Route Formation and the Choreography of Looking Back in Desert Ants (Melophorus bagoti)

Route Formation and the Choreography of Looking Back in Desert Ants (Melophorus bagoti) | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Route Formation and the Choreography of Looking Back in Desert Ants ( Melophorus bagoti ) Cody A Freas, Ken Cheng This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4670516/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract How ants, wasps and bees rapidly form visual routes represents an enduring mystery as well as a powerful example of the abilites of insect brains. Here, we analyse a previously uncharcterised behaviour, ‘lookbacks’, underlies rapid bi-directional route learning in desert ants. During these lookbacks, foragers stop forward movement to their goal location, turn and fixate their gaze to their origin, often for only 150–200ms. This turn appears to be a critical period for learning the inbound route. Route formation relies on acquiring visual cues and comparing panoramic view memories with the current view. While the nest panorama is learned during pre-foraging learning walks, during which naïve ants often fixate their gaze at the nest, route following requires separate behaviours to learn route based views. We untangle how route formation occurs in naïve Melophorus bagoti foragers during the first foraging trips by focusing on the previously uncharacterised lookback behaviours and their function in facilitating visual learning. Lookbacks were highly associated with the first few foraging trips and were concentrated in areas where the visual scene changed rapidly. Analysis of gaze directions during lookbacks show foragers clearly fixate their view to the nest direction during these behaviours (or alternatively to the feeder during inbound homing), learning the nest-aligned inbound route during their first outbound trips. We discuss lookbacks as a ‘when to learn signal’ combining visual rotation and gaze fixations to produce view-based route following. Animal Behavior Cognitive Neuroscience view comparisons pirouettes route following latent learning scanning bouts turnbacks Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Desert ants are model systems for visual navigation as they are able to efficiently navigate both to known food sites and back to the nest, rapidly forming routes between locations (Wehner et al. 1996 ; Mangan and Webb 2012; Collett et al. 2013 ; Cheng et al. 2014 ; Freas and Cheng 2022 ). As diurnal scavengers, these foragers are active during the hottest parts of the day when prolonged periods outside the nest can be fatal, thus making efficient homing and rapid spatial memory formation critical to an individual’s survival. To rapidly home to goal locations, desert ants rely on a shared navigational toolbox comprising multiple strategies (Wehner et al. 1996 , 2003; Collett et al. 2013 ; Cheng et al. 2014 ; Cheng and Freas, 2015 ; Freas and Schultheiss, 2018 ; Buehlmann et al. 2020) The two most well understood of these strategies are the maintenance of path-integration-derived vectors (Wehner, 1982 ; Collett and Collett, 2000 ; Wehner and Srinivasan, 2003 ) and learning panoramic views of the terrestrial scene, both around the nest and along known foraging routes (Wehner and Räber, 1979 ; Wehner 2003 ; Collett et al. 2006; Harris et al. 2007 ; Cheng et al. 2009; Schultheiss et al. 2016 ; Freas and Cheng 2017 , 2019 ; Freas et al. 2019 ; Collett and Hempel de Ibarra 2023). Route formation via view comparison has been extensively modelled (Zeil et al. 2003 ; Wystrach et al. 2011 ; Baddeley et al. 2011 , 2012 ; Kodzhabashev and Mangan 2015 ; Möller 2012 ; Le Möel and Wystrach 2020 , Murray et al. 2020 ). Within these models, foragers turn and travel towards attractive views by comparing the current view with memories acquired while facing goal locations, based on the current motivational context. Nest-associated view memories involve reinforcement learning of inbound views, which promotes forward movement and leads foragers back to the nest site, and is likely reinforced upon the forager’s arrival (Le Möel and Wystrach 2020 , Murray et al. 2020 ; Wystrach et al. 2020a ; Freas et al. 2022 ). Given that visual scenes can change between trips, view memories are continuously updated, and their associated valence modulated in response to scene changes (Le Möel and Wystrach 2020 ; Wystrach et al. 2020a ; Islam et al. 2020 ; Freas et al. 2022 ; Deeti et al. 2023 ; Lionetti et al. 2023 ). Yet, less well understood is how these initial view memories of foraging routes are acquired given that multiple species of desert ant are known to only need a single experience of the outbound route to acquire the necessary visual information to correctly return home (Freas and Cheng 2017 , 2018 ; Freas and Spetch 2019 ). Prior to foraging, each naïve ant, newly emerged from the nest entrance, must learn the most important visual scene, the nest’s. This is accomplished through multiple highly structured learning walks, which have been well characterised within the Cataglyphis genus (Fleischmann et al. 2016 , 2017 , 2018 , Grob et al. 2017 ; Freas et al. 2019 ; Collett and Hempel de Ibarra 2023). Within these walks, Cataglyphis noda and Cataglyphis aenescens (both species which nest in visually cluttered environments) exhibit scanning behaviours with distinct pausing or fixation phases, including what has been termed ‘pirouettes’ (Fleischmann et al. 2017 ). During these behaviours, the ant rotates in place either fully 360° or partially with a rotational reversal (left to right or right to left). Interspersed throughout this rotation are a number of fixations, where the ant neither moves forward nor rotates its gaze direction (Grob et al. 2017 ; Fleischmann et al. 2017 , 2018 ; for M. bagoti pirouettes see Fig. 1 B). Interestingly, the longest of these gaze fixations take place when the ant is oriented to the nest entrance, suggesting this period is when the ant learns the nest-aligned view memory (Fleischmann et al. 2017 ). Similar fixation behaviours have been recorded in the learning flights of bumble bees and wasps; thus, this may be a consistent view learning mechanism across Hymenoptera (Robert et al. 2018 ; Collett and Hempel de Ibarra 2023; Collett et al. 2023 ). Pirouettes, comprising both rotational movement and nest-oriented fixations (Fleischmann et al. 2017 ; Freas et al. 2019 ; Fig. 1 A) are highly associated with view acquisition and view memory formation (Fisher et al. 2022). Currently, a detailed choreography of pirouettes in Australian desert ants, Melophorus bagoti , is lacking (for learning walk structure see Deeti and Cheng 2021), though high-speed recordings suggest that naïve individuals perform similar oscillating scans during their learning walks that include both nestward gaze fixations as well as fixations away from the nest direction (for an example of these scans, see Supplemental Materials Video 1, Fig. 1 A). After performing these learning walks, ants leave the immediate area around the nest to search for food (Wehner et al. 2004; Fleischmann et al. 2016 ). Yet, visual learning continue beyond learning walks, as inexperienced foragers move into unknown scenes beyond the nest area (Freas and Cheng 2019 ). As an individual moves away from known nest views into unfamiliar areas, this translational movement causes the visual scene to change while maintaining a degree of similarity with the acquired nest views (Zeil et al. 2014). This decay in similarity occurs in a predictable fashion, with the highest degree of similarity between the ant’s current view and their memorized views occurring when they are facing the direction of their starting location (Nicholson et al. 1999; Zeil 2012 ; Zeil et al. 2014). Each view memory has a ‘catchment area’, or area around the view location where there is sufficient visual similarity between previously acquired view memories and the current view to usefully orient, though this is highly dependent on the amount of local visual clutter, given the propensity for proximal visual cues to both change rapidly and block distant visual cues during translational movement (Zeil et al. 2003 , 2014; Strüzl and Zeil 2007; Murray and Zeil 2017; Jayatilaka et al. 2018 ; Freas and Cheng 2019 ; Freas et al. 2021 ). Beyond this catchment area, new route-based views must be learned to guide the forager back along this path to the nest. Foraging ants often form routes that can extend up to hundreds of metres, necessitating the acquisition of multiple views along the foraging route and at profitable food sites for successful navigation and route formation (Mangan and Webb 2012; Graham and Cheng 2009; Schultheiss et al. 2016 ; Freas et al. 2017, 2018; Jayatilaka et al. 2018 ; Freas and Spetch 2019 , 2023 ). Similar to outbound learning in desert ants, wood ants ( Formica rufa ) are also known to exhibit bi-directional route learning during their first few trips to the feeder. Inbound views are thought to be learned during short reversal periods when the ant stops traveling to the feeder, turns back and travels a short distance back to the nest, though these path reversals do not feature any nestward fixations (Graham and Collett 2006). Wood ants are known to fixate at visual landmarks near a newly discovered feeder (Nicholson et al. 1999) and such fixations are thought to have a similar function to pirouette behaviours observed in Cataglyphis ants performing learning walks. During outbound route formation, new ant foragers are often observed performing scanning behaviours in which the forager turns back towards the nest (Fig. 1 ), a behaviour called ‘lookbacks’ (Zeil 2012 ; Zeil et al. 2014; Freas et al. 2019 ). During these scans, forward movement ceases and the ants turn in place, orienting their gaze yet often not their body (Fig. 1 B) back towards the nest, learning the nest-aligned view and inbound route during these rotations interspersed with fixations (Fig. 1 B-D). These nest-oriented gaze fixations during outbound travel align with the rapid visual route learning observed in desert ants ( Melophorus bagoti and Cataglyphis velox ), with foragers needing only a single outbound route experience (Freas and Cheng 2017 , 2018 ; Freas and Spetch 2019 ). Yet the choreography of these lookback-scans and their potential function in route formation is lacking. Here, we characterise the first foraging trips of M. bagoti foragers as individuals learn to travel between a feeder site and the nest, focusing on route formation, scanning and lookbacks to explore how foragers’ visual route learning forms and how lookbacks may aid in encoding goal-oriented views to produce rapid visual learning. Methods Study Site and Species The field site (23°45′28.12″S, 133°52′59.77″E) was located on the grounds of the Centre for Appropriate Technology campus, ~10 km south of Alice Springs, Northern Territory, Australia. Melophorus bagoti inhabits visually cluttered areas with vegetation (buffel grass and scattered Eucalyptus ). Experiments were conducted during the Australian summer (November through February) in 2022–2023. Foraging arena A plastic feeder (15 × 15 × 10 cm) was sunk into the ground 7m from the nest entrance and baited with both cookie pieces (Arnott™) and mealworm pieces. The smooth walls of the feeder prohibited escape after food collection, allowing experimenters to manually start the inbound journey by lifting foragers to the feeder ledge. Between the feeder and nest entrance, all vegetation was cleared, and a 15-cm-high plastic wall was erected, forming a ~3m wide arena (Figure 2A,B). This arena extended 5m, where a 15cm high barrier prevented further movement to the feeder except through a funnel comprising two (61cm H × 91.5cm L) white sheets of Masonite board with a 30cm gap between them (Figure 2C,E). This greatly reduced the amount of natural food resources the nest could access, motivating foragers to search the area until finding the feeder (Figure 2D,E). Experiment 1: Route Formation All foragers emerging from the nest entrance were marked on the gaster with enamel paint (Tamiya™) for five consecutive days, denoting them as experienced individuals and allowing us to exclude them from testing. On the sixth day, when non-painted naïve individuals first left the nest to complete a pre-foraging learning walk, they were collected and individually marked using a colour combination of paint (Tamiya™) on their gaster and/or thorax. Marked individuals were observed during subsequent learning walks and upon their first foraging trip. The end of learning walks and onset of foraging was determined by the distinct lack of a learning walk with alternating nestward/outward scanning behaviours. Melophorus bagoti exhibits highly structured learning walk scans (akin to pirouettes in Cataglyphis ), rhythmically alternating scanning bouts towards and away from the nest with clear nestward and outward fixations (See Supplemental Materials Video 1, Figure 1). Upon starting foraging, their outbound foraging paths to collect food and inbound returns to the nest were recorded via pencil and graph paper. We additionally marked the location along the path when foragers performed scanning behaviours and lookbacks. After the forager collected food, we also recorded the inbound path from the feeder back to the nest, marking the location of inbound scans as well as lookbacks where the forager turned back to the feeder direction. We defined scanning behaviour as the stopping of forward movement accompanied by rotational body movement on the spot (interrupted by fixations), while remaining oriented outward away from the nest. Lookbacks are structurally somewhat similar to scans (in that they entail a stopping of forward movement and rotational movement interspersed with fixations); we differentiated these from scanning rotations by constricting a ‘lookback’ to only occasions when rotational movement brought the individual’s orientation to within ±90° of the nest (during an outbound lookback) or feeder direction (during an inbound lookback). For each individual, we collected positional data of the path as well as any scans and lookbacks observed on the first 10 foraging trips conducted on the first day of foraging. To determine if route formation metrics changed after the overnight delay period when the foraging force is not active, we continued to collect these metrics on each forager’s second day of foraging for the first five trips, also noting if each forager conducted any pre-foraging re-learning walks around the nest prior to second day foraging. After completing this 5 th next-day trip, ants were marked with paint as tested. Before testing, we determined that ten foraging trips would be sufficient for a forager to become highly experienced of the route (Freas and Cheng 2018; Freas and Spetch 2019). If a forager conducted less than ten foraging trips on the first day, we continued to collect this individual’s metrics on subsequent days until they reached ten total foraging trips and then recorded their first five foraging trips on the following morning (four individuals took two days to complete their first 10 foraging trips). If a forager continued to forage after ten day-one trips, we continued to collect these paths but given these paths were highly similar to Trip 10 by every metric, they have not been used in our analysis. Route Formation Data Analysis Pathswere digitized bymarkingevery ~10cm along the path length while both scans and lookbacks counts were recorded where they occurred along the path . We assessed three metrics of path formation for both the outbound and inbound portions of the foraging route: 1.) path straightness , denoting the efficiency of the outbound/inbound journey between the nest and feeder; 2.) outbound/inbound Scan counts, associated with navigational uncertainty; and 3.) outbound/inbound Lookback counts, associated with route formation and learning. Path straightness was calculated by the ratio between the straight-line distance (nest-feeder) with the forager’s path length. All metrics were compared across trips over two days (fixed effects) using poisson loglinear General Linear Mixed Models (GLMMs) for count data and Gaussian loglinear GLMs for path straightness data with individuals as a random effect. When significant effects were uncovered, a priori within-individual contrasts akin to Helmert contrasts (with α set at 0.01 to correct for multiple non-independent comparisons) were conducted between each trip and the mean of the following trips that day (i.e. Trip 1 vs Trips 2-10; Trip 2 vs Trips 3-10, etc.). When the comparisons became insignificant, we denoted this as the within-day performance asymptote. Contrasts analysing the overnight performance decline were assessed by comparing Trip 1 of the next day with the average performance on the final three trips of the first day (Trips 8-10). Next-day performance comparisons were conducted with Helmert-style contrasts in a parallel fashion to the first foraging day (Next Day Trip 1 vs. Next Day Trips 2-5, etc.), with performance asymptote similarly determined (when p > 0.01). Additionally, within individual comparisons between performance on the outbound and inbound routes of the same trip were compared using Wilcoxon Signed Rank Tests. We separately analysed the effect of the overnight delay on route formation and maintenance by comparing the mean change (Δ) within each individual in each of the three path-based metrics separated into the outbound/inbound routes (path straightness, scans and lookbacks) between trips characterised by whether they followed a trip that occurred on the Same Day (Trips 2–10 and Next Day Trips 2–5) or after the ~16h Overnight Delay (Next Day Trip 1). For the four ants which took two days to complete the first ten trips, their first trip of day two was averaged with the Next Day Trip 1 (foraging day three for these individuals). Within-individual comparisons between Same Day and Overnight Delay trips were compared using Wilcoxon Tests. Experiment 1: Lookback Positions Image analysis of route To assess potential mechanisms underlying when foragers decide to lookback, we chose to characterise how the visual scene changed as foragers moved between the nest and feeder, looking at three factors: ‘local’ panorama change, ‘global’ panorama change, and the ant’s current global vector length (which is not based on views). We quantified the rate of local panorama change by assessing the panorama change over the last 50cm of the route and more globally by comparing visual scenes across the arena to the panorama at the nest entrance. To accomplish this, we collected 360° panoramic images of the foraging route for image analysis. These images were taken at the nest entrance and at 50cm intervals in a grid-like fashion along the arena leading up to and including the feeder panorama, totalling 72 images throughout the foraging route (Figure 2B-D, 6A). To calculate a rate panoramic change along the foraging route, we compared each panorama image within the arena with both the nest image for a metric of how the panorama has changed globally from the beginning of the foraging trip (nest), as well as comparing each image to the previous image 50cm before along the grid’s straight-line route to the nest. This gives an indication of how much the visual scene has changed recently as the ant travels to the feeder. As a final metric of change, we also calculated the vector distance (cm) from the nest for each image site. Outbound lookback counts were assigned to the closest image site within the arena (within 25cm). This allowed us to assess the panorama characteristics at sites associated with the number of lookback behaviours which occurred in the vicinity throughout the first 15 trips of the foragers’ careers. Lookbacks that occurred within 25cm of the nest entrance were excluded from this analysis as we could not assess the local or global panorama changes under this distance from the start. To calculate the amount of visual change between each of the image pairs, rotational image difference functions (rotIDFs) were conducted by using the mean pixel difference (MPD) between the focal image and comparison images for all possible rotations of the test images (in one-degree steps) using custom written scripts in MATLAB (for further details, see Zeil et al. 2003, 2014; Stürzl and Zeil 2007). Each image was down sampled to 1 pixel per degree with final images measuring 360 pixels width × 180 pixels height and converted to the blue colour channel only. When two images contain