Integrative and Comparative Biology

Robotic herding offers a promising off-animal approach to controlling the spatial distribution of grazing livestock. Nevertheless, the habituation of cattle to herding stimuli remains a critical obstacle to its practical use. This study presents the first field investigation of behavioral responses of grazing cattle to herding by a quadruped robot. We compared the herding efficacy of a drone and a quadruped robot, and tracked daily changes in responsiveness during continuous herding. Trials were conducted in a 1.1-ha pasture with Japanese Black breeding cows from June to October 2023. The quadruped robot elicited stronger avoidance responses with shorter latencies than the drone, and herding was more successful with the robot. During consecutive daily herding with the robot, behavioral responsiveness declined progressively over the initial 5 days, at a markedly slower rate than previously reported for drone herding. Following a 24-day interruption, responsiveness partially increased, but declined rapidly again over the subsequent two days. These results indicate that the quadruped robot constitutes a more persistent aversive stimulus than a drone for grazing cattle, although habituation management strategies (such as diversifying stimuli or combining sensory modalities) will be necessary for sustained control of grazing distribution.

Known for their rapid water surface locomotion, whirligig beetles (Gyrinidae) are also fast divers. They can transition from the surface to underwater swimming in under 100 ms. Diving requires them to generate sufficient hydrodynamic torque to break the water surface and reorient the body. However, the mechanism by which whirligigs create the necessary pitch-down torque is unknown. Here, we show that their diving is achieved by bending the abdomen. The inertial hydrodynamic pressure on the flexed abdomen generates the torque necessary for surface breaking and body rotation. The torque is maximized at an abdominal angle of 45{degrees}, whereas for effective diving, whirligigs display a preferential abdominal angle of around 28{degrees}, closely aligning with the angle that maximizes the torque-to-drag ratio. This abdominal angle is optimized by balancing torque generation with the drag penalty. Our findings inspire a mechanism for surface-to-underwater transition, offering design principles for multimodal robots capable of locomotion across the air-water interface.

Developing robust methods to quantify how animals allocate time across behaviours is essential for understanding energy use, habitat requirements, and responses to environmental change. For cryptic, semi-aquatic mammals such as the platypus, direct observation is difficult, creating a reliance on remote biologging approaches that can reliably infer behaviour in the wild. However, aquatic environments can both smooth acceleration signals through hydrodynamic damping and introduce noise from water movement, turbulence, and drag, potentially obscuring behavioural differences of similar magnitudes. We tested whether progressively incorporating biomechanical and frequency-domain (FFT-derived) predictors improved behavioural classification in hydrodynamically challenging aquatic environments. Tri-axial accelerometers were deployed on four ex situ platypuses, with synchronised video observations used to validate behaviour. From the acceleration data, we derived three predictor classes of increasing complexity: summary statistics describing activity level, engineered biomechanical variables capturing posture and body orientation, and FFT-derived features describing movement rhythm. These predictors were progressively incorporated into Random Forest models to classify five behaviours: burrow resting, surface resting, grooming, travelling/foraging, and diving. Model performance improved with increasing predictor complexity, although gains were behaviour specific. FFT-derived features substantially improved classification of rhythmic behaviours such as diving and foraging, while engineered biomechanical predictors improved grooming detection. In contrast, resting behaviours, particularly surface resting, showed little improvement. Overall accuracy increased from [~]75% to [~]88% when frequency-domain features were included. Misclassification was greatest among behaviours with overlapping or low-amplitude signals, and cross-individual validation revealed reduced model generalisability, indicating that individual variation in movement patterns constrained transferability. Incorporating frequency-domain features substantially improved behavioural classification in platypuses, particularly for rhythmic behaviours such as diving and foraging. This study provides the first validated accelerometry-based behavioural classification framework for the species and highlights the importance of matching predictor selection to behavioural mechanics. More broadly, the approach offers a transferable framework for aquatic and semi-aquatic taxa.

