Research articlesHumpback whale (Megaptera novaeangliae) visual acuity allows silhouette detection but not fine detail discrimination over ecological distances
Abstract
Few studies have been conducted on the visual capabilities of large cetaceans, such as the humpback whale (Megaptera novaeangliae), and understanding these capabilities provides insights into the natural history and anthropogenic vulnerabilities of these animals, which are otherwise difficult to study in situ. Here, we performed an anatomical and histological study of a subadult humpback whale eye to estimate visual acuity (i.e. spatial resolution of vision) for modelling the perception of relevant visual targets in an open ocean environment. Visual acuity was estimated at 3.95 cycles per degree (CPD)—a value that is an order of magnitude lower than what is predicted by absolute eye size. Perceptual models based on this acuity indicated that low frequency spatial information (e.g. large silhouettes) remains discriminable over ecologically relevant distances, while high frequency spatial information, perhaps critical for target identification, appears lost at 3–4 average humpback whale body lengths away. Models of horizontal sighting distance provided detection-distance thresholds that encompassed the effective range predicted for visual acuity. This study provides new insight into the visual capabilities of humpback whales, suggesting spatial vision that is suited to their open ocean ecology, but challenged by visual targets with more nuanced characteristics unless viewed at close range.
1. Introduction
For many animals, vision is a critical sensory system for finding food, finding mates and navigating the environment [1]. There are many neural and optical strategies for improving vision, but one fundamental strategy is increasing eye size [2]. Larger eyes tend to have larger apertures and longer focal lengths, having the capacity for both a greater sensitivity and resolution of vision [3]. Thus, animals with large absolute eye sizes, such as the large whales, carry the potential for having the greatest visual capability on earth.
For baleen whales (i.e. mysticetes), vision may be particularly important to survival because though they have the capacity for olfaction [4], they lack the capacity for echolocation as found in their sister taxon, the toothed whales (i.e. odontocetes) [5,6]. Despite this, only a handful of studies on baleen whale vision have been completed [7–10], primarily relating to ocular anatomy and their monochromatic spectral sensitivity of the retina [7,8]. Other aspects of vision, such as absolute sensitivity, as well as temporal and spatial resolution of vision (i.e. visual speed and acuity) also contribute to image formation and yet remain nearly unknown across cetaceans and especially among large whales. To better conserve these animals and understand their life history [11], the large whales have been the subject of a growing body of research, with an emerging call for studies of marine mammal sensory capabilities, particularly among large whales [12].
Humpback whales (Megaptera novaeangliae) are the sixth largest extant species on the planet, growing to an average mature length of 14−15 m [13]. Humpbacks reach sexual maturity at 5−10 years of age, live up to 50 years [13], and comprise 15 geographic populations over a global distribution [11,14,15]. For all but one of these populations (i.e. Arabian Sea), humpbacks undergo an annual migration cycle from their feeding grounds towards polar latitudes to calving and breeding grounds towards the equator [11,16]. These animals primarily live and forage in coastal waters, but can venture into oceanic waters, especially during migration [17,18]. In both environments, these whales employ a lunge feeding strategy to engulf large aggregations of prey (e.g. baitballs of krill or fish), a behaviour that occurs at relatively shallow depths between 23 and 160 m [19,20]. Owing to their widespread distribution and activities, humpback whales hold particularly important ecological and economic roles in the open ocean.
Humpback whales are valued for their contributions to environmental regulation [21,22], biodiversity as whale fall ecosystems [21] and the global whale-watching industry [23]. While the International Union for Conservation of Nature (IUCN) Red List status of humpback whales is ‘least concern’ [24], the National Oceanic and Atmospheric Administration and the National Marine Fisheries Service (NOAA/NMFS) [15] list at least two populations as threatened or endangered [25,26]. Today, the leading cause of humpback whale mortality is anthropogenic activities, such as from entanglement in fishing gear, vessel strikes in shipping lanes and ingestion of plastics [27–29]. In the past, sensory biologists have leveraged information on animal sensory capabilities to aid in conservation efforts surrounding anthropogenic threats. While vision has been used to reduce fatal collisions of birds (for example) with anthropogenic obstacles [30], similar strategies for marine mammals have not been tested, possibly owing to the lack of knowledge about marine mammal vision.