some degree of familiarity, this produces a rotIDF that contains a clear valley of mismatch or low Mean Pixel Difference (MPD) between the two panoramas highly associated with the direction of translation between the initial image and the image of the updated location, indicating a level of similarity the navigator can use to orient. To calculate a similarity metric corresponding to the degree to which the two panoramas were similar, we took the difference between the minimum and mean MPD of each rotIDF comparison. This created a ‘panorama change’ metric where high values corresponded with a deep valley of mismatch between panoramas, denoting a large degree of similarity and low levels of panorama change across space. In contrast, low values corresponded with a shallow valley of mismatch between panoramas in the rotIDF, denoting a low degree of similarity between the images and more rapid panorama change over space. A multiple linear regression analysis was conducted to examine the effect of the Local Panorama Similarity (last 50cm), Global Panorama Similarity and Vector Length on the location of Lookbacks during the outbound portion of the paths. Experiment 2: Lookbacks Choreography After observing individuals during the route formation experiment, we chose two locations along the route at which to record lookback behaviours for gaze direction analysis; the first ~80cm from the nest entrance and at the funnel area at 5–6m from the nest, given these two locations were associated with a high number of lookback counts during route formation testing. At these sites we spread white sand over the red soil over a 1m x 1m area in order to more clearly observe the ant’s body/head positions on video frames. Above each site we positioned a Sony Handycam 70cm above the ground, with the lense facing down (3840 x 2160, 25fps). A forager which had already been to the feeder at least once generally left the nest in the feeder directional quadrant where the camera’s FOV was positioned. As these foragers left the nest they were recorded until they left the camera frame, capturing any lookback behaviours occurring in this area. Once they left the filming area, each forager was collected and marked as tested to prevent refilming on subsequent trips. Ants’ gaze directions and positional estimates were manually extracted from the videos using SLEAP animal positioning software (Pereira et al. 2022). Two points along the ant were selected to assess gaze direction during lookback behaviours, the centre of the mouth/front of the head and the head base, where the head and body meet. There two points allowed us to assess gaze direction of the animal and directional and positional fixation phases beyond the body’s orientation, essential given the large degree of body contraction we observed during lookbacks (~55° divergence between gaze and body orientations, See Figure 1A fixation 6). We began tracking the animal’s pose estimates five frames before the ceasing of forward movement associated with the start of the lookback scan and pose estimates were stopped after the ant resumed forward travel to the goal (feeder/nest) for five frames. When assessing the structure of lookback scans, we collected the number, direction and duration of fixations throughout the lookback, determining the longest fixation duration, and its direction in relation both to the direction of travel prior to stopping and the nest/feeder direction. We additionally chose to analyse the direction at which the inflection or reversal point of the turn occurs (when the individual switches from turning left to right). Lookback Gaze Analysis Directional gaze-fixation data was analysed using circular statistics. To determine whether the longest gaze fixation and the reversal direction were directed, we first used Rayleigh tests for circular data (Fisher 2993). If headings were directed, we further analysed whether the mean direction of the longest gaze fixation was in the nest/feeder direction and the opposite (180°) of travel using the 95% confidence interval (CI) around the mean of longest gaze fixations. 95% CIs were calculated through the standard error of the mean heading direction based on the mean vector length (Fisher 2993). Within-individual directional comparisons between longest gaze fixation and reversal direction were conducted using Moore’s Paired Tests. Comparisons between the two outbound conditions, Near Nest, Funnel and the Inbound lookback conditions were conducted using poisson loglinear General Linear Models (GLM) for count data, with individuals as a random effect. If there was a significant effect of condition, Dunn-Bonferroni Post-hoc pairwise comparisons were conducted comparing conditions. Pairwise comparisons within conditions (absolute angular error comparisons and pre vs. post reversal comparisons) were compared using Wilcoxon tests. Results Route Formation Naïve foragers travelling away from the nest area for the first time exhibited highly tortuous outbound paths to the feeder (path straightness μ ± Std. = 0.30 ± 0.07; Figure 3, 4A) interrupted by multiple bouts of scanning (μ ± Std. = 11.07 ± 5.39; Figure 3, 4C) and a large number of lookbacks in the direction of the nest (μ ± Std. = 7.47 ± 0.29; Figure 3, 4E), consistent with high uncertainty beyond the nest area, until each individual found the feeder and collected food (Figure 3, 4A). In contrast, the inbound portion of Trip 1, returning to the nest with food, was significantly (Wilcoxon Signed Rank Test; Z = 4.01; p < 0.001) straighter (path straightness μ ± Std. = 0.88 ± 0.06), with significantly fewer scanning bouts (μ ± Std. = 3.20 ± 2.01; Wilcoxon Test; Z = –3.3; p = 0.001) and lookbacks to the feeder compared to the outbound portion of the same trip (μ ± Std. = 0.33 ± 0.49; Wilcoxon Test; Z = –3.4; p = 0.001: Figure 3, 3). Inbound lookbacks were remarkably rare, with only 28 observed instances across all 225 inbound routes (μ ± Std. = 0.12 ± 1.55; Figure 3, 4F). The pattern of significantly lower path straightness coupled with higher counts of both scanning bouts and lookbacks during the outbound portion (vs. inbound) persisted through the first six trips, with path straightness becoming not significantly different from its inbound counterpart at Trip 7 (Wilcoxon Test; Z = –1.93; p = 0.053; Figure 4AB), scanning bouts non-significant at Trip 9 ( Z = –0.21; p = 0.831; Figure 4CD), and lookbacks non-significant at Trip 8 ( Z = –1.29; p = 0.197; Figure 4EF). At the beginning of their second day of foraging, a majority of individuals (53%) conducted a number of ‘re-learning walks’ around the nest before beginning to forage and these walks appeared similar to those occurring prior to initial route formation. Of the eight foragers which conducted re-learning walks, four conducted one learning walk while another four conducted two walks prior to restarting foraging. Once foraging behaviour resumed on the next day, the same pattern of significantly lower outbound (vs. inbound) straightness and higher lookback counts was observed on Next Day Trip 1 and 2 ( p 0.05) on Next Day Trip 3 (Scan count was significantly higher on outbound journeys until Next Day Trip 4; Figure 4). Outbound route Both Trip and Delay had significant effects on outbound path straightness (Trip: χ 2 = 21.88; p < 0.001; Delay: χ 2 = 15.39; p < 0.001), scans (Trip: χ 2 = 24.38; p < 0.001; Delay: χ 2 = 21.76; p < 0.001) and lookbacks (Trip: χ 2 = 28.59; p < 0.001; Delay: χ2 = 37.91; p < 0.001) and there was only a significant interaction between Trip & Delay ( χ2 = 8.08; p < 0.01) on scan count. Helmert contrast comparisons showed that outbound route performance reached asymptote (Table S1) on Trip 3 (scan count), Trip 5 (lookback count) and Trip 6 (path straightness). There was a significant decrease in performance across all metrics ( p < 0.001; Table S1) after the overnight delay (Trips 8-10 vs. Next Day Trip 1). During next day foraging, Helmert contrast comparisons showed that outbound route performance returned to asymptote by Next Day Trip 3 in all metrics. Inbound route Trip ( χ 2 = 10.74; p = 0.001) but not overnight Delay ( χ 2 = 2.85; p = 0.09) had a significant effect on forager inbound path straightness, while both Trip ( χ 2 = 17.67; p < 0.001) and Delay ( χ 2 = 5.45; p = 0.02) had significant effects on scan counts and there was not a significant interaction between factors ( p = 0.540). There was no effect of Trip ( χ 2 = 2.28; p = 0.131) and Delay ( χ 2 = 0.23; p = 0.634) on inbound lookback counts. Helmert contrast comparisons showed that inbound route performance was already at asymptote (Table S1) on Trip 1 for path straightness ( p = 0.022) but scan count showed significant contrasts until Trip 3, after which no further improvements were found. Additionally, scan count did fall below asymptote during Trip 5 ( p = 0.002). There was no significant decrease in performance after the overnight delay (Trips 8-10 vs. Next Day Trip 1). During Next Day foraging, inbound route performance improved and returned to asymptote by Next Day Trip 2 (Table S1). Overnight delay An overnight delay period had a clear effect on forager performance. When comparing the mean change within individuals between all Trips which followed a same-day trip and those which followed the overnight delay, these delays corresponded with significant decreases in outbound path straightness (Wilcoxon-Test; Z = – 3.41; p = 0.001; Figure 5A) and significant increases in both outbound lookback ( p = 0.001; Figure 5B) and scan ( p = 0.001; Figure 5C) counts. A similar pattern emerged with inbound path characteristics with significant differences in mean performance on trips after same-day trips or after overnight delay trips, with significant decreases in path straightness (Wilcoxon Test; Z = –3.41; p < 0.001; Figure 5A) and significant increases in scan counts (Wilcoxon Test; –3.41; p < 0.001; Figure 5B). Lookback Spatial Positions Local Panorama Similarity (last 50cm), Global Panorama Similarity and Vector Length explained 11.92% of lookback variance across the arena and this effect was significant ( F =3.02, p = .035, R 2 = 0.12). Only Local Panorama Similarity was shown to impact the number of lookbacks which occurred spatially within the arena, with an inverse relationship of lower panorama similarity over the last 50cm being associated with a higher number of lookbacks occurring in this area ( p = 0.005). Both Global Panorama Similarity ( p = 0.234)and Vector Length ( p = 0.245)were not significantly associated with lookback numbers. Lookback chorography Three types of lookbacks were observed and recorded during testing. The first and most commonly observed was defined as the 180° turn in place (Figure 1A,C). Here, foragers cease forward movement and turn back towards the nest direction before reversing their turn direction. In a minority (~7%) of lookbacks, this reversal does not occur, and foragers continue to turn in a circle before continuing forward movement. We have defined these lookbacks as 360° circular turns (Figure 1D), and given their lack of reversals, they were not included in any pre- and post-reversal fixation direction and duration analysis. The final type of lookback which was observed consisted of the forager hooking back to the nest direction, with the turn accompanied by spurts of forward movement during the turn, typically back in the direction of the nest (see Figure 1E). All of these lookbacks consisted of multiple fixations scattered throughout the turn. The longest fixation period during each lookback in both outbound conditions (nest: μ ± Std. = 189 ± 82ms; funnel: μ ± Std. = 210 ± 107ms), was towards the nest entrance. The longest of these fixations were significantly directed (nest: Rayleigh’s Test, Z = 20.63; p < 0.001; funnel: Rayleigh’s Test, Z = 14.78; p < 0.001; Figure 7A) and the nest direction (0°) was within the 95%CI of both nest (μ ± 95%CI = 8.12° ± 9.84°) and funnel (μ ± 95%CI = 8.12° ± 9.84°) lookbacks. Reversal directions in both conditions were also significantly directed (nest: Rayleigh’s Test, Z = 19.20; p < 0.001; funnel: Rayleigh’s Test, Z = 12.74; p < 0.001; Figure 7A) with the nest direction (0°) within the 95%CI (nest: μ ± 95%CI = 5.71° ± 9.94°; funnel: μ ± 95%CI = 5.02° ± 16.14°). Inbound lookbacks mirrored outbound lookbacks with a longest gaze fixation period (μ ± Std. = 182 ± 120ms) that was towards the feeder direction. These gaze directions were significantly directed (Rayleigh’s Test, Z = 5.02; p = 0.004; Figure 7A) and the feeder direction (180°) was within the 95%CI (μ ± 95%CI = 202.56° ± 34.01°). The longest fixation direction and the reversal direction aligned in 51% (26 of 51) of analysed lookbacks and differed on average by ~10° across all conditions (Near Nest: 10.03°; Funnel: 10.32°; Inbound: 11.67°). Within-individual comparisons between longest gaze fixation and the reversal direction did not significantly differ in all three conditions (Moore’s Paired Tests; Near Nest: R’ = 0.278 ; p > 0.50; Funnel: R’ = 0.219 ; p > 0.50; Inbound: R’ = 0.897 ; p > 0.10). During inbound lookbacks, the longest gaze fixation was not significantly directed in relation to the opposite of travel direction (Rayleigh’s Test, Z = 1.68; p = 0.19). There was a significant directedness in both outbound conditions (Rayleigh’s Tests, p < 0.05). Within-individual comparisons showed that the longest-gaze-fixation direction and the reversal direction did not significantly differ in all three conditions (Moore’s Paired Tests; Near Nest: R’ = 0.278 ; p = 0.790; Funnel: R’ = 0.219 ; p = 0.710; Inbound: R’ = 0.897 ; p = 0.340). As circular statistics will generally achieve significance with data grouped in the correct hemisphere (all lookbacks by our classification sytem had to break the ±90° travel direction plane), when within-individual comparisons were significantly different, we compared the within-individual absolute angular error of the longest gaze fixation with both the nest direction and 180° of travel direction. In all three conditions absolute angular error was significantly lower against the nest direction versus the 180° of travel direction (Wilcoxon Tests, Near Nest: Z = –4.13, p < 0.001; Funnel: Z = –3.4, p = 0.001; Inbound: Z = –2.1, p = 0.036; Figure 7A,B), indicating foragers fixated closer directionally to the nest direction than to the inverse of their outbound or inbound path in all conditions. We found no significant difference between conditions in mean gaze fixation duration, longest gaze fixation duration ( p > 0.05; Figure 8A,B), but did find a significant effect of condition on the number of fixations, with inbound lookbacks exhibiting significantly fewer fixations compared to both Near Nest and Funnel outbound conditions (Figure 8C). In all three conditions, we found no significant within-individual differences in the mean gaze fixation duration before and after the reversal (Wilcoxon Tests; p > 0.05). In both the Near Nest outbound and Inbound conditions, we found a significant decrease in the number of fixations after the reversal compared to prior to the reversal of the turn although this pattern was not present in the Funnel condition. Discussion Desert ants ( Melophorus and Cataglyphis ) are known to rapidly learn visual scenes, and to successfully orient and efficiently navigate back home after only a single experience of the outbound route, even when without a corresponding vector (Freas and Cheng 2017 , 2018 ; Freas and Spetch 2019 ). Thus, this outbound portion of the first trip of the route must be a critical period for view learning to allow for rapid inbound route formation, yet the mechanisms by which ants accomplish this are unclear. Here, we show that lookbacks constitute a critical learning period of the inbound nest-aligned scene for the return trip. This behaviour includes the same nest-directed gaze fixation that is present in learning walk pirouettes in Cataglyphis desert ants, behaviours which are highly associated with view learning (Grob et al. 2017 ; Fleischmann et al. 2017 , 2018 ; Freas et al. 2019 ). Thus, nestward rotations and their accompanying gaze fixations underly view learning both around the nest, during learning walk pirouettes, and along the foraging route, during lookbacks. Furthermore, the increased tendency for foragers to engage in lookbacks when traveling through areas of high local panorama change suggests that panorama change may be a triggering mechanism, likely through increased navigational uncertainty, however that is encoded in an ant’s brain. This would allow lookbacks to be performed when the view becomes novel, encouraging the learning of new, unfamiliar route views as the forager first experiences them via panorama change. Route formation Inbound route formation generally aligned with previous work (Freas and Cheng 2017 , 2018 ), showing single-experience learning of the inbound route after an outbound route exposure, though importantly, previous work did not definitively establish this occurred on a single outbound trip rather than after multiple unsuccessful foraging trips before reaching a feeder. Individuals were well oriented returning home, with high path straightness (µ ± Std. = 0.88 ± 0.06) at asymtope even on Trip 1, with only a single outbound experience, suggesting sufficient learning had already occurred during Trip 1’s outbound search to the feeder, though note that these ants did have a corresponding global vector. Foragers exhibited non-negligible scan counts until Trip 4, suggesting further learning and performance improvements over the first few outbound trips (inbound lookbacks were rare). In contrast, the outbound route needed more exposures to complete its route formation, with performance metrics taking until Trips 3–6 to plateau. The ~ 16h overnight delay had a clear effect on route formation on both the outbound and inbound path portions, with decreases in path straightness and increases in scan and lookback counts. Coupled with this, a slight majority of foragers also conducted at least one re-learning walk around the nest on the morning of the second day of their foraging career. This suggests that both learning walks and outbound lookbacks also have a function in maintaining established routes, likely to continuously assess if the visual scene has changed. Many aspects of the visual panorama look different from the last trip of the day, typically occurring in the late afternoon (4–5pm) to the first trip of the next day the following morning (~ 9am), including shadows and light reflectivity off terrestrial cues (Wehner 1997 ). Given that these aspects of the scene change throughout the day and overnight, some level of view memory maintenance through lookbacks and relearning walks may help support route formation over multiple days. Some fading or decay of memory might have also contributed to the deterioration of navigational performance overnight. One study on the acquisition of landmark-based homing to the nest improved after 2 days of training, that is, performance on day 3 was better than on day 1 or on day 2 (Narendra et al. 2007 ). That study provided a richer array of landmarks near the nest, with the nest being surrounded by 4 black cylinders. The number of training trials was not a critical factor in acquisition, but spreading training over multiple days benefitted the ants. After 2 days of training, landmark memories seemed to last a lifetime, with the ants performing well even after 8 days of delay. This study leaves open the possibility that landmark memory, what is now more commonly called view memory, might have faded after the first day of foraging, requiring lookbacks on the next day of foraging to produce long-term route formation. Local panorama change We found that lookback counts on the outbound route are associated with spatial locations where there is a high degree of local panorama change, measured by comparing the scene to the view 50cm earlier in the route. Given that learning-walk-acquired view memories of the nest area are only useful within a catchment area, foragers should be able to modulate when to perform lookbacks based upon visual panorama metrics of change. This would allow foragers to acquire new route-based nest-aligned view memories to help return home. It remains unclear if rapid panorama change is a triggering mechanism for foragers to lookback or if this change is also typically associated with a spike in unfamiliarity, with unfamiliarity triggering the behaviour. Regardless of the mechanism, looking back to the nest in areas where the panorama changes quickly would align with