When food resources provide reward while simultaneously imposing risk, animals can regulate their behaviour to continue exploiting them while reducing associated costs. Angling presents such a feeding situation, in which obtaining bait can result in hooking and capture. Although fish become less vulnerable to angling following capture experience, the behavioural changes underlying this reduced vulnerability remain poorly understood. Here, we used fine-scale video observations to examine feeding behaviour of red sea bream (Pagrus major) toward angling gear before and after capture experience. Naive fish rapidly swallowed the bait and were subsequently captured, whereas capture-experienced fish continued to approach and feed on the gear but shifted from swallowing to biting. This feeding-mode shift may allow fish to maintain exploitation of the bait while reducing capture risk. Our observations provide a behavioural explanation for reduced angling vulnerability and suggest that direct experience with capture risk can alter how fish exploit a rewarding food resource.

Foraging hematophagous flies such as the stable fly, Stomoxys calcitrans, are shown or expected to sense multiple characteristics of visual host cues. These include relative intensity, spectral composition, and the angle and degree of polarization. Here, we investigated the role of each of these visual cue characteristics on stable fly attraction. In laboratory two-choice bioassays, we offered groups of female or male stable flies a choice between paired photostimuli that differed in their visual characteristics, recording the flies' landing responses on stimuli as the response criterion. Bioassay flies preferentially selected lower-intensity visual stimuli even when paired stimuli markedly differed in spectral composition and polarization characteristics (a high degree of horizontal or vertical polarization). When paired stimuli had identical relative intensity, flies failed to discriminate between the stimulus that presented ultraviolet, blue, green, and red wavebands and a corresponding paired stimulus lacking one of these wavebands. Our results support the hypothesis that foraging stable flies may respond to colour and polarization characteristics of visual stimuli as modulations of perceived intensity. Acquired knowledge on the mechanisms that underlie stable fly visual attraction may inform the design of vision-based traps so that they become optimally attractive to stable flies and other biting flies.

Shorebirds provide an excellent model for investigating the relationship between bill morphology and food acquisition. Food acquisition comprises three successive behavioural stages: foraging (locomotion and prey capture), feeding (food handling and transport), and swallowing. During food transport, these birds use two non-lingual mechanisms, surface-tension transport (ST) and ballistic transport (BT), whose characteristics depends on the kinematics of head and beak movements and the physical properties of the food. We investigated the behavioural flexibility of these transport mechanisms in captive and free-ranging individuals of two species with contrasting beak morphologies: the Pied avocet (Recurvirostra avosetta) and the Black-winged stilt (Himantopus himantopus). Although these species have beaks of a similar size, avocets are distinguished by their upward-curved beaks, whereas stilts have a rather straight beak. We examined the effects of food water content and the presence or absence of water in the beak on food transport by quantifying maximum gape, maximum head displacement, and maximum head velocity. Transport kinematics were jointly influenced by food properties and water availability in the beak. Moist food improved transport performance in both species, whereas dry food required compensatory increases in gape amplitude and head movements, demonstrating that neither ST nor BT constitutes a fixed behavioural sequence. Species also differed consistently in their transport strategies: Black-winged stilts relied on slower, larger-amplitude head movements, whereas Pied Avocets exhibited faster, more precise movements, particularly when water was present. These findings demonstrate that ST and BT share common biomechanical foundations while being governed by rapid kinematic adjustments to changing environmental conditions that probably correspond to flexible motor control. This behavioural flexibility in food transport is therefore likely to enhance feeding performance and ecological resilience in the heterogeneous habitats occupied by shorebirds, suggesting that context-dependent modulation of transport behaviour represents an important adaptive feature of these shorebirds.

Object detection models based on deep learning are increasingly used to reduce the manual workload associated with the processing of video recordings in behavioural research, but reproducible workflows describing how ecologists can build such models themselves remain scarce. Here, we present a step-by-step workflow for constructing a species-specific detection model and apply it as a case study to the red-legged seriema (Cariama cristata), a widely distributed but poorly studied Neotropical bird. Using 60 videos recorded with camera traps as well as non-stop recording cameras at the campus of the Universidade Federal de Vicosa (Florestal, Minas Gerais, Brazil) between 2023 and 2026, we extracted and annotated 480 images to train a YOLO26-nano detection model. On the validation set, the model achieved a precision of 0.934 and a recall of 0.786 (mAP@50 = 0.870), while performance on the independent test set was slightly lower, with a precision of 0.865 and a recall of 0.740 (mAP@50 = 0.740). When applied to unseen footage, the model processed video at roughly 4.6 times real-time speed on a standard CPU and correctly classified all four test videos for the presence or absence of seriemas. Most missed detections involved seriemas that were distant, partially out of frame, or recorded with the lower-quality non-stop cameras. These results showed that a usable detection model can be built from a comparatively small curated dataset, and the accompanying workflow and code are intended to lower the barrier for ecologists seeking to apply object detection to their own study species.