Visual acuity refers to the ability of a visual system to resolve static spatial detail [3]. It is often reported in cycles per degree (CPD), which is the number of black and white line pairs that an animal can discriminate within one degree of visual space [3]. Human visual acuity reaches a maximum of 60−100 CPD. Surprisingly, it has been found that most animals have visual acuities below the United States limit of legal human blindness (6 CPD [31,32]), including from one morphological study of a juvenile humpback whale eye predicting a maximum acuity of 3.3 CPD [10]. Morphological estimates of acuity are taken from two features of the eye. The first is the focal length, which is the distance from the point in the lens through which light rays pass without being bent to the image on the retina [1], while the second is the peak areal density of the retinal ganglion cells (RGCs), which are responsible for sending visual information to the brain for processing [32]. As focal length is typically correlated with eye size, animals with larger eyes typically have higher visual acuities [32], as well as those with higher densities of RGCs. In essence, RGCs serve as pixels of the visual image, where a greater number of pixels serves a better resolution of vision [32].
The goal of this study was to investigate humpback whale visual performance in order to model their perception of relevant visual targets in an open ocean environment. To do this, we conducted an anatomical and histological assessment of a humpback whale eye, which allowed us to estimate focal length and visualize RGCs to estimate visual acuity. Counts of RGCs were applied to a retinal topography map, allowing us to identify the corresponding lines-of-sight for the acute zones of vision. The calculated visual acuity and the R-package AcuityView [33] were then used to model the appearance of selected visual targets (i.e. baitball, gillnet) in an open ocean environment. Models of horizontal sighting distance, comprising estimates of visual physiology and environmental lighting conditions, were then used to determine the maximum detection distances of those targets relative to the effective range predicted for visual acuity. The results indicate the functional significance of vision in humpback whales, providing insights into the role of vision in their foraging ecology and the potential capacity of this species to detect and respond to natural and artificial visual targets in their open ocean environment.
2. Material and methods
(a) Tissue collection and morphometrics
A left eye (specimen ID: VGT 259) was collected from a euthanized male subadult humpback whale (8.3 m in total length) after it stranded in Thorofare Bay, North Carolina, USA. The intact eye was then fixed in 10% neutral-buffered formalin and stored for 12 years at room temperature prior to study. Digital calipers (CD-6 ASX; Mitutoyo, Sakado, Kanagawa, Japan) were used to measure (mm) external dimensions of the eye (axial, transverse and longitudinal eye and corneal diameters), as well as transverse and longitudinal diameters of the pupil. After these measurements, the eye was bisected in an oblique (diagonal) plane for measurement (mm) of the axial, transverse and longitudinal diameters of the lens, maximum eye cup depth and maximum scleral (i.e. wall) thickness.
(b) Retinal imaging
The retina was removed and cut into four pieces, representing the dorsal, ventral, temporal and nasal aspects of the eye. These retinal quarters were wholemounted onto 75 × 50 mm gelatin-coated glass microscope slides (Ted Pella INC., Redding, CA, USA), following the methods of Ullmann and colleagues [34]. Any remaining vitreous humour was then removed with the use of a Gibco™ Collagenase Type II paste (Thermo Fisher Scientific, Waltham, MA, USA). After wholemounting, cresyl violet histological stain (0.1% in 20% EtOH-H2O; pH 3.67) was used for RGC visualization. The four slides were placed in one carriage and stained simultaneously until the RGC soma contrasted strongly with the retinal background.
All four quarters of stained retina underwent automated imaging using a Leica THUNDER Imaging System with a DM6 B light microscope (Leica Microsystems, Wetzlar, Germany). The nasal, temporal and dorsal quarters of the retina were imaged using a 5× Pl Fluotar objective, producing approximately 1100 imaging frames per quarter with dimensions of 1200 μm × 1200 μm. Perhaps as an artefact of the specimen’s age, the ventral quarter stained more heavily and thus, required higher magnification images for cell counting. This quarter was imaged using a 10× Pl Fluotar objective, producing approximately 4400 imaging frames, each with dimensions of 600 µm × 600 µm. Finally, the images for each quarter were stitched together in LAS X (Leica Microsystems, Germany) for retinal visualization.