this behaviour having a view-learning function. Future work could characterise if lookbacks can be predictably triggered in specific locations in response to exposure to unfamiliar visual cues along known routes. Nest-aligned gaze fixations In the current study, outbound lookbacks contained multiple gaze fixations with the longest of these fixations being grouped in the nest direction. This makes lookbacks structurally akin to learning-walk-based pirouettes around the nest, and both likely serve the same function of acquiring nest-aligned views for inbound route formation (Grob et al. 2017 ; Fleischmann et al. 2017 , 2018 ; Freas et al. 2019 ). Nest-aligned gaze fixations were present at both outbound recording locations, near the nest and at the funnel. Across all recorded metrics of the lookback (including mean and longest fixation durations and number of fixations), we found no significant difference between these sites, suggesting the choreography of the outbound lookback is maintained across the route. Inbound lookbacks, in addition to containing a longest gaze fixation back towards the feeder instead of the nest, also showed significantly fewer gaze fixations compared to outbound lookbacks. Given these lookbacks were similar in every other metric besides their frequency, it is hard to untangle why this might be the case, though there likely is an overall lower level of uncertainty during these trips given the global vector and recency of the preceding outbound trip, which might reduce fixation numbers. Nest-aligned gaze directions and fixations appear to be closely tied to navigational view learning throughout a forager’s above-ground life, both before and after the onset of foraging. The views around the nest are known to be acquired during pre-foraging learning walks. Within these walks, naïve Cataglyphis ( C. noda and C. aenescens ) ants inhabiting cluttered environments exhibit highly choreographed turns during their pre-foraging learning walks, in the form of pirouettes, conducting either 360° or 180° turns. Scattered throughout this rotation are gaze fixations with the longest fixation in the direction of the nest, presumably to acquire this nest-aligned view (Grob et al. 2017 ; Fleischmann et al. 2017 , 2018 ). This behaviour seems dependent on the visual scene, with a species inhabiting barren habitat ( Cataglyphis fortis ) facing the nest only briefly and without fixations (called “voltes”, Fleischmann et al. 2017 ), while pirouette-like behaviours are documented in other ants within cluttered habitats (Wystrach et al. 2014; Jayatilaka et al. 2018 ). M. bagoti’s learning walks also appear to exhibit turns with distinct nest-oriented gaze fixations (Fig. 1 B). However, a comprehensive study of learning-walk pirouette gaze fixations in M. bagoti is currently absent. Given that nest-ward gaze fixations are highly associated with view acquisition during nest panorama learning, the frequency, location, and composition of lookbacks along the route all align with these behaviours being akin to learning-walk pirouettes, forming view-acquisition periods to aid foragers in acquiring the nest-aligned views of foraging routes for the ensuing inbound trips (Graham and Collett 2006; Freas and Cheng 2017 , 2018 ; Freas and Spetch 2019 ). Our description of ‘hook turn’ lookbacks bears some similarity to the path reversals and looking back in wood ants, Formica rufa (Graham and Collett 2006). Wood ants exhihit similar inbound learning during the outbound trip, which Graham and Collett (2006) ascribed to ‘reversal’ periods in which the outbound ant turned back and travelled towards the nest for a short distance before resuming its travel to the feeder. They also report that these behaviours occurred early in route formation, though they make no mention of any associated fixations or stopping periods. This similarity suggests that looking back is a consistent learning strategy for route formation in ants. While nest-aligned fixations are well established during learning walks, a key difference between gaze fixations during route-based lookbacks and pirouettes is the individual’s final goal location (being the feeder during outbound lookbacks and back to the nest during learning walks). Since during route-based lookbacks, the forager has some experience of the route and likely a long-term vector memory of the feeder site, it is worth untangling some possible alternative explanations for the nest-based gaze fixations, all of which we ultimately find unconvincing. Two alternative explanations exist for the observed nest-oriented gaze fixations during lookbacks. First, outbound foragers could be rotating 180° from their travel direction, which would typically align with the nest, which means that the longest gaze fixation should be opposite the travel direction and not the nest when these directions are different. Such a non-nest-directed view might be predicted as a forager would want to learn its outbound route even if it conflicts with the nest directionally, for example, if the route has two legs separated by a turn. Yet, in our results, foragers under-rotate to align with the nest when not travelling directly away from the nest, leading to significantly more directional error from 180° of travel direction compared to error from the nest direction (Fig. 7 B). The pattern suggests that the forager is turning until identifying that its gaze is oriented with the nest direction before reversing its turn (for example path see Fig. 1 B). Thus, it appears unlikely that these lookbacks are opposite the direction of travel and are instead tuned to stop to fixate at the nest. This theory does, however, pose interesting questions for lookbacks and route formation on non-straight-line routes, where route formation and the nest would not directionally align. The second alternative explanation is that the nest direction should typically correspond with the direction opposite of the feeder (Fig. 1 A). Consequently, when the lookback reaches 180° of the feeder, uncertainty over which way to turn should increase as both the left and right angular distances to the feeder becomes close to equal. This uncertainty should reach its maximum when 180° opposite the feeder and facing the nest and might cause the longest fixation to be towards the nest irrespective of the nest’s presence, thus accounting for this longer fixation period. However, we find this unconvincing as the nest is only directionally opposite the feeder along the straight-line route, which foragers often deviate from. When foragers are positioned laterally off this straight nest-to-feeder line, we observed the same under turns in line with fixating at the nest, suggesting that gaze fixations are to the nest and not merely opposite the goal (this under turn can be seen clearly in Fig. 1 A; Supplememtal Video 2). Yet, a part of this hypothesis may help explain why in some instances, foragers complete 360° loop turns. As stated above, angular-distance uncertainty should plateau at 180° opposite the feeder along the straight-line route from nest to feeder, meaning a forager near this line may become more uncertain of which is the most efficient turning direction to return to outbound travel, resulting in foragers sometimes continuing to turn in a loop rather than reversing. In support of this hypothesis, all four observed examples of loops took place when foragers were very close to the straight-line route between nest and feeder, where deciding to loop 360° may be a similar directional distance to reversing back 180°. Thus, lookback instances where foragers are near the nest-feeder line might result in foragers being equally likely to conduct 180° or 360° turns given the equidistant angular difference, a phenomenon evident in one instance of a forager traveling almost straight to the feeder, which conducts two 360° lookbacks followed by one 180° lookback turn (See Fig. 1 C for first 360° turn; Supplemental Materials Video 3). This would suggest that after the forager completes its gaze fixation of the nest, to return to face along its outbound route, it likely initially relies on a food vector memory of the feeder location (Bolek et al. 2012 ; Wolf et al. 2012 ; Bolek and Wolf 2015 ) to decide which way to turn (akin to backwards facing navigation; Schwarz et al. 2017 ). When to learn views? The nest-directed gaze-fixation-duration evidence clearly points to ants aligning to the nest, both during learning walks (Fleischmann et al. 2017 ) and now during route formation. How the lookback and pirouette behaviours’ structures may promote visual learning is an interesting query, especially given the multiple rotation and fixation intervals which make up each behaviour as well as lookbacks’ occurrence in association with panorama change. Additionally, lookbacks should occur regardless of whether an outbound search is successful, leaving open the question of how the ant’s brain learns the nest-aligned route views acquired during lookbacks. Our current understanding is that long term view-based memories are stored within the mushroom bodies along with valence based on positive and negative experiences with these views (Ardin et al. 2016 ; Buehlmann et al. 2020; Kamhi et al. 2020; Webb and Wystrach 2016 , 2020; Le Moël and Wystrach 2020a; Freas et al. 2022 ). View memories and visual route formation in ants involve memory storage within the ants’ mushroom bodies (Buehlmann et al. 2020; Kamhi et al. 2020). Individual views of the route can be represented neurally within the mushroom bodies through specific Kenyon cell (KCs) activation patterns. Whether a view memory is associated with the goal location and thus has an attractive valence is accomplished through the projection of these activation patterns onto multiple motor output neurons (MBON), with memory valence controlled by the activation of dopaminergic neurons, which mediates the association between a view memory and the associated valence stemming from negative or positive experiences (Cohn et al. 2015; Aso et al. 2015). However, lookbacks do not follow or precede any positive or negative reinforcement, thus leaving open the question of how the nest-aligned fixations become attractive. Recent work in the head-direction network of Drosophila (Fisher et al. 2022) suggests that visual rotation may provide a ‘when to learn signal’, thus enhancing latent or ‘unsupervised’ visual learning during lookbacks. Dopamine is released when a visual cue rotates, thus strengthening visual learning during periods when cue rotation provides a rich stream of visual information. The authors theorise that constrains visual learning to these periods and may protect established visual memories during other periods such as straight-line movement. Given that visual rotation should enhance learning, the observed large rotational changes that occur during lookback turns should be clear periods in which view learning is enhanced. The intermittent fixations may also aid in this function, with short fixations periods allowing for better encoding of a stable view. Learning during the rapid-rotation periods may lead the ant to learn degraded blurry views of the route, thus decreasing navigational performance. Thus, the alternation of rotations and gaze fixations within the lookback behaviour together facilitate view learning and route formation. Furthermore, reliance on dopamine for learning may require gaze fixations to be reasonably short in duration so that the visual-system signals of rotation that promote learning still linger. Longer fixations in contrast might break this coupling, leading to a lack of dopamine presence during fixation. While highspeed video analysis of fixation duration and interval during lookbacks is ongoing, fixation durations in M. bagoti scanning bouts shows a strong leftward skew indicating many short fixation periods (Deeti et al. 2023 ). The coupling of rotation as a ‘when to learn signal’ and short gaze fixations may explain the left-skewed distribution of fixation durations in relearning walks (Deeti et al. 2023 ) as well as the prevalence of short fixations during lookbacks in this study. Finally, the benefit of learning the nest-aligned views through rotation and nest centred gaze fixations appears clear, yet it presents an interesting conflict with the modelling of route formation. Modelling has shown that correctly goal-oriented views alone are insufficient to produce efficient route formation and that only when familiarity estimates corresponding to orientations to the left and right of the goal are included does route formation become robust (Wystrach et al. 2020b ). To bring these ideas into agreement, it is important to note that the nest-aligned fixations, while the longest, represent only one of many fixations that occur throughout the lookback. This suggests that these non-nest fixations within the lookback also represent learning periods, primed by rotation, for views along the route acquired in conjunction with heading representations of goal locations within the central complex (Seelig et al. 2015; Stone et al. 2017 ). Despite the heavy focus on the longest nest-aligned gaze fixations when discussing learning walks and flights (Grob et al. 2017 ; Fleischmann et al. 2017 , 2018 Robert et al. 2018 ; Collett and Hempel de Ibarra 2023; Collett et al. 2023 ; Freas et al. 2019 ), non-nest-aligned fixations are also well known to occur during these learning events, just as with our lookbacks. Our speculation is that such non-nest-aligned views become associated with turn signals for turning in order to reach a nest-aligned view, signals that will become crucial for steering homeward later. Wasps have also been shown to rhythmically alternate between fixations to the left and right of a goal site (Stürzl et al. 2016). Cataglyphis ants as well as M. bagoti (Fig. 1 b) have been shown to rotate between nest and anti-nest directions with fixations in multiple directions in between these reversals (Fleischmann et al. 2017 ). Thus, all the fixations observed during these periods of rotation may constitute periods for learning views, critical for both learning the nest views and for route formation via lookbacks. Conclusions Their frequency at the beginning of route formation, association with local panorama change and structural similarity to learning walk pirouettes all point to lookback behaviours supporting route-based view learning of areas beyond the nest at the beginning of route formation. We show that lookbacks contain multiple gaze fixations of different orientations of route-based views with the longest fixation being associated with the nest direction. This nest-ward fixation likely serves as a critical period of learning the inbound route to allow the forager to quickly acquire the visual information needed to return home. Declarations Acknowledgments Trevor Murray helped with creating code to run batch rotIDF comparisons. We thank him for his assistance. Funding This project was funded by a Macquarie University Research Fellowship (MQRF0001094), by Macquarie University, and by an ARC Discovery Grant (DP200102337). Land Acknowledgment This work was conducted upon the grounds of the Centre for Appropriate Technology whom we thank for access to the nest. This work was conducted upon land traditionally owned by the Arrernte. We acknowledge the traditional custodians of the land on which our field site sits. Their culture and customs have nurtured and sustained this land since the Dreamtime and continue to do so. We pay our respects to their Elders, past, present, and future. Conflicts of interest The authors declare no conflicts of interest associated with this work. Ethics There are no state or federal governmental regulations guiding the research of invertebrates in Australia. References Ardin, P. Peng, F. Mangan, M. Lagogiannis, K. & Webb, B. (2016). Using an insect mushroom body circuit to encode route memory in complex natural environments. PLoS computational biology, 12(2), e1004683. Baddeley, B. Graham, P. Philippides, A. & Husbands, P. (2011). Models of visually guided routes in ants: Embodiment simplifies route acquisition. In S. Jeschke, H. Liu, D. Schilberg (Eds.), Proceedings of the international conference on intelligent robotics and applications (ICIRA) part II, lecture notes in artificial intelligence (pp. 75–84). Heidelberg, Germany: Springer-Verlag. Baddeley, B. Graham, P. Husbands, P. & Philippides, A. (2012). A model of ant route navigation driven by scene familiarity. PLoS computational biology, 8(1), e1002336. Bolek, S. & Wolf, H. (2015). Food searches and guiding structures in North African desert ants, Cataglyphis. Journal of Comparative Physiology A, 201, 631-644. Bolek, S. Wittlinger, M. & Wolf, H. (2012). Establishing food site vectors in desert ants. Journal of Experimental Biology, 215(4), 653-656. Cheng, K. & Freas, C. A. (2015). Path integration, views, search, and matched filters: the contributions of Rüdiger Wehner to the study of orientation and navigation. Journal of Comparative Physiology A, 201, 517-532. Cheng, K. Schultheiss, P. Schwarz, S. Wystrach, A. & Wehner, R. (2014). Beginnings of a synthetic approach to desert ant navigation. Behavioural processes, 102, 51-61. Collett TS, Collett M 2000. Path integration in insects. Current Opinion in Neurobiology 10:757–762. Collett, T. S. & Hempel de Ibarra, N. (2023). An ‘instinct for learning’: the learning flights and walks of bees, wasps and ants from the 1850s to now. Journal of Experimental Biology, 226(6), jeb245278. Collett, M. Chittka, L. and Collett, T. S. (2013). Spatial memory in insect navigation. Curr. Biol. 23, R789-R800. https://doi.org/10.1016/j.cub.2013.07.020 Collett, T. S. Robert, T. Frasnelli, E. Philippides, A. & Hempel de Ibarra, N. (2023). How bumblebees coordinate path integration and body orientation at the start of their first learning flight. Journal of Experimental Biology, 226(8), jeb245271. Deeti, S. Islam, M. Freas, C. Murray, T. & Cheng, K. (2023). Intricacies of running a route without success in night-active bull ants ( Myrmecia midas ). Journal of Experimental Psychology: Animal Learning and Cognition, 49(2), 111. Fleischmann, P. N. Christian, M. Müller, V. L. Rössler, W. & Wehner, R. (2016). Ontogeny of learning walks and the acquisition of landmark information in desert ants, Cataglyphis fortis. The Journal of Experi- mental Biology, 219, 3137–3145. http://dx.doi.org/10.1242/jeb.140459 Fleischmann, P. N. Grob, R. Wehner, R. & Rössler, W. (2017). Species- specific differences in the fine structure of learning walk elements in Cataglyphis ants. The Journal of Experimental Biology, 220, 2426–2435. Fleischmann, P. N. Rössler, W. & Wehner, R. (2018). Early foraging life: spatial and temporal aspects of landmark learning in the ant Cataglyphis noda. Journal of Comparative Physiology A, 204(6), 579-592. Freas, C. A. & Cheng, K. (2017). Learning and time‐dependent cue choice in the desert ant, melophorus bagoti. Ethology, 123(8), 503-515. Freas, C. A. & Cheng, K. (2018). Landmark learning, cue conflict, and outbound view sequence in navigating desert ants. Journal of Experimental Psychology: Animal Learning and Cognition, 44(4), 409. Freas, C. A. & Cheng, K. (2019). Panorama similarity and navigational knowledge in the nocturnal bull ant Myrmecia midas. Journal of Experimental Biology, 222(11), jeb193201. Freas, C. A. & Cheng, K. (2022). The basis of navigation across species. Annual review of psychology, 73, 217-241. Freas, C. A. & Spetch, M. L. (2019). Terrestrial cue learning and retention during the outbound and inbound foraging trip in the desert ant, Cataglyphis velox. Journal of Comparative Physiology A, 205(2), 177-189. Freas, C. A. & Spetch, M. L. (2023). Varieties of visual navigation in insects. Animal cognition, 26(1), 319-342. Freas, C. A. & Schultheiss, P. (2018). How to navigate in different environments and situations: lessons from ants. Frontiers in psychology, 9, 369115. Freas, C. A. Fleischmann, P. N. & Cheng, K. (2019). Experimental ethology of learning in desert ants: becoming expert navigators. Behavioural Processes, 158, 181-191. Freas, C. A. Plowes, N. J. & Spetch, M. L. (2021). Traveling through light clutter: Path integration and panorama guided navigation in the Sonoran Desert ant, Novomessor cockerelli. Behavioural Processes, 186, 104373. Freas, C. A. Wystrach, A. Schwarz, S. & Spetch, M. L. (2022). Aversive view memories and risk perception in navigating ants. Scientific Reports, 12(1), 2899. Grob, R. Fleischmann, P. N. Grübel, K. Wehner, R. & Rössler, W. (2017). The role of celestial compass information in Cataglyphis ants during learning walks and for neuroplasticity in the central complex and mushroom bodies. Frontiers in behavioral neuroscience, 11, 226. Harris, R. A. Graham, P. & Collett, T. S. (2007). Visual cues for the retrieval of landmark memories by navigating wood ants. Current biology, 17(2), 93-102. Islam, M. Freas, C. A. & Cheng, K. (2020). Effect of large visual changes on the navigation of the nocturnal bull ant, Myrmecia midas. Animal Cognition, 23, 1071-1080. Jayatilaka P, Murray T, Narendra A, Zeil J (2018) The choreography of learning walks in the Australian jack jumper ant Myrmecia croslandi. J Exp Biol 221:jeb185306. 