Chemosensory signaling is crucial for organisms to respond to their environment and dysfunctions in these complex pathways have been implicated in diseases such as neurological disorders. The nematode C. elegans is a useful tool for studying chemosensation as it offers a fully-mapped nervous system with relatively simple chemosensory circuits. One of the most common methods to quantify chemosensory function in worms is through chemotaxis assays, which require extensive time and manual effort. In this work, we sought to capture novel chemosensory metrics and increase the throughput of these assays by combining rapid image acquisition with bioinformatics processing. In doing so, we have developed a Technique for Automating chemotaXis Investigations using Simple-Capture Approaches in Nematodes (TAXISCAN), a rapid acquisition platform that can be used to study natural chemosensory behavior as well as to characterize behavioral and locomotion phenotypes in half the time of standard assays. HighlightsO_LIThe TAXISCAN platform provides a rapid, accurate, and inexpensive means of conducting unbiased chemotaxis assays in C. elegans C_LIO_LIIn addition to the Chemotaxis Index (CI), the TAXISCAN platform can quantify behavioral metrics such as movement speed, distance travelled, and dispersion patterns C_LIO_LIWe successfully use TAXISCAN to discriminate behavioral phenotypes in neurosensory mutants C_LI

Octopuses exhibit remarkable regenerative capacity, as shown by their ability to fully restore a whole arm. During this process, a complex appendage containing muscle, connective tissue, vasculature, suckers, sensory structures, peripheral nerves and the axial nerve cord (ANC) is regrown. Although the morphology of octopus regeneration is increasingly well described, the physiological variables that determine regenerative speed, including the potential contribution of ion-channel-dependent processes, remain poorly understood. In this work we examined the influence of reproductive state, injury severity, biological sex and magnesium chloride (MgCl2) anesthesia, on the rate of longitudinal arm regrowth. Reproductive stage was the strongest determinant of regenerative performance: pre-reproductive stage (pre-RS) animals showed robust regrowth, whereas reproductive stage (RS) animals regenerated significantly slower. In pre-RS animals, regeneration speed scaled strongly with injury severity, whereas this relationship was absent in RS animals. Biological sex did not independently predict regeneration speed, and MgCl2 exposure showed no long-term inhibitory effect on regeneration. Our final model explained 81.4% of variance in regeneration speed, with reproductive state and injury severity having major effects. These findings demonstrate that sexual maturation and optic-gland-mediated senescence act as an irreversible physiological block against somatic regeneration, and provide empirical support for mechanosensory and bioelectric-dependent scaling models of cephalopod limb restoration.

Flying insects process motion across the visual scene to regulate their ground speed, correct course deviations, and behaviorally relevant stimuli such as avoiding predators or finding hosts. The yellow-fever mosquito Aedes aegypti provides a useful model for studying these visually guided behaviors, as both sexes use vision to track moving features. Approaching a moving target presents mosquitoes with a demanding visual problem: they must extract target motion from the changing visual scene while simultaneously controlling their own flight and responding to potentially dangerous movements. To understand this, we first studied free-flight responses to static and moving vertical targets in a wind tunnel system capable of projecting dynamic visual stimuli, and found that males and females are sexually dimorphic with respect to their responses to target motion. When the feature was static, both males and females advanced toward it during flight; however, when it oscillated laterally at 2 Hz, only females continued to pursue it. In a subsequent yaw-unrestrained magnetic tethering paradigm, we showed that this divergence cannot be simply explained by sex-specific differences in visuomotor tuning dynamics, while micro-CT reconstructions revealed only modest sexual differences in eye morphology and broadly similar patterns of estimated spatial acuity. Together, these results indicate that males retain the biomechanical capacity to track oscillating targets but actively suppress approach behavior during free flight, whereas females pursue salient features independently of target motion. This pattern aligns with sex-divergent visual ecologies because the risk of approaching a large moving feature too closely likely outweighs the benefits for males seeking mates, whereas for host-seeking females the reproductive necessity of blood feeding overrides this avoidance.