Fiji ImageJ software [35] was used to count RGCs in each image within counting frames scaled to 500 µm × 500 µm. To ensure adequate sampling of RGCs, we calculated the coefficient of error (CE) [36]. A CE value lower than 0.1 is considered a sufficient amount of stereological sampling. We used the Schaeffer estimator [37,38] and found CE values between 0.01 and 0.02 for our wholemounted quarters. After cell counts were completed (see data in electronic supplementary material), we created a topographical map of RGC distribution across this retina. To do this we created a scalable vector graphics (.svg) file in Adobe Illustrator (Adobe Inc. San Jose, California, USA) for our retina and followed an approach of Garza-Gisholt and colleagues [39] to generate retinal cell topography maps in R.
(c) Visual acuity analysis
We calculated visual acuity following Collin & Pettigrew [40] and Murayama and colleagues [41]. We estimated focal length—an approximation of posterior nodal distance—by multiplying the axial length of the eye by 0.57. This ratio of 0.57 refers to the relationship between the posterior nodal distance and axial length of the eye for animals with arrhythmic or crepuscular behaviour [42] and has been used in at least two previous studies of baleen whale vision [10,41]. We then used this value in the following calculations as in other studies of visual acuity [40,41]. First, the angle α, which is the angle that subtends 1 mm of the retina, was calculated from the following equations:
The resulting value was then inputted into the following equation along with the square root of the peak RGC density in cells mm−2 to calculate the number of cells per degree.
Two RGCs are required to differentiate the black and white bars that make up one cycle. Thus, to calculate visual acuity in CPD, the value of cells per degree must be divided by two.
(d) AcuityView modelling
After calculating the visual acuity in CPD, we used this value to investigate humpback whale spatial perception of relevant targets in their environment using the R-package AcuityView [33]. AcuityView uses Fourier transform and a modulation transfer function (MTF) to render visual scenes as they may be perceived by an animal with a lower visual acuity than that of humans [33]. Two visual targets were selected for analysis: (i) a baitball of pelagic schooling fish to address the ecology of foraging and (ii) a commercial drift gillnet to represent a relevant anthropogenic threat. Additionally, two other images were tested: scenes of conspecific and an approaching predator, an orca (Orcinus orca; see electronic supplementary material).
Selected images were licensed through Adobe Stock Images (Adobe Inc. San Jose, California, USA) and edited in Adobe Photoshop (Adobe Inc. San Jose, California, USA) to meet the size and shape requirements of AcuityView. Humpback whales have a single channel spectral sensitivity peaking at 492 nm [7]. Thus, the blue colour channel of each image was isolated and converted to greyscale in Adobe Photoshop. For our images, the ‘real world’ width was used to estimate the width using objects within the image. For the baitball image, an estimated length of one of the fish was used to estimate the width of the image to be approximately 5 m. For the net image, one of the floats on the top of the net was used similarly and the width of the image was estimated to be approximately 3 m. These values were inputted into the AcuityView code along with their respected images after conversion of acuity to CPD to minimum resolvable angle (see electronic supplementary material). To examine perceptual changes over a visual range, we selected five distances scaled to the average humpback whale body length (i.e. 15 m [11]), running the analysis at distances of 15, 30, 45, 60 and 75 m away. Following a precedent in the literature (e.g. [43]), we then visually assessed the changes in spatial resolution of the rendered AcuityView images to identify the distance at which fine-scale spatial information necessary for object identification becomes lost. As the spatial resolution required by a humpback whale to identify an object in a scene is unknown, data were interpreted conservatively as a range of possible distance thresholds.
