Corneal sensitivity is required for orientation in free-flying migratory bats

Magnetoreception has been demonstrated behaviorally in several mammals, but the location of the underlying receptors remains unresolved. Evidence across vertebrates indicates that some species can use a magnetic polarity compass, hypothesized to rely on iron oxide (magnetite) particles that could transduce magnetic information via sensory pathways, potentially through the trigeminal system. Prior work in mole rats suggested the cornea may host primary magnetoreceptors, as bilateral corneal anaesthesia disrupted polarity-dependent behaviors. In bats, non-migratory species show polarity-sensitive magnetic orientation, and migratory bats calibrate compass systems to solar azimuth and other cues. However, the role of specific sensory structures in long-distance navigation of free-flying migratory bats is unknown. This study tests whether corneal sensation is necessary for orientation in migrating Nathusius’ bats after displacement by comparing vanishing bearings of bats with unilateral or bilateral topical corneal anaesthesia to sham-treated controls, and validates that corneal anaesthesia does not abolish basic light detection.

The manuscript reviews evidence for magnetoreception across vertebrates, including mammals, fish, and sea turtles, noting polarity-sensitive orientation in darkness. It discusses hypothesized magnetite-based mechanisms that could modulate ion channels depending on field alignment and emphasizes the trigeminal pathway as a conduit for magnetic information processing. A key prior result is that bilateral corneal anaesthesia in mole rats disrupted normal polarity-dependent nesting, implicating the cornea as a potential magnetoreceptive site. In bats, studies have shown compass calibration to solar and olfactory cues and polarity sensitivity in some species, but the mechanisms for migratory navigation remain unclear. Work in birds and homing pigeons is cited, where magnetic anomalies or trigeminal manipulations disrupt orientation or map sense, supporting a role for iron-particle-based receptors and trigeminal transmission in magnetic mapping. Magnetite has been detected in bat heads, but linkage to sensory pathways remains unestablished. The authors also note that prior experiments manipulating visual systems in mammals have not provided evidence for an ocular visual mechanism as a primary orientation cue, underscoring the need to test non-visual corneal sensing hypotheses in free-ranging animals.

Study species and capture: Over three late-summer migration seasons at the Baltic Sea coast near Pāvilosta, Latvia, 100 adult Nathusius’ bats (Pipistrellus nathusii) were captured using a directional funnel trap adjacent to the shoreline. Animals were held temporarily in dark boxes, fed and watered, and managed under permits issued by the Latvian Nature Conservation Agency. Individuals were marked and maintained until experimental procedures and release at night. Corneal anaesthesia treatments: Bats received topical corneal treatments immediately prior to testing. Experimental groups underwent unilateral or bilateral corneal anaesthesia using an ophthalmic topical anaesthetic (oxybuprocaine 0.4% is referenced in the text and supporting citations), applied as eye drops to one or both eyes. Sham control groups received saline drops. Treatments were randomized and administered by personnel blinded to group assignments; the field experimenter was masked to the treatment via coded droppers. Photoreception control (Y‑maze): To verify that corneal anaesthesia did not abolish basic visual light detection, 76 bats were tested in a Y‑maze with one lit exit (~120 lux from visible LEDs) and one dark exit. The bifurcation region was kept dark. Bats were placed at the entrance under a low cap, allowed to move, and their exit choice (lit vs dark) and latency were recorded. Groups included unilateral anaesthesia, bilateral anaesthesia, and corresponding sham controls. Maze arms and lighted sides were alternated between trials, and surfaces were cleaned between tests. Navigation test after translocation: For orientation testing, adult bats were equipped with VHF radio transmitters (<5% body mass; frequencies ~150–152 MHz; models from Holohil, Telemeter-Service Dessau, and Torcklab). Transmitters were glued to dorsal fur with silicone adhesive. Bats were translocated from their migratory corridor and released at unfamiliar coastal sites with clear horizons between 23:00 and 00:00 h, typically in pairs. Release sites were ~1 km apart. Upon release, vanishing bearings and vanishing times were recorded via radio-tracking until signal loss; subsequent checks confirmed no immediate returns. Groups comprised: two sham control cohorts (released in different seasons), a unilateral anaesthesia group, and a bilateral anaesthesia group. Total sample size for orientation analyses was n = 76. Statistical analyses: Circular orientation was assessed using Rayleigh’s test for non-uniformity, with mean vector bearings and lengths (r) computed. Group differences in mean direction and dispersion were evaluated with Mardia–Watson–Wheeler tests and Levene’s tests on absolute angular residuals. Likelihood-based circular models (CircMLE) compared distributions (uniform, unimodal, symmetric/axial bimodal variants) via AICc and model weights. Bootstrapping (up to 100,000 resamples) generated confidence intervals for r to compare directedness between groups. Vanishing times were compared via ANOVA after checking distributional assumptions.

