing on the ring of bare skin around the eye (Eis.2). This zone was chosen from preliminary observations when we observed that blushing was not diffused homogeneously throughout the entire bare skin patch surface. This area around the eye blushed the most frequently. The percentage of inter-observers accordance was 88.15%. the bare white cheek skin, 100 images were taken with 20 images per bird (1 per phase x 10 ses- sions). Three images were discarded due to blurring (all with the caretaker’s back to the bird). A panel of four naive observers visually assessed the presence or absence of blushing on the ring of bare skin around the eye (Fig 2). This zone was chosen from preliminary observations when we observed that blushing was not diffused homogeneously throughout the entire bare skin patch surface. This area around the eye blushed the most frequently. The percentage of inter-observers accordance was 88.15%. Fig 2 Examples of photographs. A) Target zone used to estimate the naive observers, B) Photograph where all observers judged blushing to be present in the target zone. C) Photograph where no observers judged blushing to be present in the target zone. (All photographs taken by A. Beraud). or absence of blushing by a panel of 4 EI https dol.01 10 137ournat pone 0201762 August 22, 2018 5/12