Perceptual models of spatial vision are robust to variation in the colour and contrast of tested images because spatial information exists independently of these other visual features. Contrast, colour and spatial information however, can all contribute to the likelihood of detecting and identifying a visual target. Thus, AcuityView models must be interpreted in regards to changes in available spatial information within static visual scenes alone. Further, AcuityView produces models of spatial perception as predicted by the capacities of an eye and not an intact animal. Thus, AcuityView cannot account for the brain’s capacity for image processing, such as by edge enhancement [33], which can sharpen images (although, it cannot recover spatial detail) beyond what is detected at the level of the eye. Finally, AcuityView does not account for changes in light attenuation in water owing to environmental conditions, which is a relevant factor in visual performance. Despite these limitations, AcuityView provides a fundamental understanding of how the capacity to resolve spatial detail diminishes with distance and thus, it provides a conservative estimate for the effective visual acuity range for these whales.
(e) Modelling of horizontal sighting distance
To determine the maximum detection distance by a humpback whale for a relevant visual target (i.e. pelagic baitball), we modelled horizontal sighting distances using an approach originally developed by Nilsson et al. [44] and modified by Ruxton & Johnsen [45] (see electronic supplementary material). We further modified that from Ruxton & Johnsen [45] to directly set a diameter of the baitball and not to estimate school size based on the size and number of fish. Incorporating information on pupil diameter, retinal physiology (i.e. integration rate), environmental light and optical properties of the target, this approach is used to estimate the maximum distances over which a target can be detected in a given environment, assuming optical acuity for seeing the target. In each model, we set pupil diameter A = 0.015 m, and then looked to Warrant [2] to set reasonable estimates for vertebrate ocular transmittance τ = 0.8 and quantum efficiency q = 0.34. The rod photoreceptors of humpback whales have a peak absorbance at 492 nm and may have an absorption coefficient of 0.035 μm−1 and an outer segment length of 57 μm [46,47]. The integration time of the eye is unknown for any cetacean, so we used three values over a large range (0.05, 0.20 and 1.00 s), incorporating a value previously estimated for another baleen whale with rod monochromacy (i.e. North Atlantic right whale; Eubalaena glacialis [47]). A range of Weber contrast values relevant to pelagic schooling fish (0.2, 1, 10, 50) was taken from a study by Pertzelan and colleagues [48] and tested. The width of the target (e.g. baitball) was set to 7 m, a previously reported median value for baitball diameters relevant to humpback whale prey [49].
Underwater radiance distributions were calculated at 20 and 50 m depth for a near-zenith sun. See electronic supplementary material for how background radiance spectra Lb(λ) and underwater light field were modelled. The sky and sea were considered to be cloudless and calm, respectively. The sky irradiance was calculated using the Radtran model [50] and the sky radiance distribution was calculated using the model given in Harrison & Coombes [51]. Pure water absorption was taken from Pope & Fry [52]. The scattering phase function was a Fournier–Forand function chosen to fit the measured backscatter ratio [53]. Chlorophyll fluorescence was calculated from chlorophyll absorption estimated according to Prieur & Sathyendranath [54] and a fluorescence efficiency of 0.02. At each depth, radiance was calculated from 400 to 700 nm at 15 nm intervals with an angular resolution of 15° (azimuth) by 10° (elevation).
Depths were selected to represent the shallow end of the range of reported daytime foraging depths for humpback whales [19]. The sky and sea were considered to be cloudless and calm, respectively. The sky irradiance was calculated using the Radtran model [50] and the sky radiance distribution was calculated using the model given in Harrison & Coombes [51]. Pure water absorption was taken from Pope & Fry [52]. The scattering phase function was a Fournier–Forand function chosen to fit the measured backscatter ratio [53]. Chlorophyll fluorescence was calculated from chlorophyll absorption estimated according to Prieur & Sathyendranath [54] and a fluorescence efficiency of 0.02. At both depths, radiance was calculated from 400 to 700 nm at 15 nm intervals with an angular resolution of 15° (azimuth) by 10° (elevation).