10.1242/jeb.185306 Kodzhabashev, A. & Mangan, M. (2015). Route following without scanning. In Biomimetic and Biohybrid Systems: 4th International Conference, Living Machines 2015, Barcelona, Spain, July 28-31, 2015, Proceedings 4 (pp. 199-210). Springer International Publishing. Le Möel, F. & Wystrach, A. (2020). Opponent processes in visual memories: A model of attraction and repulsion in navigating insects’ mushroom bodies. PLoS computational biology, 16(2), e1007631. Lionetti, V. A. Deeti, S. Murray, T. & Cheng, K. (2023). Resolving conflict between aversive and appetitive learning of views: how ants shift to a new route during navigation. Learning & Behavior, 51(4), 446-457. Möller, R. (2012). A model of ant navigation based on visual prediction. Journal of theoretical biology, 305, 118-130. Murray, T. Kocsi, Z. Dahmen, H. Narendra, A. Le Möel, F. Wystrach, A. & Zeil, J. (2020). The role of attractive and repellent scene memories in ant homing (Myrmecia croslandi). Journal of Experimental Biology, 223(3), jeb210021. Narendra, A. Si, A. Sulikowski, D. & Cheng, K. (2007). Learning, retention and coding of nest-associated visual cues by the Australian desert ant, Melophorus bagoti . Behavioral Ecology and Sociobiology, 61, 1543-1553. Pfeiffer, K. & Homberg, U. (2014). Organization and functional roles of the central complex in the insect brain. Annual review of entomology, 59, 165-184. Robert, T. Frasnelli, E. Hempel de Ibarra, N. & Collett, T. S. (2018). Variations on a theme: bumblebee learning flights from the nest and from flowers. Journal of Experimental Biology, 221(4), jeb172601. Schultheiss, P. Wystrach, A. Schwarz, S. Tack, A. Delor, J. Nooten, S. S. ... & Cheng, K. (2016). Crucial role of ultraviolet light for desert ants in determining direction from the terrestrial panorama. Animal Behaviour, 115, 19-28. Schwarz, S. Mangan, M. Zeil, J. Webb, B. and Wystrach, A. (2017). How Ants Use Vision When Homing Backward. Curr. Biol. 27, 401–407. https://doi.org/10.1016/j.cub.2016.12.019 Seelig, J. D. & Jayaraman, V. (2015). Neural dynamics for landmark orientation and angular path integration. Nature, 521(7551), 186-191. Stone, T. Webb, B. Adden, A. Weddig, N. B. Honkanen, A. Templin, R. ... & Heinze, S. (2017). An anatomically constrained model for path integration in the bee brain. Current Biology, 27(20), 3069-3085. Webb, B. & Wystrach, A. (2016). Neural mechanisms of insect navigation. Current Opinion in Insect Science, 15, 27-39. Wehner R (1982) Himmelsnavigation bei Insekten: Neurophysiologie und Verhalten. Orell Fussli, Zurich Wehner, R. (1997). The ant's celestial compass system: spectral and polarisation channels. In Orientation and Communication in Arthropods (ed. M. Lehrer), pp.145-185. Basel: Birkhäuser Verlag. Wehner, R. (2003). Desert ant navigation: how miniature brains solve complex tasks. Journal of Comparative Physiology A, 189, 579-588. Wehner, R. & Räber, F. (1979). Visual spatial memory in desert ants, Cataglyphis bicolor (Hymenoptera: Formicidae). Experientia, 35(12), 1569-1571. Wehner, R. & Srinivasan, M. V. (2003). Path integration in insects. The neurobiology of spatial behaviour, 9-30. Wehner, R. Michel, B. and Antonsen, P. (1996). Visual navigation in insects: coupling of egocentric and geocentric information. J. Exp. Biol. 199, 129–140. Wolf, H. Wittlinger, M. & Bolek, S. (2012). Re-visiting of plentiful food sources and food search strategies in desert ants. Frontiers in neuroscience, 6, 102. Wystrach, A. Buehlmann, C. Schwarz, S. Cheng, K. & Graham, P. (2020a). Rapid aversive and memory trace learning during route navigation in desert ants. Current biology, 30(10), 1927-1933. Wystrach, A. Le Moël, F. Clement, L. & Schwarz, S. (2020b). A lateralised design for the interaction of visual memories and heading representations in navigating ants. BioRxiv, 2020-08. Wystrach, A. Cheng, K. Sosa, S. & Beugnon, G. (2011). Geometry, features, and panoramic views: ants in rectangular arenas. Journal of Experimental Psychology: Animal Behavior Processes, 37(4), 420. Zeil, J. (2012). Visual homing: an insect perspective. Curr. Opin. Neurobiol. 22(2), 285–293. Zeil, J. Hofmann, M. I. & Chahl, J. S. (2003). Catchment areas of panoramic snapshots in outdoor scenes. JOSA A, 20(3), 450-469. Additional Declarations The authors declare no competing interests. Supplementary Files SupplementalTable1.pdf Supplemental Table 1 SupplmentalVideo1.mp4 Supplemental Video 1. M. bagoti Learning Walk (Pirouettes) SupplementalVideo2.mp4 Supplemental Video 2. M bagoti lookback (180º turn) from Figure 1a SupplementalVideo3.mp4 Supplemental Video 3. M bagoti lookback loops and 180º turn Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4670516","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":321459363,"identity":"8279d607-54c2-43ad-862f-6077577e7c4f","order_by":0,"name":"Cody A Freas","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4klEQVRIiWNgGAWjYHACAwYGGxDNfABEMjYQpyUNRLMlkKyFx4A4LfztzdskGBLuyfH3n/m64Q2DjeyGA8zPJPBpkThzrAyopdhY4kbutptzGNKMNxxgM8OrxUAix0yC8UdCYsMN3m23eRgOJ244wECEFoaEhPr55888A2r5D9TC/o0oLQkGB3LYgFoOALXw4LcF6JdiC6AOw4030sxuzjFINp55mKfYAp8WYIhtvPEhIUFe7vzhZzfeVNjJ9h1v33gDnxYgYJFIgDHBUcNMQD1IyQc4k4ew6lEwCkbBKBiBAABK+Uqlr3pQ4AAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0001-7026-1255","institution":"Macquarie University","correspondingAuthor":true,"prefix":"","firstName":"Cody","middleName":"A","lastName":"Freas","suffix":""},{"id":321459364,"identity":"c55ed4c4-54d6-42a3-8a0f-a663d6149480","order_by":1,"name":"Ken Cheng","email":"","orcid":"https://orcid.org/0000-0002-4913-2691","institution":"Macquarie University","correspondingAuthor":false,"prefix":"","firstName":"Ken","middleName":"","lastName":"Cheng","suffix":""}],"badges":[],"createdAt":"2024-07-02 00:07:59","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-4670516/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4670516/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":59564028,"identity":"33adf644-d817-4a91-8dfc-849c3c0a817b","added_by":"auto","created_at":"2024-07-03 08:47:10","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":582535,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImage and coordinate based representations of nest gaze fixations during learning walk-based pirouettes and the different types of lookback behaviours in \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eMelophorus bagoti.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eImages extracted and compiled from a highspeed video of (\u003cstrong\u003eA\u003c/strong\u003e) a lookback behaviour showing eight fixation phases in an outbound \u003cem\u003eM. bagoti \u003c/em\u003eforager travelling to the feeder and (\u003cstrong\u003eB\u003c/strong\u003e) learning walk-based pirouettes (for videos see supplemental materials). Videos were taken at 600fps at 1080p using a Chronos 2.1HD camera. The orange and red arrows in panel b denote the gaze fixations in the nest and anti-nest direcitons during each pirouette. For representations of lookbacks, dots denote the position of the head base (black dot) and mandibles (grey dot) during the three different observed types of lookbacks (see illustrated forager in panel B) recorded every 40ms; (\u003cstrong\u003eC\u003c/strong\u003e) the ~180° turn in place (\u003cstrong\u003eD\u003c/strong\u003e) the 360° circular turn and a (\u003cstrong\u003eE\u003c/strong\u003e) hook turn. Translational and rotational movements are denoted by the white arrows. Fixation phases (\u0026gt;40ms), defined as when the ant’s gaze does not shift (within 2°) between frames while also remaining in a 5° window over successive frames of the fixation period are indicated by consecutively numbered arrows. For the pirouettes (A), orange arrows denote the nest fixation while red arrows denote the angular maximum outward gaze fixation. For lookback panels (A,C-E), green arrows denote fixations which occur prior to the turning reversal, while blue arrows denote fixations occurring post reversal. The orange arrow in each panel indicates the longest fixation period and the grey arrow denotes the reversal fixation if the reversal does not align with the longest pause direction.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/e5ff72564845c7183aeb9059.png"},{"id":59564021,"identity":"6ef1c354-396d-4460-91ba-a6509ac240b0","added_by":"auto","created_at":"2024-07-03 08:47:09","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":212832,"visible":true,"origin":"","legend":"\u003cp\u003ePhotos and\u003cstrong\u003e \u003c/strong\u003eDiagram of the experimental set-up for route formation. (A) Photo of setup illustrating the enclosed route from the nest through the funnel to the feeder (small shrub was removed from nest area prior to testing). Panoramic 360° photos of the surrounding visual cues during testing at (B) the nest, (C) the funnel and (D) the feeder. (E) Diagram of the arena. Each photo is oriented to the feeder direction (photo centre). Foragers were allowed to travel freely on the outbound trip to the feeder and return to the nest with the collected food piece.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/97dd0014b7f6932b7310b902.png"},{"id":59564025,"identity":"f907d98c-71b2-44d9-85ea-8f7c2bee5979","added_by":"auto","created_at":"2024-07-03 08:47:09","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":54502,"visible":true,"origin":"","legend":"\u003cp\u003eForagers’ outbound and inbound paths on Trip 1, Trip 5 and Trip 10 of each forager’s first foraging day and the first two trips of the next day of foraging. A single example path is shown in black while all other paths are in grey. For the example path, closed black circles denote locations where the forager’s stopped and performed scanning bouts while open circles denote lookback locations. *Note: a single forager began its foraging career during the local termite populations nuptial flights meaning during this period wingless termite queens would make it into the arena just above the funnel at 5m from the nest and this forager developed a route to this area to collect these abundant food pieces over its first two days of foraging.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/c669479bc799904b2e10408a.png"},{"id":59564023,"identity":"99075e58-9064-445b-8ccb-61250ae7bfef","added_by":"auto","created_at":"2024-07-03 08:47:09","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":110863,"visible":true,"origin":"","legend":"\u003cp\u003eGraphs denoting the change in the performance metrics during route formation. (A) Outbound and (B) Inbound path straightness, (C) Outbound and (D) Inbound Scan count and (E) Outbound and (F) Inbound Lookback counts are shown across all fifteen recorded foraging trips. The overnight delay is referenced by a black line. * denote when each metric reaches asymptope calculated via Helmert contrasts (with \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01).\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/b904bd38a47f05224b7c4ca9.png"},{"id":59564027,"identity":"e3dda771-8162-41ce-88d9-458bdabd75e1","added_by":"auto","created_at":"2024-07-03 08:47:10","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":294482,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMean path characteristic changes (by preceding trip: that day or after an overnight delay) for both the outbound and inbound portion of the foraging trip. \u003c/strong\u003eMean Change calculated per individual in (\u003cstrong\u003eA\u003c/strong\u003e) path straightness (\u003cstrong\u003eB\u003c/strong\u003e) scans and (\u003cstrong\u003eC\u003c/strong\u003e) lookbacks across all trips based on the day of the preceding trip. Trip 1 for each individual was excluded given it had no preceding foraging trip. Box plots show the median hesitation change (dotted line), mean hesitation change (solid line) and 25th and 75th percentile (box) while the whiskers extend to min and max values. Raw data points for each individual are reported next to each box plot with within individual comparisons linked with a grey line, to illustrate general trends. *Denotes a significant difference between conditions using Wilcoxon ranked sum tests for paired data (p \u0026lt; 0.05).\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/db0cd34e38f1906959a56537.png"},{"id":59564031,"identity":"f311ddff-a563-4025-a959-dec364c60194","added_by":"auto","created_at":"2024-07-03 08:47:10","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1370496,"visible":true,"origin":"","legend":"\u003cp\u003eQuantifying the rate of visual panorama change which occurs locally and globally along the foraging route. (\u003cstrong\u003eA\u003c/strong\u003e) Foraging arena with locations where 360° panoramic images were collected along 50cm distances in a grid like fashion. Grey filled dots denote areas where the local panorama changes rapidly over space (last 50cm), corresponding with low local similarity. Of this grid we chose two areas of the route where the local panorama rapidly changes to illustrate rotIDFs. (\u003cstrong\u003eB\u003c/strong\u003e) Panoramic images beginning at the nest entrance in 50cm increments and (\u003cstrong\u003eC\u003c/strong\u003e) images beginning at 500cm from the nest illustrating the change occurring as the forager passes through the funnel. RotIDFs compare the root mean square pixel difference between each panorama with the nest (blue) panorama and the 50cm preceding panorama (red). This illustrated both the global similarity in the panorama across the entire route (notice the lack of a clear valley of mismatch in the B panels), as well as how much similarity is maintained over short distances (high local similarity), which are associated with lookbacks. (\u003cstrong\u003eD\u003c/strong\u003e) Graphs plotting the lookbacks at given spatial locations along the route against the local panorama similarity, global panorama similarity to the nest and the foragers’ vector length (distance to the nest).\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/add1b79bd8b907a97f0b04a4.png"},{"id":59564026,"identity":"0f962f27-d43b-4815-a24e-9bbe3ffede79","added_by":"auto","created_at":"2024-07-03 08:47:09","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":162262,"visible":true,"origin":"","legend":"\u003cp\u003eCircular histograms indicating (\u003cstrong\u003eA\u003c/strong\u003e) longest gaze fixation and the reversal direction (absent in 360° lookbacks) of individual loopback behaviours in the three conditions: Outbound Near Nest; Outbound Funnel, and Inbound. Grey arrows denote the nest direction while black arrows denote the feeder direction in the inbound condition. The red arrow indicates the direction and length of the mean vector. \u003cem\u003en\u003c/em\u003e, number of lookbacks; Ø, mean vector; \u003cem\u003er\u003c/em\u003e, mean vector length. (B) Within-individual mean absolute angular error of the longest gaze fixation between the opposite of travel direction (180°) in blue and the nest direction in red. Box plots show the median hesitation change (dotted line), mean hesitation change (solid line) and 25th and 75th percentile (box) while the whiskers extend to min and max values (excluding outliers). Raw data points for each lookback are reported next to each box plot with within-individual comparisons linked with a grey line, to illustrate general trends.\u003cstrong\u003e \u003c/strong\u003e*Denotes a significant difference between conditions (p \u0026lt; 0.05).\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/937b5f56b974c0889964dabe.png"},{"id":59564707,"identity":"9fdfcd83-abaa-4454-bc88-04440e054c8e","added_by":"auto","created_at":"2024-07-03 08:55:09","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":510028,"visible":true,"origin":"","legend":"\u003cp\u003eGaze fixation duration and counts comparisons within and across the three lookback conditions.\u003cstrong\u003e \u003c/strong\u003e(\u003cstrong\u003eA\u003c/strong\u003e) Mean fixation duration (\u003cstrong\u003eB\u003c/strong\u003e) longest gaze fixation duration and (C) fixation count between conditions. Within individual lookbacks comparisons (\u003cstrong\u003eD\u003c/strong\u003e) mean fixation durations and (\u003cstrong\u003eE\u003c/strong\u003e) number fixations between pre and post reversal fixations. Box plots shows the median hesitation change (dotted line), mean hesitation change (solid line) and 25th and 75th percentile (box) while the whiskers extend to min and max values (excluding outliers). Raw data points for each individual are reported next to each box plot. When paired data points showed a significant within individual trend, each data pair was linked with a grey line, to illustrate the general trend pre and post reversal.\u003cstrong\u003e \u003c/strong\u003e*Denotes a significant difference between conditions (p \u0026lt; 0.05).\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/c0b594b73a75ea724ecaa6c6.png"},{"id":59565388,"identity":"c7af1fe3-2427-4c35-b8ff-99dd2a587a59","added_by":"auto","created_at":"2024-07-03 09:03:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4491708,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/2711997c-1217-4e7c-885d-54d023ff69a0.pdf"},{"id":59564022,"identity":"e20ab3c0-0723-47bf-a325-cf475f24eb8f","added_by":"auto","created_at":"2024-07-03 08:47:09","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":110966,"visible":true,"origin":"","legend":"\u003cp\u003eSupplemental Table 1\u003c/p\u003e","description":"","filename":"SupplementalTable1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/d6d0be26a9ba3e187ed42dcf.pdf"},{"id":59564035,"identity":"a753b767-8107-41c9-85d8-a6cb8ec437ce","added_by":"auto","created_at":"2024-07-03 08:47:15","extension":"mp4","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":493404347,"visible":true,"origin":"","legend":"\u003cp\u003eSupplemental Video 1. M. bagoti Learning Walk (Pirouettes)\u003c/p\u003e","description":"","filename":"SupplmentalVideo1.mp4","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/67b51c2b17a397ca2ec9233d.mp4"},{"id":59564033,"identity":"907f2858-8978-487b-900e-82be89925e66","added_by":"auto","created_at":"2024-07-03 08:47:11","extension":"mp4","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":153159044,"visible":true,"origin":"","legend":"\u003cp\u003eSupplemental Video 2. M bagoti lookback (180º turn) from Figure 1a\u003c/p\u003e","description":"","filename":"SupplementalVideo2.mp4","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/e1808687dc78fe7d9989385d.mp4"},{"id":59564034,"identity":"5909a74f-944d-41ec-8f6c-f525e8f06ab7","added_by":"auto","created_at":"2024-07-03 08:47:11","extension":"mp4","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":103138033,"visible":true,"origin":"","legend":"\u003cp\u003eSupplemental Video 3. M bagoti lookback loops and 180º turn\u003c/p\u003e","description":"","filename":"SupplementalVideo3.mp4","url":"https://assets-eu.researchsquare.com/files/rs-4670516/v1/74ce98aa9b6418ccfe14c73f.mp4"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eRoute Formation and the Choreography of Looking Back in Desert Ants (\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eMelophorus bagoti\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e)\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDesert ants are model systems for visual navigation as they are able to efficiently navigate both to known food sites and back to the nest, rapidly forming routes between locations (Wehner et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e1996\u003c/span\u003e; Mangan and Webb 2012; Collett et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Cheng et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Freas and Cheng \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). As diurnal scavengers, these foragers are active during the hottest parts of the day when prolonged periods outside the nest can be fatal, thus making efficient homing and rapid spatial memory formation critical to an individual\u0026rsquo;s survival.