Phenotypic plasticity exemplifies how environmental signals can shape organismal development, yet the cellular mechanisms translating ecological cues into adaptive morphologies remain incompletely characterized. Many species of Daphnia (freshwater crustaceans) develop a diverse array of inducible defenses in response to predator chemical cues (kairomones), providing a tractable system for examining how ecological pressures are transduced through developmental mechanisms. Daphnia lumholtzi produce elongated head and tailspines in response to kairomones. How this occurs at the cellular level, through changes in cell size, proliferation, or both, is currently unknown. To address this question, we quantified the temporal dynamics of cell proliferation in D. lumholtzi across 72 hours following kairomone exposure using EdU incorporation (marking proliferating cells) and DAPI staining (to quantify total nuclei). Predator cue exposure induced a three-phase proliferative response: initiation within 24 hours (slightly increased proliferation and total cells), a transitional plateau at 48 hours (minimal effects in both treatments), and commitment by 72 hours (strong increases in both proliferation and cell accumulation). Headspines exhibited higher proliferation than tailspines, suggesting anterior- posterior developmental prioritization. Treated animals maintained smaller average cell sizes throughout the response, consistent with continuous addition of newly divided cells rather than cell enlargement. Unlike the delayed-division strategy in D. longicephala or bilayer formation in D. pulex, D. lumholtzi employs sustained hyperplasia, demonstrating that the same selective pressure produces similar ecological outcomes through mechanistically distinct developmental programs. Our findings bridge ecological signals with cellular responses, exemplifying eco-evo- devo integration whereby environmental pressures (predation) are transduced through developmental processes (cell proliferation dynamics) to generate adaptive and heritable morphological diversity across ecological and evolutionary timescales.

Predators pursuing agile prey must coordinate locomotion with rapidly changing visual information, yet how they control head orientation during high-speed terrestrial pursuit remains poorly understood. We reconstructed three-dimensional head, torso, and lure target trajectories from synchronised multi-camera recordings of freely running cheetahs (Acinonyx jubatus) using model-based markerless reconstruction, and propagated reconstruction uncertainty through Monte Carlo analysis. Across 1,373 valid reconstructed pursuit frames, head pitch was more closely aligned with the target than torso pitch, and is tightly stabilised relative to target elevation. Head-on-torso counter-rotation nearly cancelled concurrent torso-pitch rotations, preserving target-relative head pitch despite locomotor oscillations. In yaw, cheetahs predominantly oriented their heads ahead of the target's instantaneous position while maintaining a distinct locomotor heading and continuing to close range. During steady-lateral pursuit, head yaw was best described by a short horizon constant-velocity extrapolation model with a cross-validated median prediction horizon of 45 ms. Together, these findings reveal complementary head orientation strategies during pursuit: robust sagittal stabilisation and predictive lateral tracking keep the head target-referenced while the torso generates a closing trajectory towards intersection.

When sensory modalities yield conflicting information, animals must rapidly reassess stimuli to select their actions. We induced auditory-visual conflict in free-flying echolocating Egyptian fruit bats, by fitting animals with prisms that shifted the perceived visual location of a landing perch while echoes returned from its veridical location. Bats that course-corrected within a single goal-directed flight did so by decoupling sonar gaze from steering, to enable rapid reweighting of visual and auditory cues. We designed artificial agents that used Bayesian inference to construct estimates of goal locations in their environment. When competing estimates directed active-sensing behaviors distinctly from steering, agents course-corrected more rapidly. Consistent with this idea, when bats were fit with prisms and earplugs that attenuated auditory localization cues, they were unable to course-correct. Removing prisms produced no systematic after-effects. Our framework suggests that instead of correcting their behavior after failure, animals could efficiently employ active sensing to resolve sensory conflict before failure occurs. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=178 SRC="FIGDIR/small/743555v1_ufig1.gif" ALT="Figure 1"> View larger version (41K): org.highwire.dtl.DTLVardef@16262edorg.highwire.dtl.DTLVardef@4cd82aorg.highwire.dtl.DTLVardef@103e6fforg.highwire.dtl.DTLVardef@13292bf_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOCover figure.C_FLOATNO Bat wearing helmet with clear-glasses and tracking markers. Photograph (C) 2026 Nikita M. Finger / Moss Laboratory. C_FIG