3. Results
(a) Morphometrics
Viewed from the front, the humpback whale eye was circular in shape, with the transverse and longitudinal diameters being nearly identical (table 1, figure 1). The eye itself was not spherical, as the axial diameter (or eye depth) was only 70% of the diameter in the transverse plane (table 1, figure 1). Both the cornea and pupil were oval in shape, having a larger diameter in the transverse than longitudinal planes (table 1). Like the eye, the lens was circular when viewed from the front, and not spherical, as the axial diameter (lens depth) was shorter than those in the transverse and longitudinal planes (table 1). The eyecup depth, measured from the opening of the pupil to the back of the eyecup, was found to be much shorter than the axial diameter of the eye. In fact, the eye cup only accounted for 61% of the total axial diameter of the eye (table 1). This was owing to the thickness of the sclera, which at its thickest point made up 39% of the axial length of the eye (table 1, figure 1).
Figure 1. Humpback whale (M. novaeangliae) eye used in this study. Frontal plane (a) and sagittal plane (b) of the eye are shown, as well as the bisected eye (c) following removal of the retina. V = ventral; N = nasal; A = anterior; P = posterior. All scale bars are 10 mm in length. Sample collected by University of North Carolina Wilmington (UNCW) through a Stranding Agreement with NOAA/NMFS and Institutional Animal Care and Use Committee (IACUC) #A0809−019.
eyea | eye cup | cornea | lens | pupil | sclera | |
|---|---|---|---|---|---|---|
transverse diameter | 84.3 | — | 34.0 | 15.2 | 15.5 | — |
longitudinal diameter | 84.7 | — | 26.6 | 15.2 | 14.5 | — |
axial diameter | 59.2 | 36.3 | — | 11.1 | — | — |
maximum thickness | — | — | 2.13 | — | — | 27.6 |
(b) Retinal ganglion cell topography
Cresyl violet staining yielded high-contrast imaging of the retinal cell field, permitting RGC differentiation, identification and counting (figure 2). Cell counts were completed across each retinal quarter, all of which were examined over at least 250 counting sites, with a total of 1385 counting sites across the entire retina (electronic supplementary material, figure S1 and table S1). The range of RGC densities documented across the retina was 16−180 cells mm−2, with areas of high and low RGC densities over this range existing across all four retinal quarters (table 2, figure 2). Low RGC densities were found mostly at the periphery and around the optic nerve. Further, half of the retinal surface, primarily over the ventral and dorsal quarters, had RGC densities of ≤50 cells mm−2. Average cell counts ranged from 26 cells mm−2 in the ventral quarter to 73 cells mm−2 in the temporal quarter (table 2). In many places across the retina, this vasculature would take up to 1/5 of the counting frame, with no RGCs overlaying the vasculature.
Figure 2. Topography map showing retinal ganglion cell (RGC) distribution across the left retina of the humpback whale (M. novaeangliae). (a) Cresyl violet stained retina with insets displaying regions of high and low cell density, with scale bars equal to 100 μm. Sample collected by UNCW through a Stranding Agreement with NOAA/NMFS and IACUC #A0809−019. (b) Two areas of higher cell density can be seen across the naso-temporal axis of the retina and a peak RGC density of 180 cell mm−2 can be seen in the temporal quarter of the retina. N = nasal; V = ventral.
retinal quarter | peak cells mm–2 | average cells mm–2 |
|---|---|---|
nasal quarter | 124 | 65 |
dorsal quarter | 72 | 42 |
temporal quarter | 180 | 73 |
ventral quarter | 72 | 26 |
Across the retinal topography of RGCs, two high-density acute zones were identified (figure 2). The acute zone with the highest RGC density was in the temporal quarter with a maximum cell count of 180 cells mm−2, while that of the nasal quarter had a maximum count of 124 cells mm−2 (table 2, figure 2). Owing to the concavity of the eye, the acute zone in the temporal quarter corresponded to the animal’s forward- or rostral-facing line-of-sight, while that of the nasal quarter corresponded to the rear- or caudal-facing line-of-sight for vision.