\u003c/p\u003e \u003cp\u003eTo rapidly home to goal locations, desert ants rely on a shared navigational toolbox comprising multiple strategies (Wehner et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e1996\u003c/span\u003e, 2003; Collett et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Cheng et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Cheng and Freas, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Freas and Schultheiss, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Buehlmann et al. 2020) The two most well understood of these strategies are the maintenance of path-integration-derived vectors (Wehner, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1982\u003c/span\u003e; Collett and Collett, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Wehner and Srinivasan, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2003\u003c/span\u003e) and learning panoramic views of the terrestrial scene, both around the nest and along known foraging routes (Wehner and R\u0026auml;ber, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e1979\u003c/span\u003e; Wehner \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Collett et al. 2006; Harris et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Cheng et al. 2009; Schultheiss et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Freas and Cheng \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Collett and Hempel de Ibarra 2023).\u003c/p\u003e \u003cp\u003eRoute formation via view comparison has been extensively modelled (Zeil et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Wystrach et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Baddeley et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2011\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Kodzhabashev and Mangan \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; M\u0026ouml;ller \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Le M\u0026ouml;el and Wystrach \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Murray et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Within these models, foragers turn and travel towards attractive views by comparing the current view with memories acquired while facing goal locations, based on the current motivational context. Nest-associated view memories involve reinforcement learning of inbound views, which promotes forward movement and leads foragers back to the nest site, and is likely reinforced upon the forager\u0026rsquo;s arrival (Le M\u0026ouml;el and Wystrach \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Murray et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wystrach et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2020a\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Given that visual scenes can change between trips, view memories are continuously updated, and their associated valence modulated in response to scene changes (Le M\u0026ouml;el and Wystrach \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wystrach et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2020a\u003c/span\u003e; Islam et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Deeti et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Lionetti et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Yet, less well understood is how these initial view memories of foraging routes are acquired given that multiple species of desert ant are known to only need a single experience of the outbound route to acquire the necessary visual information to correctly return home (Freas and Cheng \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Freas and Spetch \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003ePrior to foraging, each na\u0026iuml;ve ant, newly emerged from the nest entrance, must learn the most important visual scene, the nest\u0026rsquo;s. This is accomplished through multiple highly structured learning walks, which have been well characterised within the \u003cem\u003eCataglyphis\u003c/em\u003e genus (Fleischmann et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, Grob et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Collett and Hempel de Ibarra 2023). Within these walks, \u003cem\u003eCataglyphis noda\u003c/em\u003e and \u003cem\u003eCataglyphis aenescens\u003c/em\u003e (both species which nest in visually cluttered environments) exhibit scanning behaviours with distinct pausing or fixation phases, including what has been termed \u0026lsquo;pirouettes\u0026rsquo; (Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). During these behaviours, the ant rotates in place either fully 360\u0026deg; or partially with a rotational reversal (left to right or right to left). Interspersed throughout this rotation are a number of fixations, where the ant neither moves forward nor rotates its gaze direction (Grob et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; for \u003cem\u003eM. bagoti\u003c/em\u003e pirouettes see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). Interestingly, the longest of these gaze fixations take place when the ant is oriented to the nest entrance, suggesting this period is when the ant learns the nest-aligned view memory (Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Similar fixation behaviours have been recorded in the learning flights of bumble bees and wasps; thus, this may be a consistent view learning mechanism across Hymenoptera (Robert et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Collett and Hempel de Ibarra 2023; Collett et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Pirouettes, comprising both rotational movement and nest-oriented fixations (Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA) are highly associated with view acquisition and view memory formation (Fisher et al. 2022). Currently, a detailed choreography of pirouettes in Australian desert ants, \u003cem\u003eMelophorus bagoti\u003c/em\u003e, is lacking (for learning walk structure see Deeti and Cheng 2021), though high-speed recordings suggest that na\u0026iuml;ve individuals perform similar oscillating scans during their learning walks that include both nestward gaze fixations as well as fixations away from the nest direction (for an example of these scans, see Supplemental Materials Video 1, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA).\u003c/p\u003e \u003cp\u003eAfter performing these learning walks, ants leave the immediate area around the nest to search for food (Wehner et al. 2004; Fleischmann et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Yet, visual learning continue beyond learning walks, as inexperienced foragers move into unknown scenes beyond the nest area (Freas and Cheng \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). As an individual moves away from known nest views into unfamiliar areas, this translational movement causes the visual scene to change while maintaining a degree of similarity with the acquired nest views (Zeil et al. 2014). This decay in similarity occurs in a predictable fashion, with the highest degree of similarity between the ant\u0026rsquo;s current view and their memorized views occurring when they are facing the direction of their starting location (Nicholson et al. 1999; Zeil \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Zeil et al. 2014). Each view memory has a \u0026lsquo;catchment area\u0026rsquo;, or area around the view location where there is sufficient visual similarity between previously acquired view memories and the current view to usefully orient, though this is highly dependent on the amount of local visual clutter, given the propensity for proximal visual cues to both change rapidly and block distant visual cues during translational movement (Zeil et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2003\u003c/span\u003e, 2014; Str\u0026uuml;zl and Zeil 2007; Murray and Zeil 2017; Jayatilaka et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Freas and Cheng \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Beyond this catchment area, new route-based views must be learned to guide the forager back along this path to the nest. Foraging ants often form routes that can extend up to hundreds of metres, necessitating the acquisition of multiple views along the foraging route and at profitable food sites for successful navigation and route formation (Mangan and Webb 2012; Graham and Cheng 2009; Schultheiss et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Freas et al. 2017, 2018; Jayatilaka et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Freas and Spetch \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSimilar to outbound learning in desert ants, wood ants (\u003cem\u003eFormica rufa\u003c/em\u003e) are also known to exhibit bi-directional route learning during their first few trips to the feeder. Inbound views are thought to be learned during short reversal periods when the ant stops traveling to the feeder, turns back and travels a short distance back to the nest, though these path reversals do not feature any nestward fixations (Graham and Collett 2006). Wood ants are known to fixate at visual landmarks near a newly discovered feeder (Nicholson et al. 1999) and such fixations are thought to have a similar function to pirouette behaviours observed in \u003cem\u003eCataglyphis\u003c/em\u003e ants performing learning walks. During outbound route formation, new ant foragers are often observed performing scanning behaviours in which the forager turns back towards the nest (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), a behaviour called \u0026lsquo;lookbacks\u0026rsquo; (Zeil \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Zeil et al. 2014; Freas et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). During these scans, forward movement ceases and the ants turn in place, orienting their gaze yet often not their body (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB) back towards the nest, learning the nest-aligned view and inbound route during these rotations interspersed with fixations (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB-D). These nest-oriented gaze fixations during outbound travel align with the rapid visual route learning observed in desert ants (\u003cem\u003eMelophorus bagoti\u003c/em\u003e and \u003cem\u003eCataglyphis velox\u003c/em\u003e), with foragers needing only a single outbound route experience (Freas and Cheng \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Freas and Spetch \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Yet the choreography of these lookback-scans and their potential function in route formation is lacking. Here, we characterise the first foraging trips of \u003cem\u003eM. bagoti\u003c/em\u003e foragers as individuals learn to travel between a feeder site and the nest, focusing on route formation, scanning and lookbacks to explore how foragers\u0026rsquo; visual route learning forms and how lookbacks may aid in encoding goal-oriented views to produce rapid visual learning.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cem\u003eStudy Site and Species\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe field site (23°45′28.12″S, 133°52′59.77″E) was located on the grounds of the Centre for Appropriate Technology campus, ~10 km south of Alice Springs, Northern Territory, Australia. \u003cem\u003eMelophorus bagoti\u003c/em\u003e inhabits visually cluttered areas with vegetation (buffel grass and scattered \u003cem\u003eEucalyptus\u003c/em\u003e). Experiments were conducted during the Australian summer (November through February) in 2022–2023.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eForaging arena\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eA plastic feeder (15 × 15 × 10 cm) was sunk into the ground 7m from the nest entrance and baited with both cookie pieces (Arnott™) and mealworm pieces. The smooth walls of the feeder prohibited escape after food collection, allowing experimenters to manually start the inbound journey by lifting foragers to the feeder ledge. Between the feeder and nest entrance, all vegetation was cleared, and a 15-cm-high plastic wall was erected, forming a ~3m wide arena (Figure 2A,B). This arena extended 5m, where a 15cm high barrier prevented further movement to the feeder except through a funnel comprising two (61cm H × 91.5cm L) white sheets of Masonite board with a 30cm gap between them (Figure 2C,E). This greatly reduced the amount of natural food resources the nest could access, motivating foragers to search the area until finding the feeder (Figure 2D,E).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eExperiment 1: Route Formation\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAll foragers emerging from the nest entrance were marked on the gaster with enamel paint (Tamiya™) for five consecutive days, denoting them as experienced individuals and allowing us to exclude them from testing. On the sixth day, when non-painted naïve individuals first left the nest to complete a pre-foraging learning walk, they were collected and individually marked using a colour combination of paint (Tamiya™) on their gaster and/or thorax. Marked individuals were observed during subsequent learning walks and upon their first foraging trip. The end of learning walks and onset of foraging was determined by the distinct lack of a learning walk with alternating nestward/outward scanning behaviours. \u003cem\u003eMelophorus bagoti\u003c/em\u003e exhibits highly structured learning walk scans (akin to pirouettes in \u003cem\u003eCataglyphis\u003c/em\u003e), rhythmically alternating scanning bouts towards and away from the nest with clear nestward and outward fixations (See Supplemental Materials Video 1, Figure 1). Upon starting foraging, their outbound foraging paths to collect food and inbound returns to the nest were recorded via pencil and graph paper. We additionally marked the location along the path when foragers performed scanning behaviours and lookbacks. After the forager collected food, we also recorded the inbound path from the feeder back to the nest, marking the location of inbound scans as well as lookbacks where the forager turned back to the feeder direction.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe defined scanning behaviour as the stopping of forward movement accompanied by rotational body movement on the spot (interrupted by fixations), while remaining oriented outward away from the nest. Lookbacks are structurally somewhat similar to scans (in that they entail a stopping of forward movement and rotational movement interspersed with fixations); we differentiated these from scanning rotations by constricting a ‘lookback’ to only occasions when rotational movement brought the individual’s orientation to within ±90° of the nest (during an outbound lookback) or feeder direction (during an inbound lookback).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;For each individual, we collected positional data of the path as well as any scans and lookbacks observed on the first 10 foraging trips conducted on the first day of foraging. To determine if route formation metrics changed after the overnight delay period when the foraging force is not active, we continued to collect these metrics on each forager’s second day of foraging for the first five trips, also noting if each forager conducted any pre-foraging re-learning walks around the nest prior to second day foraging. After completing this 5\u003csup\u003eth\u003c/sup\u003e next-day trip, ants were marked with paint as tested. Before testing, we determined that ten foraging trips would be sufficient for a forager to become highly experienced of the route (Freas and Cheng 2018; Freas and Spetch 2019). If a forager conducted less than ten foraging trips on the first day, we continued to collect this individual’s metrics on subsequent days until they reached ten total foraging trips and then recorded their first five foraging trips on the following morning (four individuals took two days to complete their first 10 foraging trips). If a forager continued to forage after ten day-one trips, we continued to collect these paths but given these paths were highly similar to Trip 10 by every metric, they have not been used in our analysis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u003cem\u003eRoute Formation Data Analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003ePathswere digitized bymarkingevery ~10cm along the path length while both scans and lookbacks counts were recorded where they occurred along the path\u003cem\u003e.\u0026nbsp;\u003c/em\u003eWe assessed three metrics of path formation for both the outbound and inbound portions of the foraging route: 1.) \u003cem\u003epath straightness\u003c/em\u003e, denoting the efficiency of the outbound/inbound journey between the nest and feeder; 2.) outbound/inbound \u003cem\u003eScan\u003c/em\u003e counts, associated with navigational uncertainty; and 3.) outbound/inbound \u003cem\u003eLookback\u0026nbsp;\u003c/em\u003ecounts, associated with route formation and learning. Path straightness was calculated by the ratio between the straight-line distance (nest-feeder) with the forager’s path length. All metrics were compared across trips over two days (fixed effects) using poisson loglinear General Linear Mixed Models (GLMMs) for count data and Gaussian loglinear GLMs for path straightness data with individuals as a random effect. When significant effects were uncovered, a priori within-individual contrasts akin to Helmert contrasts (with α set at 0.01 to correct for multiple non-independent comparisons) were conducted between each trip and the mean of the following trips that day (i.e. Trip 1 vs Trips 2-10; Trip 2 vs Trips 3-10, etc.). When the comparisons became insignificant, we denoted this as the within-day performance asymptote. Contrasts analysing the overnight performance decline were assessed by comparing Trip 1 of the next day with the average performance on the final three trips of the first day (Trips 8-10). Next-day performance comparisons were conducted with Helmert-style contrasts in a parallel fashion to the first foraging day (Next Day Trip 1 vs. Next Day Trips 2-5, etc.), with performance asymptote similarly determined (when \u003cem\u003ep\u003c/em\u003e \u0026gt; 0.01). Additionally, within individual comparisons between performance on the outbound and inbound routes of the same trip were compared using Wilcoxon Signed Rank Tests.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;We separately analysed the effect of the overnight delay on route formation and maintenance by comparing the mean change (Δ) within each individual in each of the three path-based metrics separated into the outbound/inbound routes (path straightness, scans and lookbacks) between trips characterised by whether they followed a trip that occurred on the \u003cem\u003eSame Day\u003c/em\u003e (Trips 2–10 and Next Day Trips 2–5) or after the ~16h \u003cem\u003eOvernight Delay\u003c/em\u003e (Next Day Trip 1). For the four ants which took two days to complete the first ten trips, their first trip of day two was averaged with the Next Day Trip 1 (foraging day three for these individuals). Within-individual comparisons between \u003cem\u003eSame Day\u003c/em\u003e and \u003cem\u003eOvernight Delay\u003c/em\u003e trips were compared using Wilcoxon Tests.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eExperiment 1: Lookback Positions\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eImage analysis of route\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo assess potential mechanisms underlying when foragers decide to lookback, we chose to characterise how the visual scene changed as foragers moved between the nest and feeder, looking at three factors: ‘local’ panorama change, ‘global’ panorama change, and the ant’s current global vector length (which is not based on views). We quantified the rate of local panorama change by assessing the panorama change over the last 50cm of the route and more globally by comparing visual scenes across the arena to the panorama at the nest entrance. To accomplish this, we collected 360° panoramic images of the foraging route for image analysis. These images were taken at the nest entrance and at 50cm intervals in a grid-like fashion along the arena leading up to and including the feeder panorama, totalling 72 images throughout the foraging route (Figure 2B-D, 6A).