Cruising speed is a key factor affecting prey-search efficiency and migration range in continuously swimming animals. Tunas and lamnid sharks (e.g., white sharks) have convergently evolved traits for high-speed cruising, including the ability to maintain slow-twitch, aerobic red muscle (RM) warmer than ambient water, known as RM endothermy. Despite their well-known high cruising speeds, kinematic features underlying their elevated speeds remain unclear. Swim speed is the product of tailbeat frequency (TBF; Hz) and stride length (SL, the absolute distance traveled per tailbeat; m). RM endothermy is expected to elevate TBF by enhancing muscle contraction performance. Furthermore, within RM-endothermic fishes, tunas and lamnid sharks may exhibit distinct kinematic features because of differences in caudal fin morphology and tailbeat amplitude. Here, we compiled kinematic parameters from 20 fish species, including five RM-endothermic species, measured in the wild using animal-borne sensors. Comparative analyses showed that, for a given body mass and water temperature, RM-endothermic fishes exhibited 1.9 times higher cruising speed and TBF than ectothermic fishes, while SL remained similar. Within RM-endothermic fishes, tunas exhibited 2.3 times higher TBF than similar-sized lamnid sharks, whereas lamnid sharks showed 1.7 times longer SL than similar-sized tunas. These results indicate that RM endothermy is generally associated with higher TBF, while significant kinematic differences remain between tunas and lamnid sharks. This divergence may be partly explained by the greater caudal fin area and tailbeat amplitude in lamnid sharks. It may also reflect contrasting skeletal types of teleosts and elasmobranchs, which potentially influence body stiffness and swimming kinematics.

Spiders arrest their fall by abseiling on dragline silk, varying silk release rate to avoid excessive strains that could break the thread. To understand how this is achieved, we develop a minimal dynamical model of abseiling and formulate a time-optimal control problem subject to physical limits on silk strain and dragline spinning rate. We find that the optimal trajectory includes an interval over which the strain is constant, resulting in uniform deceleration consistent with published experiments. To test the generality of this strategy, we then turn to a distinct extreme motor behavior seen in jumping spiders: using a dragline anchored at the takeoff site to correct their orientation while jumping at speeds exceeding 100 body lengths per second. We show that the same constant-strain strategy suffices to counteract the initial body rotation and ensure an upright landing without significant loss of range. Together, our theoretical framework suggests how strain regulation may serve as a strategy for controlling extreme movements such as abseiling seen in many spiders and may be specialized to tasks such as aerial braking and reorientation in jumping spiders.

Gliding enables mammals to forage and escape from predators by moving between discontinuous forests. Its benefits depend not only on glide distance but also on the ability to decelerate and land safely. This study developed a theoretical framework linking glide distance, gliding velocity, aerodynamic braking, body mass, and braking distance. Twenty-two representative distance-velocity observations from eight studies and five species were compiled. Among distance-velocity models, log-distance and saturated with V0 models received nearly equivalent support. Both models predicted increasing velocity with glide distance, with the rate of increase declining at longer distances. For a 1 kg animal undergoing a 60% reduction in velocity, predicted kinetic energy remaining immediately before contact increased from 4.80-5.14J at 20 m to 8.13-8.90 J at 80 m. This velocity reduction corresponded to a dissipation of 84% of approach kinetic energy before contact. Over a braking distance of 1 m, the required mean deceleration increased from 2.57-2.75 g at 20 m to 4.35-4.77 g at 80 m. At 80 m, shortening the braking distance from 4 to 0.5 m increased the required deceleration from 1.09-1.19 to 8.70-9.53 g. These results indicate that the absolute energetic and deceleration requirements of landing increase with glide distance, even when velocity is close to an asymptote. These results provide a quantitative basis for considering aerodynamic braking and landing requirements alongside conventional measures of glide performance. Summary statementModels quantify how glide distance, aerodynamic braking, and braking distance affect pre-contact kinetic energy and deceleration requirements in gliding mammals.