(c) Visual acuity and AcuityView modelling
Visual acuity was calculated using the estimated focal length of the eye and peak RGC areal density within the retina. We estimated focal length from the eye’s axial diameter (59 mm), calculating a length of 34 mm. The focal length and the peak density of RGCs (180 cells mm−2) were then used to calculate the humpback whale visual acuity of 3.95 CPD. The calculated visual acuity was converted from CPD to minimum resolvable angle (Δφ), which resulted in a value of approximately 0.25°. This value was then used for the AcuityView [33] modelling.
The AcuityView modelling showed how the capacity of the humpback whale to resolve spatial detail decreased as viewing distance increased (figure 3). The consequences of viewing distance for image resolution are dependent on the spatial scale of information available in the image [1]. Small spatial scale information in the tested images, such as individual bodies of the schooling fish and the mesh architecture of the net, was lost prior to the large spatial scale information of the baitball contour (figure 3) when viewed over increasing distance (figure 3, electronic supplementary material, figure S2). While the fine-scale characteristics, perhaps critical for target identification, appear lost at roughly 45−60 m away (figure 3), the large-scale information permitting target detection appear conserved across all distances tested. Thus, the capacity for target identification may become limited over 3 to 4 average humpback whale body lengths away from a visual target.
Figure 3. Modelling the perception of visual targets in the water column based on the humpback whale (M. novaeangliae) visual acuity estimate (3.95 CPD). The two targets (i.e. bait ball and gillnet) were modelled over distances of 15 m (1 body length) to 75 m (5 body lengths). X represents the visual target, and the humpback whale silhouette is scaled to an average body length of mature humpback whales at 15 m. Note that each baitball has been scaled to the same size so changes in spatial detail could be compared more easily.
(d) Horizontal sighting distance modelling
The daytime horizontal sighting distances of a humpback whale viewing a 7 m diameter school of prey at 20 and 50 m depth are shown in table 3. Depending on the assumed temporal resolution of the eye, the sighting distances ranged from 36 to 41 m and 31 to 37 m at 20 and 50 m depth, respectively, for low contrast targets. For high contrast targets, the distances range from 56 to 61 m and 51 to 56 m at 20 and 50 m depth, respectively. Overall, the sighting ranges vary only about twofold over the large ranges of target contrasts and temporal resolutions tested, as well as over the approximately 20-fold drop in illumination at 50 m depth versus 20 m depth. As expected, longer temporal resolutions and higher target contrasts increased sighting distance, but less than might be expected, because the LambertW function in equation S1 has a very low slope (see electronic supplementary material). Based on the structure of electronic supplementary material, equation S1, the most influential parameter affecting sighting distance is the clarity of the water, with sighting distance being inversely proportional to the beam attenuation coefficient c, which is in turn a measure of the murkiness of the water. Thus, individuals at higher latitudes and closer to the coast, where c increases substantially, will have considerably shorter sighting distances.
20 metres depth | CFF (Hz) | ||
|---|---|---|---|
contrast | 20 | 5 | 1 |
0.2 | 36 | 39 | 41 |
1 | 42 | 44 | 47 |
10 | 50 | 52 | 55 |
50 | 56 | 58 | 61 |
50 metres depth | CFF (Hz) | ||
contrast | 20 | 5 | 1 |
0.2 | 31 | 34 | 37 |
1 | 37 | 39 | 42 |
10 | 45 | 47 | 50 |
50 | 51 | 53 | 56 |
4. Discussion
We performed an anatomical and histological assessment of a humpback whale eye to calculate peak visual acuity and model the perception of relevant visual targets in an open ocean environment. Peak densities of RGCs were identified in the temporal and nasal aspects of the retina, conferring the greatest acuity of vision to the rostral-facing and caudal-facing lines-of-sight, respectively. The maximum visual acuity (estimated from the temporal aspect of the retina) was roughly an order of magnitude lower than what might be predicted by the animal’s absolute eye size at 3.95 CPD [32]. Perceptual models based on this acuity indicated that low spatial frequency information appears discriminable at any distance within the maximum target detection ranges we predicted for humpback whales, up to five average humpback whale body lengths away. However, high spatial frequency information, perhaps critical for target identification, appears limited to roughly two to four body lengths away. These findings provide new information about humpback whale vision, indicating their acuity may be suited to detecting large silhouettes (e.g. school of prey), but not to visualizing fine detail in the environment, except when viewed at close range.