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;To calculate a rate panoramic change along the foraging route, we compared each panorama image within the arena with both the nest image for a metric of how the panorama has changed globally from the beginning of the foraging trip (nest), as well as comparing each image to the previous image 50cm before along the grid’s straight-line route to the nest. This gives an indication of how much the visual scene has changed recently as the ant travels to the feeder. As a final metric of change, we also calculated the vector distance (cm) from the nest for each image site. Outbound lookback counts were assigned to the closest image site within the arena (within 25cm). This allowed us to assess the panorama characteristics at sites associated with the number of lookback behaviours which occurred in the vicinity throughout the first 15 trips of the foragers’ careers. Lookbacks that occurred within 25cm of the nest entrance were excluded from this analysis as we could not assess the local or global panorama changes under this distance from the start.\u003c/p\u003e\n\u003cp\u003eTo calculate the amount of visual change between each of the image pairs, rotational image difference functions (rotIDFs) were conducted by using the mean pixel difference (MPD) between the focal image and comparison images for all possible rotations of the test images (in one-degree steps) using custom written scripts in MATLAB (for further details, see Zeil et al. 2003, 2014; Stürzl and Zeil 2007). Each image was down sampled to 1 pixel per degree with final images measuring 360 pixels width × 180 pixels height and converted to the blue colour channel only. When two images contain some degree of familiarity, this produces a rotIDF that contains a clear valley of mismatch or low Mean Pixel Difference (MPD) between the two panoramas highly associated with the direction of translation between the initial image and the image of the updated location, indicating a level of similarity the navigator can use to orient.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo calculate a similarity metric corresponding to the degree to which the two panoramas were similar, we took the difference between the minimum and mean MPD of each rotIDF comparison. This created a ‘panorama change’ metric where high values corresponded with a deep valley of mismatch between panoramas, denoting a large degree of similarity and low levels of panorama change across space. In contrast, low values corresponded with a shallow valley of mismatch between panoramas in the rotIDF, denoting a low degree of similarity between the images and more rapid panorama change over space. A multiple linear regression analysis was conducted to examine the effect of the \u003cem\u003eLocal Panorama Similarity\u0026nbsp;\u003c/em\u003e(last 50cm), \u003cem\u003eGlobal Panorama Similarity\u003c/em\u003e and \u003cem\u003eVector Length\u003c/em\u003e on the location of Lookbacks during the outbound portion of the paths.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eExperiment 2: Lookbacks Choreography\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;After observing individuals during the route formation experiment, we chose two locations along the route at which to record lookback behaviours for gaze direction analysis; the first ~80cm from the nest entrance and at the funnel area at 5–6m from the nest, given these two locations were associated with a high number of lookback counts during route formation testing. At these sites we spread white sand over the red soil over a 1m x 1m area in order to more clearly observe the ant’s body/head positions on video frames. Above each site we positioned a Sony Handycam 70cm above the ground, with the lense facing down (3840 x 2160, 25fps). A forager which had already been to the feeder at least once generally left the nest in the feeder directional quadrant where the camera’s FOV was positioned. As these foragers left the nest they were recorded until they left the camera frame, capturing any lookback behaviours occurring in this area. Once they left the filming area, each forager was collected and marked as tested to prevent refilming on subsequent trips.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Ants’ gaze directions and positional estimates were manually extracted from the videos using SLEAP animal positioning software (Pereira et al. 2022). Two points along the ant were selected to assess gaze direction during lookback behaviours, the centre of the mouth/front of the head and the head base, where the head and body meet. There two points allowed us to assess gaze direction of the animal and directional and positional fixation phases beyond the body’s orientation, essential given the large degree of body contraction we observed during lookbacks (~55° divergence between gaze and body orientations, See Figure 1A fixation 6). We began tracking the animal’s pose estimates five frames before the ceasing of forward movement associated with the start of the lookback scan and pose estimates were stopped after the ant resumed forward travel to the goal (feeder/nest) for five frames.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhen assessing the structure of lookback scans, we collected the number, direction and duration of fixations throughout the lookback, determining the longest fixation duration, and its direction in relation both to the direction of travel prior to stopping and the nest/feeder direction. We additionally chose to analyse the direction at which the inflection or reversal point of the turn occurs (when the individual switches from turning left to right).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLookback Gaze Analysis\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eDirectional gaze-fixation data was analysed using circular statistics. To determine whether the longest gaze fixation and the reversal direction were directed, we first used Rayleigh tests for circular data (Fisher 2993). If headings were directed, we further analysed whether the mean direction of the longest gaze fixation was in the nest/feeder direction and the opposite (180°) of travel using the 95% confidence interval (CI) around the mean of longest gaze fixations. 95% CIs were calculated through the standard error of the mean heading direction based on the mean vector length (Fisher 2993). Within-individual directional comparisons between longest gaze fixation and reversal direction were conducted using Moore’s Paired Tests.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Comparisons between the two outbound conditions, Near Nest, Funnel and the Inbound lookback conditions were conducted using poisson loglinear General Linear Models (GLM) for count data, with individuals as a random effect. If there was a significant effect of condition, Dunn-Bonferroni Post-hoc pairwise comparisons were conducted comparing conditions. Pairwise comparisons within conditions (absolute angular error comparisons and pre vs. post reversal comparisons) were compared using Wilcoxon tests.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cem\u003eRoute Formation \u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eNaïve foragers travelling away from the nest area for the first time exhibited highly tortuous outbound paths to the feeder (path straightness μ ± Std. = 0.30 ± 0.07; Figure 3, 4A) interrupted by multiple bouts of scanning (μ ± Std. = 11.07 ± 5.39; Figure 3, 4C) and a large number of lookbacks in the direction of the nest (μ ± Std. = 7.47 ± 0.29; Figure 3, 4E), consistent with high uncertainty beyond the nest area, until each individual found the feeder and collected food (Figure 3, 4A). In contrast, the inbound portion of Trip 1, returning to the nest with food, was significantly (Wilcoxon Signed Rank Test; \u003cem\u003eZ \u003c/em\u003e= 4.01; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001) straighter (path straightness μ ± Std. = 0.88 ± 0.06), with significantly fewer scanning bouts (μ ± Std. = 3.20 ± 2.01; Wilcoxon Test; \u003cem\u003eZ \u003c/em\u003e= –3.3; \u003cem\u003ep\u003c/em\u003e = 0.001) and lookbacks to the feeder compared to the outbound portion of the same trip (μ ± Std. = 0.33 ± 0.49; Wilcoxon Test; \u003cem\u003eZ \u003c/em\u003e= –3.4; \u003cem\u003ep\u003c/em\u003e = 0.001: Figure 3, 3). Inbound lookbacks were remarkably rare, with only 28 observed instances across all 225 inbound routes (μ ± Std. = 0.12 ± 1.55; Figure 3, 4F).\u003c/p\u003e\n\u003cp\u003eThe pattern of significantly lower path straightness coupled with higher counts of both scanning bouts and lookbacks during the outbound portion (vs. inbound) persisted through the first six trips, with path straightness becoming not significantly different from its inbound counterpart at Trip 7 (Wilcoxon Test; \u003cem\u003eZ \u003c/em\u003e= –1.93; \u003cem\u003ep\u003c/em\u003e = 0.053; Figure 4AB), scanning bouts non-significant at Trip 9 (\u003cem\u003eZ \u003c/em\u003e= –0.21; \u003cem\u003ep\u003c/em\u003e = 0.831; Figure 4CD), and lookbacks non-significant at Trip 8 (\u003cem\u003eZ \u003c/em\u003e= –1.29; \u003cem\u003ep\u003c/em\u003e = 0.197; Figure 4EF). \u003c/p\u003e\n\u003cp\u003eAt the beginning of their second day of foraging, a majority of individuals (53%) conducted a number of ‘re-learning walks’ around the nest before beginning to forage and these walks appeared similar to those occurring prior to initial route formation. Of the eight foragers which conducted re-learning walks, four conducted one learning walk while another four conducted two walks prior to restarting foraging. Once foraging behaviour resumed on the next day, the same pattern of significantly lower outbound (vs. inbound) straightness and higher lookback counts was observed on Next Day Trip 1 and 2 (\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.027) before becoming nonsignificant (\u003cem\u003ep\u003c/em\u003e \u0026gt; 0.05) on Next Day Trip 3 (Scan count was significantly higher on outbound journeys until Next Day Trip 4; Figure 4). \u003c/p\u003e\n\u003cp\u003e\u003cem\u003eOutbound route\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eBoth \u003cem\u003eTrip\u003c/em\u003e and \u003cem\u003eDelay\u003c/em\u003e had significant effects on outbound path straightness (Trip: \u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 21.88; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001; Delay: \u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 15.39; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001), scans (Trip: \u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 24.38; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001; Delay: \u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 21.76; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001) and lookbacks (Trip: \u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 28.59; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001; Delay: \u003cem\u003eχ2\u003c/em\u003e = 37.91; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001) and there was only a significant interaction between \u003cem\u003eTrip\u003c/em\u003e \u0026amp; \u003cem\u003eDelay\u003c/em\u003e (\u003cem\u003eχ2\u003c/em\u003e = 8.08; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01) on scan count. Helmert contrast comparisons showed that outbound route performance reached asymptote (Table S1) on Trip 3 (scan count), Trip 5 (lookback count) and Trip 6 (path straightness). There was a significant decrease in performance across all metrics (\u003cem\u003ep \u0026lt; \u003c/em\u003e0.001; Table S1) after the overnight delay (Trips 8-10 vs. Next Day Trip 1). During next day foraging, Helmert contrast comparisons showed that outbound route performance returned to asymptote by Next Day Trip 3 in all metrics.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eInbound route\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eTrip\u003c/em\u003e (\u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 10.74; \u003cem\u003ep\u003c/em\u003e = 0.001) but not overnight \u003cem\u003eDelay \u003c/em\u003e(\u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 2.85; \u003cem\u003ep\u003c/em\u003e = 0.09) had a significant effect on forager inbound path straightness, while both \u003cem\u003eTrip \u003c/em\u003e(\u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 17.67; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001) and \u003cem\u003eDelay \u003c/em\u003e(\u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 5.45; \u003cem\u003ep\u003c/em\u003e = 0.02) had significant effects on scan counts and there was not a significant interaction between factors (\u003cem\u003ep\u003c/em\u003e = 0.540). There was no effect of \u003cem\u003eTrip \u003c/em\u003e(\u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 2.28; \u003cem\u003ep\u003c/em\u003e = 0.131) and \u003cem\u003eDelay\u003c/em\u003e (\u003cem\u003eχ\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 0.23; \u003cem\u003ep\u003c/em\u003e = 0.634) on inbound lookback counts. Helmert contrast comparisons showed that inbound route performance was already at asymptote (Table S1) on Trip 1 for path straightness (\u003cem\u003ep\u003c/em\u003e = 0.022) but scan count showed significant contrasts until Trip 3, after which no further improvements were found. Additionally, scan count did fall below asymptote during Trip 5 (\u003cem\u003ep\u003c/em\u003e = 0.002). There was no significant decrease in performance after the overnight delay (Trips 8-10 vs. Next Day Trip 1). During Next Day foraging, inbound route performance improved and returned to asymptote by Next Day Trip 2 (Table S1).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eOvernight delay\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAn overnight delay period had a clear effect on forager performance. When comparing the mean change within individuals between all Trips which followed a same-day trip and those which followed the overnight delay, these delays corresponded with significant decreases in outbound path straightness (Wilcoxon-Test; \u003cem\u003eZ = –\u003c/em\u003e3.41; \u003cem\u003ep = \u003c/em\u003e0.001; Figure 5A) and significant increases in both outbound lookback (\u003cem\u003ep = \u003c/em\u003e0.001; Figure 5B) and scan (\u003cem\u003ep = \u003c/em\u003e0.001; Figure 5C) counts. A similar pattern emerged with inbound path characteristics with significant differences in mean performance on trips after same-day trips or after overnight delay trips, with significant decreases in path straightness (Wilcoxon Test; \u003cem\u003eZ \u003c/em\u003e= –3.41; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001; Figure 5A) and significant increases in scan counts (Wilcoxon Test; –3.41; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001; Figure 5B).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLookback Spatial Positions\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLocal Panorama Similarity \u003c/em\u003e(last 50cm), \u003cem\u003eGlobal Panorama Similarity\u003c/em\u003e and \u003cem\u003eVector Length\u003c/em\u003e explained 11.92% of lookback variance across the arena and this effect was significant (\u003cem\u003eF\u003c/em\u003e=3.02, \u003cem\u003ep\u003c/em\u003e = .035, \u003cem\u003eR\u003c/em\u003e\u003cem\u003e\u003csup\u003e2\u003c/sup\u003e\u003c/em\u003e = 0.12). Only \u003cem\u003eLocal Panorama Similarity \u003c/em\u003ewas shown to impact the number of lookbacks which occurred spatially within the arena, with an inverse relationship of lower panorama similarity over the last 50cm being associated with a higher number of lookbacks occurring in this area (\u003cem\u003ep =\u003c/em\u003e 0.005). Both \u003cem\u003eGlobal Panorama Similarity\u003c/em\u003e (\u003cem\u003ep\u003c/em\u003e = 0.234)and \u003cem\u003eVector Length \u003c/em\u003e(\u003cem\u003ep\u003c/em\u003e = 0.245)were not significantly associated with lookback numbers. \u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLookback chorography\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThree types of lookbacks were observed and recorded during testing. The first and most commonly observed was defined as the 180° turn in place (Figure 1A,C). Here, foragers cease forward movement and turn back towards the nest direction before reversing their turn direction. In a minority (~7%) of lookbacks, this reversal does not occur, and foragers continue to turn in a circle before continuing forward movement. We have defined these lookbacks as 360° circular turns (Figure 1D), and given their lack of reversals, they were not included in any pre- and post-reversal fixation direction and duration analysis. The final type of lookback which was observed consisted of the forager hooking back to the nest direction, with the turn accompanied by spurts of forward movement during the turn, typically back in the direction of the nest (see Figure 1E). All of these lookbacks consisted of multiple fixations scattered throughout the turn. \u003c/p\u003e\n\u003cp\u003eThe longest fixation period during each lookback in both outbound conditions (nest: μ ± Std. = 189 ± 82ms; funnel: μ ± Std. = 210 ± 107ms), was towards the nest entrance. The longest of these fixations were significantly directed (nest: Rayleigh’s Test, \u003cem\u003eZ \u003c/em\u003e= 20.63; \u003cem\u003ep \u003c/em\u003e\u0026lt; 0.001; funnel: Rayleigh’s Test, \u003cem\u003eZ \u003c/em\u003e= 14.78; \u003cem\u003ep \u003c/em\u003e\u0026lt; 0.001; Figure 7A) and the nest direction (0°) was within the 95%CI of both nest (μ ± 95%CI = 8.12° ± 9.84°) and funnel (μ ± 95%CI = 8.12° ± 9.84°) lookbacks. Reversal directions in both conditions were also significantly directed (nest: Rayleigh’s Test, \u003cem\u003eZ \u003c/em\u003e= 19.20; \u003cem\u003ep \u003c/em\u003e\u0026lt; 0.001; funnel: Rayleigh’s Test, \u003cem\u003eZ \u003c/em\u003e= 12.74; \u003cem\u003ep \u003c/em\u003e\u0026lt; 0.001; Figure 7A) with the nest direction (0°) within the 95%CI (nest: μ ± 95%CI = 5.71° ± 9.94°; funnel: μ ± 95%CI = 5.02° ± 16.14°). Inbound lookbacks mirrored outbound lookbacks with a longest gaze fixation period (μ ± Std. = 182 ± 120ms) that was towards the feeder direction. These gaze directions were significantly directed (Rayleigh’s Test, \u003cem\u003eZ \u003c/em\u003e= 5.02; \u003cem\u003ep \u003c/em\u003e= 0.004; Figure 7A) and the feeder direction (180°) was within the 95%CI (μ ± 95%CI = 202.56° ± 34.01°).\u003c/p\u003e\n\u003cp\u003eThe longest fixation direction and the reversal direction aligned in 51% (26 of 51) of analysed lookbacks and differed on average by ~10° across all conditions (Near Nest: 10.03°; Funnel: 10.32°; Inbound: 11.67°). Within-individual comparisons between longest gaze fixation and the reversal direction did not significantly differ in all three conditions (Moore’s Paired Tests; Near Nest: \u003cem\u003eR’ = \u003c/em\u003e0.278\u003cem\u003e; p \u0026gt; \u003c/em\u003e0.50; Funnel: \u003cem\u003eR’ = \u003c/em\u003e0.219\u003cem\u003e; p \u0026gt; \u003c/em\u003e0.50; Inbound: \u003cem\u003eR’ = \u003c/em\u003e0.897\u003cem\u003e; p \u0026gt; \u003c/em\u003e0.10).\u003c/p\u003e\n\u003cp\u003eDuring inbound lookbacks, the longest gaze fixation was not significantly directed in relation to the opposite of travel direction (Rayleigh’s Test, \u003cem\u003eZ \u003c/em\u003e= 1.68; \u003cem\u003ep \u003c/em\u003e= 0.19). There was a significant directedness in both outbound conditions (Rayleigh’s Tests, \u003cem\u003ep \u003c/em\u003e\u0026lt; 0.05). Within-individual comparisons showed that the longest-gaze-fixation direction and the reversal direction did not significantly differ in all three conditions (Moore’s Paired Tests; Near Nest: \u003cem\u003eR’ = \u003c/em\u003e0.278\u003cem\u003e; p = \u003c/em\u003e0.790; Funnel: \u003cem\u003eR’ = \u003c/em\u003e0.219\u003cem\u003e; p = \u003c/em\u003e0.710; Inbound: \u003cem\u003eR’ = \u003c/em\u003e0.897\u003cem\u003e; p = \u003c/em\u003e0.340).\u003c/p\u003e\n\u003cp\u003eAs circular statistics will generally achieve significance with data grouped in the correct hemisphere (all lookbacks by our classification sytem had to break the ±90° travel direction plane), when within-individual comparisons were significantly different, we compared the within-individual absolute angular error of the longest gaze fixation with both the nest direction and 180° of travel direction. In all three conditions absolute angular error was significantly lower against the nest direction versus the 180° of travel direction (Wilcoxon Tests, Near Nest: \u003cem\u003eZ \u003c/em\u003e= –4.13, \u003cem\u003ep \u003c/em\u003e\u0026lt; 0.001; Funnel: \u003cem\u003eZ \u003c/em\u003e= –3.4, \u003cem\u003ep \u003c/em\u003e= 0.001; Inbound: \u003cem\u003eZ \u003c/em\u003e= –2.1, \u003cem\u003ep \u003c/em\u003e= 0.036; Figure 7A,B), indicating foragers fixated closer directionally to the nest direction than to the inverse of their outbound or inbound path in all conditions.