- Photoreception intact under corneal anaesthesia: In the Y‑maze, both anaesthetized and sham-treated bats preferred the lit exit, indicating preserved light detection. Reported proportions choosing lit exits: sham control group(s): 75–77% (n = 16–22); unilateral anaesthesia: 81.3% (n = 16); bilateral anaesthesia: 86% (n = 22), with significant preference over chance in most groups (e.g., χ2 = 11.64, p < 0.001 for bilateral group). Exit latencies did not differ between bilateral anaesthesia and its sham control (Mann–Whitney U = 199.5, n = 44, p = 0.321). - Orientation after release: Both sham control groups oriented seasonally southward with significant mean vectors (Rayleigh’s test: 183° ± 34°, n = 20, r = 0.495, p = 0.006; and 187° ± 34°, n = 19, r = 0.522, p = 0.007). Unilateral corneal anaesthesia did not differ from sham controls in mean direction or dispersion. In contrast, bilateral corneal anaesthesia led to random vanishing directions (Rayleigh’s test: mean 240°, n = 18, r = 0.061, z = 0.206, p = 0.973), best described by a uniform model. Bootstrapped 99.9% confidence intervals for r in oriented groups did not overlap the bilateral anaesthesia group’s r = 0.06 (p < 0.001), confirming significantly reduced directedness. - Variance comparisons: Variances differed significantly between the bilateral anaesthesia group and its sham control (Levene’s F1,35 = 5.824, p = 0.021) and between the two anaesthesia groups (F1,35 = 3.310, p = 0.027). Sham control groups did not differ (Mardia–Watson–Wheeler U = 0.189, p = 0.91). - Vanishing times: No significant differences among groups (means ~16–19 min; ANOVA, p = 0.135).

These field experiments demonstrate that intact corneal sensation is required for accurate orientation in free-flying migratory bats following displacement, without altering environmental cues. Sham-treated bats departed in seasonally appropriate southward directions, as did bats with unilateral corneal anaesthesia, indicating that one functional cornea sufficed. Bilateral corneal anaesthesia abolished directed vanishing bearings while leaving departure speed unchanged. Y‑maze results show that corneal anaesthesia did not eliminate light detection, arguing against a gross loss of retinal function as the cause of disorientation. The findings are consistent with a corneal-based sensory input contributing to navigation, potentially within a magnetic compass or map sense mediated via the ophthalmic branch of the trigeminal nerve, paralleling evidence from mole rats and migratory birds where trigeminal manipulations or magnetic anomalies disrupt orientation. Alternative explanations, such as visual acuity changes due to anaesthetic, cannot be fully excluded but are considered less likely given common ophthalmic use of the anaesthetic and preserved phototaxis. Overall, the data support a bilateral corneal sensory mechanism that can operate unilaterally to sustain orientation, and they highlight the cornea as a promising site for magnetoreception in mammals.

The study provides the first evidence in a free-ranging mammal that corneal sensitivity is necessary for maintaining seasonally appropriate orientation during migration. Bilateral topical corneal anaesthesia led to random vanishing directions, whereas unilateral anaesthesia and sham treatments preserved directed southward departures. Phototaxis remained intact under anaesthesia, suggesting the effect is not due to loss of basic light perception. These results implicate the cornea as a potential site for magnetoreception and suggest a bilateral, trigeminally mediated mechanism that can function with input from a single eye. Future work should directly test magnetic sensing in this species, identify the cellular and anatomical substrates (e.g., iron-particle receptors) within the cornea or associated pathways, and disentangle any residual visual side effects of anaesthetics.

Direct evidence linking the observed disorientation specifically to disruption of magnetic sensing is not yet available. Potential non-magnetic effects of the topical anaesthetic (e.g., on visual acuity or other ocular physiology) cannot be entirely excluded, despite preserved phototaxis. The study focuses on a single migratory bat species and a specific migratory context, which may limit generalizability. Some methodological descriptions indicate variability across field seasons and groups, and certain measurements (e.g., exact visual acuity) were not assessed. Further controlled experiments are needed to localize receptors and confirm the trigeminal magnetic pathway.