The Siamese fighting fish (Betta splendens) is well-known for its high levels of aggression and complex, stereotyped displays, making it an ideal model for testing deep learning-based tools for high-throughput, unbiased quantification of aggressive behaviour. A deep-learning object recognition model (BettaAI) was developed to detect the species' aggressive displays, including threat (opercula opening and fin distension), attack (bites) behaviours, as well as surface air breathing, a behaviour correlated with the frequency and intensity of aggression. Videos of males confronting their mirror image were acquired simultaneously with synchronised top and side cameras. Two separate models were trained, with some behaviours identified from top-view frames and others from side-view frames; outputs from both models were combined and post-processed for final quantification. BettaAI performed with high accuracy: F1 scores were 0.99 for the top-view model and 0.93 for the side-view model. Average Precision (AP) at 0.5 IoU exceeded 0.98 for all categories except bites, a rare behaviour in the dataset, which reached 0.865. Comparisons with a manually curated "ground-truth" dataset from videos of 10 males showed strong concordance between automated quantification and manual annotations by expert observers across all behaviours. A case study on 461 males reacting either to a mirror or to a video playback of a conspecific demonstrated the high-throughput and unbiased advantages of BettaAI, with threat displays being more frequent toward the mirror image and attack rates not differing between treatments. This work underscores the potential of BettaAI for automated, bias-free quantification of aggressive displays in B. splendens.

Animal movements can transport materials across ecosystem boundaries, while simultaneously modifying habitats through physical engineering. We integrated GPS telemetry, camera trapping, sediment measurements and physicochemical sampling of ponds on Santa Cruz Island, Galapagos, to evaluate how giant tortoises connect and shape terrestrial and freshwater ecosystems. Telemetry revealed 2,254 visits by 21 tortoises to 12 ponds, with visits lasting up to 21.38 h and showing strong seasonal and diel variation, while camera traps showed peak pond use in January and during early morning and late afternoon. Greater tortoise abundance was associated with higher ammonium, dissolved organic carbon and total phosphorus concentrations, while ponds with persistently greater tortoise abundance had higher total suspended solids. Departing tortoises carried a mean 505.9 g of sediment, yielding a conservative potential export of 8.20 kg per tortoise per year and tens of tonnes annually. These findings identify giant tortoises as mobile zoogeochemical engineers linking terrestrial and freshwater ecosystems through physical and biogeochemical pathways.

As foraging increases exposure to predators, individuals must allocate time to vigilance in addition to food acquisition. Although those behaviors are assumed to be mutually exclusive, this is often hard to disentangle, given that multiple sensory modalities are often applied. However, when foraging, animals often have one sensory modality more impaired (e.g. vision), than others (e.g. auditory); thus, a level of compromise on vigilance capabilities is assumed while foraging. Here, we test the idea that when a food item requires manipulation, thus reducing vigilance capabilities, social foraging should promote shared costs of vigilance. We conducted video recordings to investigate whether there is a trade-off between foraging and vigilance behavior in red-legged seriemas (Cariama cristata) when foraging in pairs. We studied the social foraging behavioral patterns while seriemas were feeding on macauba palm (Acrocomia aculeata) fruit, which requires a series of throwing maneuvers to crack the hard outer shell. Our results showed that within a pair recording one individual was observed to spend more time vigilant than their partner, while the other individual spent more time throwing fruit than their partner. However, we found no evidence for synchronization or coordination of vigilance and foraging between individuals within a feeding bout. These results provide insight into the foraging ecology of red-legged seriemas and the trade-off between foraging and vigilance in social contexts.

1. Organisms rely on environmental cues, such as moonlight illumination and weather, to modulate nocturnal behaviors that reduce predation risk and conserve energy during demanding periods. Urbanization can alter the reliability of cues, potentially impairing organisms' ability to respond appropriately with consequences in increasingly urbanized landscapes. 2. We leveraged the Motus Wildlife Tracking System to identify roost initiation, roost departure and nocturnal restlessness of dark-eyed juncos (Junco hyemalis) overwintering along an urban to rural gradient to determine if activity patterns are affected by weather, moonlight illumination, and urbanization. 3. Nocturnal behavior of overwintering juncos varied with environmental factors, including moonlight illumination and temperature. However, the strength of these responses differed among sites, suggesting that natural cues of moonlight illumination were masked in urban areas, where individuals were more responsive to temperature. 4. Our results suggest that urbanization weakens birds' responsiveness to natural light cues, likely because artificial lighting reduces the visibility of moonlight; representing a key mechanism by which human-altered landscapes can disrupt nocturnal ecology in birds.