\u003c/p\u003e\n\u003cp\u003eWe found no significant difference between conditions in mean gaze fixation duration, longest gaze fixation duration (\u003cem\u003ep \u003c/em\u003e\u0026gt; 0.05; Figure 8A,B), but did find a significant effect of condition on the number of fixations, with inbound lookbacks exhibiting significantly fewer fixations compared to both Near Nest and Funnel outbound conditions (Figure 8C).\u003c/p\u003e\n\u003cp\u003eIn all three conditions, we found no significant within-individual differences in the mean gaze fixation duration before and after the reversal (Wilcoxon Tests; \u003cem\u003ep \u0026gt; \u003c/em\u003e0.05). In both the Near Nest outbound and Inbound conditions, we found a significant decrease in the number of fixations after the reversal compared to prior to the reversal of the turn although this pattern was not present in the Funnel condition.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eDesert ants (\u003cem\u003eMelophorus\u003c/em\u003e and \u003cem\u003eCataglyphis\u003c/em\u003e) are known to rapidly learn visual scenes, and to successfully orient and efficiently navigate back home after only a single experience of the outbound route, even when without a corresponding vector (Freas and Cheng \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Freas and Spetch \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Thus, this outbound portion of the first trip of the route must be a critical period for view learning to allow for rapid inbound route formation, yet the mechanisms by which ants accomplish this are unclear. Here, we show that lookbacks constitute a critical learning period of the inbound nest-aligned scene for the return trip. This behaviour includes the same nest-directed gaze fixation that is present in learning walk pirouettes in \u003cem\u003eCataglyphis\u003c/em\u003e desert ants, behaviours which are highly associated with view learning (Grob et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Thus, nestward rotations and their accompanying gaze fixations underly view learning both around the nest, during learning walk pirouettes, and along the foraging route, during lookbacks. Furthermore, the increased tendency for foragers to engage in lookbacks when traveling through areas of high local panorama change suggests that panorama change may be a triggering mechanism, likely through increased navigational uncertainty, however that is encoded in an ant\u0026rsquo;s brain. This would allow lookbacks to be performed when the view becomes novel, encouraging the learning of new, unfamiliar route views as the forager first experiences them via panorama change.\u003c/p\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eRoute formation\u003c/h2\u003e \u003cp\u003eInbound route formation generally aligned with previous work (Freas and Cheng \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), showing single-experience learning of the inbound route after an outbound route exposure, though importantly, previous work did not definitively establish this occurred on a single outbound trip rather than after multiple unsuccessful foraging trips before reaching a feeder. Individuals were well oriented returning home, with high path straightness (\u0026micro;\u0026thinsp;\u0026plusmn;\u0026thinsp;Std. = 0.88\u0026thinsp;\u0026plusmn;\u0026thinsp;0.06) at asymtope even on Trip 1, with only a single outbound experience, suggesting sufficient learning had already occurred during Trip 1\u0026rsquo;s outbound search to the feeder, though note that these ants did have a corresponding global vector. Foragers exhibited non-negligible scan counts until Trip 4, suggesting further learning and performance improvements over the first few outbound trips (inbound lookbacks were rare). In contrast, the outbound route needed more exposures to complete its route formation, with performance metrics taking until Trips 3\u0026ndash;6 to plateau.\u003c/p\u003e \u003cp\u003eThe ~\u0026thinsp;16h overnight delay had a clear effect on route formation on both the outbound and inbound path portions, with decreases in path straightness and increases in scan and lookback counts. Coupled with this, a slight majority of foragers also conducted at least one re-learning walk around the nest on the morning of the second day of their foraging career. This suggests that both learning walks and outbound lookbacks also have a function in maintaining established routes, likely to continuously assess if the visual scene has changed. Many aspects of the visual panorama look different from the last trip of the day, typically occurring in the late afternoon (4\u0026ndash;5pm) to the first trip of the next day the following morning (~\u0026thinsp;9am), including shadows and light reflectivity off terrestrial cues (Wehner \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e1997\u003c/span\u003e). Given that these aspects of the scene change throughout the day and overnight, some level of view memory maintenance through lookbacks and relearning walks may help support route formation over multiple days.\u003c/p\u003e \u003cp\u003eSome fading or decay of memory might have also contributed to the deterioration of navigational performance overnight. One study on the acquisition of landmark-based homing to the nest improved after 2 days of training, that is, performance on day 3 was better than on day 1 or on day 2 (Narendra et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). That study provided a richer array of landmarks near the nest, with the nest being surrounded by 4 black cylinders. The number of training trials was not a critical factor in acquisition, but spreading training over multiple days benefitted the ants. After 2 days of training, landmark memories seemed to last a lifetime, with the ants performing well even after 8 days of delay. This study leaves open the possibility that landmark memory, what is now more commonly called view memory, might have faded after the first day of foraging, requiring lookbacks on the next day of foraging to produce long-term route formation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eLocal panorama change\u003c/h2\u003e \u003cp\u003eWe found that lookback counts on the outbound route are associated with spatial locations where there is a high degree of local panorama change, measured by comparing the scene to the view 50cm earlier in the route. Given that learning-walk-acquired view memories of the nest area are only useful within a catchment area, foragers should be able to modulate when to perform lookbacks based upon visual panorama metrics of change. This would allow foragers to acquire new route-based nest-aligned view memories to help return home. It remains unclear if rapid panorama change is a triggering mechanism for foragers to lookback or if this change is also typically associated with a spike in unfamiliarity, with unfamiliarity triggering the behaviour. Regardless of the mechanism, looking back to the nest in areas where the panorama changes quickly would align with this behaviour having a view-learning function. Future work could characterise if lookbacks can be predictably triggered in specific locations in response to exposure to unfamiliar visual cues along known routes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eNest-aligned gaze fixations\u003c/h2\u003e \u003cp\u003eIn the current study, outbound lookbacks contained multiple gaze fixations with the longest of these fixations being grouped in the nest direction. This makes lookbacks structurally akin to learning-walk-based pirouettes around the nest, and both likely serve the same function of acquiring nest-aligned views for inbound route formation (Grob et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Nest-aligned gaze fixations were present at both outbound recording locations, near the nest and at the funnel. Across all recorded metrics of the lookback (including mean and longest fixation durations and number of fixations), we found no significant difference between these sites, suggesting the choreography of the outbound lookback is maintained across the route. Inbound lookbacks, in addition to containing a longest gaze fixation back towards the feeder instead of the nest, also showed significantly fewer gaze fixations compared to outbound lookbacks. Given these lookbacks were similar in every other metric besides their frequency, it is hard to untangle why this might be the case, though there likely is an overall lower level of uncertainty during these trips given the global vector and recency of the preceding outbound trip, which might reduce fixation numbers.\u003c/p\u003e \u003cp\u003eNest-aligned gaze directions and fixations appear to be closely tied to navigational view learning throughout a forager\u0026rsquo;s above-ground life, both before and after the onset of foraging. The views around the nest are known to be acquired during pre-foraging learning walks. Within these walks, na\u0026iuml;ve \u003cem\u003eCataglyphis\u003c/em\u003e (\u003cem\u003eC. noda\u003c/em\u003e and \u003cem\u003eC. aenescens\u003c/em\u003e) ants inhabiting cluttered environments exhibit highly choreographed turns during their pre-foraging learning walks, in the form of pirouettes, conducting either 360\u0026deg; or 180\u0026deg; turns. Scattered throughout this rotation are gaze fixations with the longest fixation in the direction of the nest, presumably to acquire this nest-aligned view (Grob et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). This behaviour seems dependent on the visual scene, with a species inhabiting barren habitat (\u003cem\u003eCataglyphis fortis\u003c/em\u003e) facing the nest only briefly and without fixations (called \u0026ldquo;voltes\u0026rdquo;, Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), while pirouette-like behaviours are documented in other ants within cluttered habitats (Wystrach et al. 2014; Jayatilaka et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). \u003cem\u003eM. bagoti\u0026rsquo;s\u003c/em\u003e learning walks also appear to exhibit turns with distinct nest-oriented gaze fixations (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). However, a comprehensive study of learning-walk pirouette gaze fixations in \u003cem\u003eM. bagoti\u003c/em\u003e is currently absent.\u003c/p\u003e \u003cp\u003eGiven that nest-ward gaze fixations are highly associated with view acquisition during nest panorama learning, the frequency, location, and composition of lookbacks along the route all align with these behaviours being akin to learning-walk pirouettes, forming view-acquisition periods to aid foragers in acquiring the nest-aligned views of foraging routes for the ensuing inbound trips (Graham and Collett 2006; Freas and Cheng \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Freas and Spetch \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Our description of \u0026lsquo;hook turn\u0026rsquo; lookbacks bears some similarity to the path reversals and looking back in wood ants, \u003cem\u003eFormica rufa\u003c/em\u003e (Graham and Collett 2006). Wood ants exhihit similar inbound learning during the outbound trip, which Graham and Collett (2006) ascribed to \u0026lsquo;reversal\u0026rsquo; periods in which the outbound ant turned back and travelled towards the nest for a short distance before resuming its travel to the feeder. They also report that these behaviours occurred early in route formation, though they make no mention of any associated fixations or stopping periods. This similarity suggests that looking back is a consistent learning strategy for route formation in ants.\u003c/p\u003e \u003cp\u003eWhile nest-aligned fixations are well established during learning walks, a key difference between gaze fixations during route-based lookbacks and pirouettes is the individual\u0026rsquo;s final goal location (being the feeder during outbound lookbacks and back to the nest during learning walks). Since during route-based lookbacks, the forager has some experience of the route and likely a long-term vector memory of the feeder site, it is worth untangling some possible alternative explanations for the nest-based gaze fixations, all of which we ultimately find unconvincing.\u003c/p\u003e \u003cp\u003eTwo alternative explanations exist for the observed nest-oriented gaze fixations during lookbacks. First, outbound foragers could be rotating 180\u0026deg; from their travel direction, which would typically align with the nest, which means that the longest gaze fixation should be opposite the travel direction and not the nest when these directions are different. Such a non-nest-directed view might be predicted as a forager would want to learn its outbound route even if it conflicts with the nest directionally, for example, if the route has two legs separated by a turn. Yet, in our results, foragers under-rotate to align with the nest when not travelling directly away from the nest, leading to significantly more directional error from 180\u0026deg; of travel direction compared to error from the nest direction (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB). The pattern suggests that the forager is turning until identifying that its gaze is oriented with the nest direction before reversing its turn (for example path see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). Thus, it appears unlikely that these lookbacks are opposite the direction of travel and are instead tuned to stop to fixate at the nest. This theory does, however, pose interesting questions for lookbacks and route formation on non-straight-line routes, where route formation and the nest would not directionally align.\u003c/p\u003e \u003cp\u003eThe second alternative explanation is that the nest direction should typically correspond with the direction opposite of the feeder (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Consequently, when the lookback reaches 180\u0026deg; of the feeder, uncertainty over which way to turn should increase as both the left and right angular distances to the feeder becomes close to equal. This uncertainty should reach its maximum when 180\u0026deg; opposite the feeder and facing the nest and might cause the longest fixation to be towards the nest irrespective of the nest\u0026rsquo;s presence, thus accounting for this longer fixation period. However, we find this unconvincing as the nest is only directionally opposite the feeder along the straight-line route, which foragers often deviate from. When foragers are positioned laterally off this straight nest-to-feeder line, we observed the same under turns in line with fixating at the nest, suggesting that gaze fixations are to the nest and not merely opposite the goal (this under turn can be seen clearly in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA; Supplememtal Video 2). Yet, a part of this hypothesis may help explain why in some instances, foragers complete 360\u0026deg; loop turns. As stated above, angular-distance uncertainty should plateau at 180\u0026deg; opposite the feeder along the straight-line route from nest to feeder, meaning a forager near this line may become more uncertain of which is the most efficient turning direction to return to outbound travel, resulting in foragers sometimes continuing to turn in a loop rather than reversing. In support of this hypothesis, all four observed examples of loops took place when foragers were very close to the straight-line route between nest and feeder, where deciding to loop 360\u0026deg; may be a similar directional distance to reversing back 180\u0026deg;. Thus, lookback instances where foragers are near the nest-feeder line might result in foragers being equally likely to conduct 180\u0026deg; or 360\u0026deg; turns given the equidistant angular difference, a phenomenon evident in one instance of a forager traveling almost straight to the feeder, which conducts two 360\u0026deg; lookbacks followed by one 180\u0026deg; lookback turn (See Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC for first 360\u0026deg; turn; Supplemental Materials Video 3).\u003c/p\u003e \u003cp\u003eThis would suggest that after the forager completes its gaze fixation of the nest, to return to face along its outbound route, it likely initially relies on a food vector memory of the feeder location (Bolek et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Wolf et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Bolek and Wolf \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) to decide which way to turn (akin to backwards facing navigation; Schwarz et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eWhen to learn views?\u003c/h2\u003e \u003cp\u003eThe nest-directed gaze-fixation-duration evidence clearly points to ants aligning to the nest, both during learning walks (Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) and now during route formation. How the lookback and pirouette behaviours\u0026rsquo; structures may promote visual learning is an interesting query, especially given the multiple rotation and fixation intervals which make up each behaviour as well as lookbacks\u0026rsquo; occurrence in association with panorama change. Additionally, lookbacks should occur regardless of whether an outbound search is successful, leaving open the question of how the ant\u0026rsquo;s brain learns the nest-aligned route views acquired during lookbacks.\u003c/p\u003e \u003cp\u003eOur current understanding is that long term view-based memories are stored within the mushroom bodies along with valence based on positive and negative experiences with these views (Ardin et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Buehlmann et al. 2020; Kamhi et al. 2020; Webb and Wystrach \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, 2020; Le Mo\u0026euml;l and Wystrach 2020a; Freas et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). View memories and visual route formation in ants involve memory storage within the ants\u0026rsquo; mushroom bodies (Buehlmann et al. 2020; Kamhi et al. 2020). Individual views of the route can be represented neurally within the mushroom bodies through specific Kenyon cell (KCs) activation patterns. Whether a view memory is associated with the goal location and thus has an attractive valence is accomplished through the projection of these activation patterns onto multiple motor output neurons (MBON), with memory valence controlled by the activation of dopaminergic neurons, which mediates the association between a view memory and the associated valence stemming from negative or positive experiences (Cohn et al. 2015; Aso et al. 2015). However, lookbacks do not follow or precede any positive or negative reinforcement, thus leaving open the question of how the nest-aligned fixations become attractive.\u003c/p\u003e \u003cp\u003eRecent work in the head-direction network of \u003cem\u003eDrosophila\u003c/em\u003e (Fisher et al. 2022) suggests that visual rotation may provide a \u0026lsquo;when to learn signal\u0026rsquo;, thus enhancing latent or \u0026lsquo;unsupervised\u0026rsquo; visual learning during lookbacks. Dopamine is released when a visual cue rotates, thus strengthening visual learning during periods when cue rotation provides a rich stream of visual information. The authors theorise that constrains visual learning to these periods and may protect established visual memories during other periods such as straight-line movement. Given that visual rotation should enhance learning, the observed large rotational changes that occur during lookback turns should be clear periods in which view learning is enhanced. The intermittent fixations may also aid in this function, with short fixations periods allowing for better encoding of a stable view. Learning during the rapid-rotation periods may lead the ant to learn degraded blurry views of the route, thus decreasing navigational performance. Thus, the alternation of rotations and gaze fixations within the lookback behaviour together facilitate view learning and route formation. Furthermore, reliance on dopamine for learning may require gaze fixations to be reasonably short in duration so that the visual-system signals of rotation that promote learning still linger. Longer fixations in contrast might break this coupling, leading to a lack of dopamine presence during fixation. While highspeed video analysis of fixation duration and interval during lookbacks is ongoing, fixation durations in \u003cem\u003eM. bagoti\u003c/em\u003e scanning bouts shows a strong leftward skew indicating many short fixation periods (Deeti et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The coupling of rotation as a \u0026lsquo;when to learn signal\u0026rsquo; and short gaze fixations may explain the left-skewed distribution of fixation durations in relearning walks (Deeti et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) as well as the prevalence of short fixations during lookbacks in this study.\u003c/p\u003e \u003cp\u003eFinally, the benefit of learning the nest-aligned views through rotation and nest centred gaze fixations appears clear, yet it presents an interesting conflict with the modelling of route formation. Modelling has shown that correctly goal-oriented views alone are insufficient to produce efficient route formation and that only when familiarity estimates corresponding to orientations to the left and right of the goal are included does route formation become robust (Wystrach et al. \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2020b\u003c/span\u003e). To bring these ideas into agreement, it is important to note that the nest-aligned fixations, while the longest, represent only one of many fixations that occur throughout the lookback. This suggests that these non-nest fixations within the lookback also represent learning periods, primed by rotation, for views along the route acquired in conjunction with heading representations of goal locations within the central complex (Seelig et al. 2015; Stone et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Despite the heavy focus on the longest nest-aligned gaze fixations when discussing learning walks and flights (Grob et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2018\u003c/span\u003e Robert et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Collett and Hempel de Ibarra 2023; Collett et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Freas et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), non-nest-aligned fixations are also well known to occur during these learning events, just as with our lookbacks. Our speculation is that such non-nest-aligned views become associated with turn signals for turning in order to reach a nest-aligned view, signals that will become crucial for steering homeward later. Wasps have also been shown to rhythmically alternate between fixations to the left and right of a goal site (St\u0026uuml;rzl et al. 2016). \u003cem\u003eCataglyphis\u003c/em\u003e ants as well as \u003cem\u003eM. bagoti\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb) have been shown to rotate between nest and anti-nest directions with fixations in multiple directions in between these reversals (Fleischmann et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Thus, all the fixations observed during these periods of rotation may constitute periods for learning views, critical for both learning the nest views and for route formation via lookbacks.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eTheir frequency at the beginning of route formation, association with local panorama change and structural similarity to learning walk pirouettes all point to lookback behaviours supporting route-based view learning of areas beyond the nest at the beginning of route formation. We show that lookbacks contain multiple gaze fixations of different orientations of route-based views with the longest fixation being associated with the nest direction. This nest-ward fixation likely serves as a critical period of learning the inbound route to allow the forager to quickly acquire the visual information needed to return home.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTrevor Murray helped with creating code to run batch rotIDF comparisons. We thank him for his assistance.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFunding\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis project was funded by a Macquarie University Research Fellowship (MQRF0001094), by Macquarie University, and by an ARC Discovery Grant (DP200102337).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLand Acknowledgment\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis work was conducted upon the grounds of the Centre for Appropriate Technology whom we thank for access to the nest. This work was conducted upon land traditionally owned by the Arrernte. We acknowledge the traditional custodians of the land on which our field site sits. Their culture and customs have nurtured and sustained this land since the Dreamtime and continue to do so. We pay our respects to their Elders, past, present, and future.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eConflicts of interest\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflicts of interest associated with this work.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eEthics\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThere are no state or federal governmental regulations guiding the research of invertebrates in Australia.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eArdin, P. Peng, F. Mangan, M. Lagogiannis, K. \u0026amp; Webb, B. (2016). Using an insect mushroom body circuit to encode route memory in complex natural environments. PLoS computational biology, 12(2), e1004683.\u003c/li\u003e\n \u003cli\u003eBaddeley, B. Graham, P. Philippides, A. \u0026amp; Husbands, P. (2011). Models of visually guided routes in ants: Embodiment simplifies route acquisition. In S. Jeschke, H. Liu, D. Schilberg (Eds.), Proceedings of the international conference on intelligent robotics and applications (ICIRA) part II, lecture notes in artificial intelligence (pp. 75\u0026ndash;84). Heidelberg, Germany: Springer-Verlag.\u003c/li\u003e\n \u003cli\u003eBaddeley, B. Graham, P. Husbands, P. \u0026amp; Philippides, A. (2012). A model of ant route navigation driven by scene familiarity. PLoS computational biology, 8(1), e1002336.\u003c/li\u003e\n \u003cli\u003eBolek, S. \u0026amp; Wolf, H. (2015). Food searches and guiding structures in North African desert ants, Cataglyphis. Journal of Comparative Physiology A, 201, 631-644.\u003c/li\u003e\n \u003cli\u003eBolek, S. Wittlinger, M. \u0026amp; Wolf, H. (2012). Establishing food site vectors in desert ants. Journal of Experimental Biology, 215(4), 653-656.\u003c/li\u003e\n \u003cli\u003eCheng, K. \u0026amp; Freas, C. A. (2015). Path integration, views, search, and matched filters: the contributions of R\u0026uuml;diger Wehner to the study of orientation and navigation. Journal of Comparative Physiology A, 201, 517-532.\u003c/li\u003e\n \u003cli\u003eCheng, K. Schultheiss, P. Schwarz, S. Wystrach, A. \u0026amp; Wehner, R. (2014). Beginnings of a synthetic approach to desert ant navigation. Behavioural processes, 102, 51-61.\u003c/li\u003e\n \u003cli\u003eCollett TS, Collett M 2000. Path integration in insects. Current Opinion in Neurobiology 10:757\u0026ndash;762.\u003c/li\u003e\n \u003cli\u003eCollett, T. S. \u0026amp; Hempel de Ibarra, N. (2023). An \u0026lsquo;instinct for learning\u0026rsquo;: the learning flights and walks of bees, wasps and ants from the 1850s to now. Journal of Experimental Biology, 226(6), jeb245278.\u003c/li\u003e\n \u003cli\u003eCollett, M. Chittka, L. and Collett, T. S. (2013). Spatial memory in insect navigation. Curr. Biol. 23, R789-R800. https://doi.org/10.1016/j.cub.2013.07.020\u003c/li\u003e\n \u003cli\u003eCollett, T. S. Robert, T. Frasnelli, E. Philippides, A. \u0026amp; Hempel de Ibarra, N. (2023). How bumblebees coordinate path integration and body orientation at the start of their first learning flight. Journal of Experimental Biology, 226(8), jeb245271.\u003c/li\u003e\n \u003cli\u003eDeeti, S. Islam, M. Freas, C. Murray, T. \u0026amp; Cheng, K. (2023). Intricacies of running a route without success in night-active bull ants (\u003cem\u003eMyrmecia midas\u003c/em\u003e). Journal of Experimental Psychology: Animal Learning and Cognition, 49(2), 111.\u003c/li\u003e\n \u003cli\u003eFleischmann, P. N. Christian, M. M\u0026uuml;ller, V. L. R\u0026ouml;ssler, W. \u0026amp; Wehner, R. (2016). Ontogeny of learning walks and the acquisition of landmark information in desert ants, Cataglyphis fortis. The Journal of Experi- mental Biology, 219, 3137\u0026ndash;3145. http://dx.doi.org/10.1242/jeb.140459\u003c/li\u003e\n \u003cli\u003eFleischmann, P. N. Grob, R. Wehner, R. \u0026amp; R\u0026ouml;ssler, W. (2017). Species- specific differences in the fine structure of learning walk elements in Cataglyphis ants. The Journal of Experimental Biology, 220, 2426\u0026ndash;2435.\u003c/li\u003e\n \u003cli\u003eFleischmann, P. N. R\u0026ouml;ssler, W. \u0026amp; Wehner, R. (2018). Early foraging life: spatial and temporal aspects of landmark learning in the ant Cataglyphis noda. Journal of Comparative Physiology A, 204(6), 579-592.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. \u0026amp; Cheng, K. (2017). Learning and time‐dependent cue choice in the desert ant, melophorus bagoti. Ethology, 123(8), 503-515.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. \u0026amp; Cheng, K. (2018). Landmark learning, cue conflict, and outbound view sequence in navigating desert ants. Journal of Experimental Psychology: Animal Learning and Cognition, 44(4), 409.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. \u0026amp; Cheng, K. (2019). Panorama similarity and navigational knowledge in the nocturnal bull ant Myrmecia midas. Journal of Experimental Biology, 222(11), jeb193201.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. \u0026amp; Cheng, K. (2022). The basis of navigation across species. Annual review of psychology, 73, 217-241.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. \u0026amp; Spetch, M. L. (2019). Terrestrial cue learning and retention during the outbound and inbound foraging trip in the desert ant, Cataglyphis velox. Journal of Comparative Physiology A, 205(2), 177-189.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. \u0026amp; Spetch, M. L. (2023). Varieties of visual navigation in insects. Animal cognition, 26(1), 319-342.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. \u0026amp; Schultheiss, P. (2018). How to navigate in different environments and situations: lessons from ants. Frontiers in psychology, 9, 369115.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. Fleischmann, P. N. \u0026amp; Cheng, K. (2019). Experimental ethology of learning in desert ants: becoming expert navigators. Behavioural Processes, 158, 181-191.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. Plowes, N. J. \u0026amp; Spetch, M. L. (2021). Traveling through light clutter: Path integration and panorama guided navigation in the Sonoran Desert ant, Novomessor cockerelli. Behavioural Processes, 186, 104373.\u003c/li\u003e\n \u003cli\u003eFreas, C. A. Wystrach, A. Schwarz, S. \u0026amp; Spetch, M. L. (2022). Aversive view memories and risk perception in navigating ants. Scientific Reports, 12(1), 2899.\u003c/li\u003e\n \u003cli\u003eGrob, R. Fleischmann, P. N. Gr\u0026uuml;bel, K. Wehner, R. \u0026amp; R\u0026ouml;ssler, W. (2017). The role of celestial compass information in Cataglyphis ants during learning walks and for neuroplasticity in the central complex and mushroom bodies. Frontiers in behavioral neuroscience, 11, 226.\u003c/li\u003e\n \u003cli\u003eHarris, R. A. Graham, P. \u0026amp; Collett, T. S. (2007). Visual cues for the retrieval of landmark memories by navigating wood ants. Current biology, 17(2), 93-102.\u003c/li\u003e\n \u003cli\u003eIslam, M. Freas, C. A. \u0026amp; Cheng, K. (2020). Effect of large visual changes on the navigation of the nocturnal bull ant, Myrmecia midas. Animal Cognition, 23, 1071-1080.\u003c/li\u003e\n \u003cli\u003eJayatilaka P, Murray T, Narendra A, Zeil J (2018) The choreography of learning walks in the Australian jack jumper ant Myrmecia croslandi. J Exp Biol 221:jeb185306. 10.1242/jeb.185306\u003c/li\u003e\n \u003cli\u003eKodzhabashev, A. \u0026amp; Mangan, M. (2015). Route following without scanning. In Biomimetic and Biohybrid Systems: 4th International Conference, Living Machines 2015, Barcelona, Spain, July 28-31, 2015, Proceedings 4 (pp. 199-210). Springer International Publishing.\u003c/li\u003e\n \u003cli\u003eLe M\u0026ouml;el, F. \u0026amp; Wystrach, A. (2020). Opponent processes in visual memories: A model of attraction and repulsion in navigating insects\u0026rsquo; mushroom bodies. PLoS computational biology, 16(2), e1007631.\u003c/li\u003e\n \u003cli\u003eLionetti, V. A. Deeti, S. Murray, T. \u0026amp; Cheng, K. (2023). Resolving conflict between aversive and appetitive learning of views: how ants shift to a new route during navigation. Learning \u0026amp; Behavior, 51(4), 446-457.\u003c/li\u003e\n \u003cli\u003eM\u0026ouml;ller, R. (2012). A model of ant navigation based on visual prediction. Journal of theoretical biology, 305, 118-130.\u003c/li\u003e\n \u003cli\u003eMurray, T. Kocsi, Z. Dahmen, H. Narendra, A. Le M\u0026ouml;el, F. Wystrach, A. \u0026amp; Zeil, J. (2020). The role of attractive and repellent scene memories in ant homing (Myrmecia croslandi). Journal of Experimental Biology, 223(3), jeb210021.\u003c/li\u003e\n \u003cli\u003eNarendra, A. Si, A. Sulikowski, D. \u0026amp; Cheng, K. (2007). Learning, retention and coding of nest-associated visual cues by the Australian desert ant, \u003cem\u003eMelophorus bagoti\u003c/em\u003e. Behavioral Ecology and Sociobiology, 61, 1543-1553.\u003c/li\u003e\n \u003cli\u003ePfeiffer, K. \u0026amp; Homberg, U. (2014). Organization and functional roles of the central complex in the insect brain. Annual review of entomology, 59, 165-184.\u003c/li\u003e\n \u003cli\u003eRobert, T. Frasnelli, E. Hempel de Ibarra, N. \u0026amp; Collett, T. S. (2018). Variations on a theme: bumblebee learning flights from the nest and from flowers. Journal of Experimental Biology, 221(4), jeb172601.\u003c/li\u003e\n \u003cli\u003eSchultheiss, P. Wystrach, A. Schwarz, S. Tack, A. Delor, J. Nooten, S. S. ... \u0026amp; Cheng, K. (2016). Crucial role of ultraviolet light for desert ants in determining direction from the terrestrial panorama. Animal Behaviour, 115, 19-28.\u003c/li\u003e\n \u003cli\u003eSchwarz, S. Mangan, M. Zeil, J. Webb, B. and Wystrach, A. (2017). How Ants Use Vision When Homing Backward. Curr. Biol. 27, 401\u0026ndash;407. https://doi.org/10.1016/j.cub.2016.12.019\u003c/li\u003e\n \u003cli\u003eSeelig, J. D. \u0026amp; Jayaraman, V. (2015). Neural dynamics for landmark orientation and angular path integration. Nature, 521(7551), 186-191.\u003c/li\u003e\n \u003cli\u003eStone, T. Webb, B. Adden, A. Weddig, N. B. Honkanen, A. Templin, R. ... \u0026amp; Heinze, S. (2017). An anatomically constrained model for path integration in the bee brain. Current Biology, 27(20), 3069-3085.\u003c/li\u003e\n \u003cli\u003eWebb, B. \u0026amp; Wystrach, A. (2016). Neural mechanisms of insect navigation. Current Opinion in Insect Science, 15, 27-39.\u003c/li\u003e\n \u003cli\u003eWehner R (1982) Himmelsnavigation bei Insekten: Neurophysiologie und Verhalten. Orell Fussli, Zurich\u003c/li\u003e\n \u003cli\u003eWehner, R. (1997). The ant\u0026apos;s celestial compass system: spectral and polarisation channels. In Orientation and Communication in Arthropods (ed. M. Lehrer), pp.145-185. Basel: Birkhäuser Verlag.\u003c/li\u003e\n \u003cli\u003eWehner, R. (2003). Desert ant navigation: how miniature brains solve complex tasks. Journal of Comparative Physiology A, 189, 579-588.\u003c/li\u003e\n \u003cli\u003eWehner, R. \u0026amp; R\u0026auml;ber, F. (1979). Visual spatial memory in desert ants, Cataglyphis bicolor (Hymenoptera: Formicidae). Experientia, 35(12), 1569-1571.\u003c/li\u003e\n \u003cli\u003eWehner, R. \u0026amp; Srinivasan, M. V. (2003). Path integration in insects. The neurobiology of spatial behaviour, 9-30.\u003c/li\u003e\n \u003cli\u003eWehner, R. Michel, B. and Antonsen, P. (1996). Visual navigation in insects: coupling of egocentric and geocentric information. J. Exp. Biol. 199, 129\u0026ndash;140.\u003c/li\u003e\n \u003cli\u003eWolf, H. Wittlinger, M. \u0026amp; Bolek, S. (2012). Re-visiting of plentiful food sources and food search strategies in desert ants. Frontiers in neuroscience, 6, 102.\u003c/li\u003e\n \u003cli\u003eWystrach, A. Buehlmann, C. Schwarz, S. Cheng, K. \u0026amp; Graham, P. (2020a). Rapid aversive and memory trace learning during route navigation in desert ants. Current biology, 30(10), 1927-1933.\u003c/li\u003e\n \u003cli\u003eWystrach, A. Le Mo\u0026euml;l, F. Clement, L. \u0026amp; Schwarz, S. (2020b). A lateralised design for the interaction of visual memories and heading representations in navigating ants. BioRxiv, 2020-08.\u003c/li\u003e\n \u003cli\u003eWystrach, A. Cheng, K. Sosa, S. \u0026amp; Beugnon, G. (2011). Geometry, features, and panoramic views: ants in rectangular arenas. Journal of Experimental Psychology: Animal Behavior Processes, 37(4), 420.\u003c/li\u003e\n \u003cli\u003eZeil, J. (2012). Visual homing: an insect perspective. Curr. Opin. Neurobiol. 22(2), 285\u0026ndash;293.\u003c/li\u003e\n \u003cli\u003eZeil, J. Hofmann, M. I. \u0026amp; Chahl, J. S. (2003). Catchment areas of panoramic snapshots in outdoor scenes. JOSA A, 20(3), 450-469.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"b647dbf5-87a9-4ef3-b022-63088baa662d","identifier":"10.13039/501100001230","name":"Macquarie University","awardNumber":"MQRF0001094","order_by":0},{"identity":"cc5cfbf1-8e2e-43fe-9328-077d50dcb803","identifier":"10.13039/501100000923","name":"Australian Research Council","awardNumber":"DP200102337","order_by":1}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Macquarie University","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"view comparisons, pirouettes, route following, latent learning, scanning bouts, turnbacks","lastPublishedDoi":"10.21203/rs.3.rs-4670516/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4670516/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eHow ants, wasps and bees rapidly form visual routes represents an enduring mystery as well as a powerful example of the abilites of insect brains. Here, we analyse a previously uncharcterised behaviour, \u0026lsquo;lookbacks\u0026rsquo;, underlies rapid bi-directional route learning in desert ants. During these lookbacks, foragers stop forward movement to their goal location, turn and fixate their gaze to their origin, often for only 150\u0026ndash;200ms. This turn appears to be a critical period for learning the inbound route. Route formation relies on acquiring visual cues and comparing panoramic view memories with the current view. While the nest panorama is learned during pre-foraging learning walks, during which na\u0026iuml;ve ants often fixate their gaze at the nest, route following requires separate behaviours to learn route based views. We untangle how route formation occurs in na\u0026iuml;ve \u003cem\u003eMelophorus bagoti\u003c/em\u003e foragers during the first foraging trips by focusing on the previously uncharacterised lookback behaviours and their function in facilitating visual learning. Lookbacks were highly associated with the first few foraging trips and were concentrated in areas where the visual scene changed rapidly. Analysis of gaze directions during lookbacks show foragers clearly fixate their view to the nest direction during these behaviours (or alternatively to the feeder during inbound homing), learning the nest-aligned inbound route during their first outbound trips. We discuss lookbacks as a \u0026lsquo;when to learn signal\u0026rsquo; combining visual rotation and gaze fixations to produce view-based route following.\u003c/p\u003e","manuscriptTitle":"Route Formation and the Choreography of Looking Back in Desert Ants (Melophorus bagoti)","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-07-03 08:47:03","doi":"10.21203/rs.3.rs-4670516/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"394125d2-a47f-47d2-94ce-e67183fdd8ff","owner":[],"postedDate":"July 3rd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":33985008,"name":"Animal Behavior"},{"id":33985009,"name":"Cognitive Neuroscience"}],"tags":[],"updatedAt":"2024-07-03T08:47:04+00:00","versionOfRecord":[],"versionCreatedAt":"2024-07-03 08:47:03","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4670516","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4670516","identity":"rs-4670516","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}