(a) Morphometrics, topography and acuity
The relative morphometrics of the eye examined here were consistent with those observed in other cetaceans [10,55,56]. Like in other mysticetes and odontocetes, the cornea and pupil of the eye specimen were longer in the transverse than longitudinal axes. Transverse elongation of the cornea and pupil could relate to having a greater field-of-view in the horizontal plane, as seen commonly in other animals with laterally placed eyes [57]. Cetaceans also have a shorter axial length of the eye (i.e. eye depth) than that of the transverse and longitudinal diameters [10,56]. Reducing eye depth decreases focal length and thus, may decrease spatial resolution [1,32]. This is further exacerbated in humpback whales and other cetaceans where over one-third of the eye’s axial depth is occupied by an unusually thickened sclera [10,55,56]. It remains unknown why cetaceans have such thickened sclera relative to other mammals [55]. Here, the maximum scleral thickness recorded for our specimen was 27.6 mm, amounting to 39% of the axial depth of the eye. In turn, the sclera reduces the depth of the eye cup, diminishing the focal length (and thus, visual acuity) from what otherwise could be supported by the axial length of the eye.
The RGC areal densities observed over the retina were similar to those found in the previous study of a juvenile humpback whale by Lisney & Collin [10]. While we found a peak RGC density of 180 cells/mm−2, they found a slightly lower peak value of 160 cells/mm−2 in the juvenile. Alignment of our findings to those of Lisney & Collin [10] was encouraging as any assessment of a single specimen, especially that preserved for an extended period, are problematic for histology. Though perhaps minimal for ocular tissues [58], formalin fixation is known to shrink soft tissue volume [59], and while this should not change the number of RGCs present in the retina, it could increase RGC density due to retinal shrinkage [60]. Any artifact of retinal shrinkage would only increase visual acuity, which was already documented here (and by [10]) as exceptionally low relative to that predicted by humpback whale eye size [32].
Further like Lisney & Collin [10], we found two acute zones of RGC density, with that in the temporal aspect having a higher density of RGCs relative to the acute zone in the nasal aspect. These findings are in line with acute zone positions known among other cetaceans [55], which due to the concavity of the retina, can serve two directions of vision. The acute zone in the temporal aspect of cetacean eyes serves rostral-facing vision, predicted to aid in ecological tasks such as navigation and foraging [55]. Whereas that of the nasal aspect serves caudal-facing vision, possibly aiding in conspecific and calf visualization and predator avoidance [55]. Like the acute zones of other animals, these localized areas of high RGC density may optimize acuity across the visual field for visualizing relevant targets in the environment [55]. This cellular topography of the retina, along with binocular field-of-view, potential for stereoscopic vision, and capacity for lens accommodation (especially as it relates to aerial versus underwater vision [55]), for example, all contribute to how humpback whales may experience their visual world. The latter three however, have yet to be determined for this species.
Calculation of visual acuity yielded a value of 3.95 CPD for our humpback whale specimen. This fell in line with the findings of Lisney & Collin [10] for their juvenile specimen at 3.3 CPD. This difference was driven by both a shorter focal length and a lower peak RGC density in the juvenile relative to our subadult specimen, suggesting that visual acuity can change ontogenetically as maturing animals experience increases in eye size [61]. Ontogenetic change in visual acuity has not been thoroughly investigated in marine mammals, however it has been shown in other species, primarily fish, that it can increase during development [61,62].
Here, the acuity found for the humpback whale (3.95 CPD) was surprising because it was roughly an order of magnitude lower than what might be predicted by the animal’s absolute eye size (i.e. axial length of the eye [32,43]). Because focal length affects visual acuity, this aspect of visual performance typically scales with changes in eye size [32,43]. In fact, at least for some taxonomic groups, eye size has been shown to be a greater predictor of visual acuity than any other ecological variable (e.g. [43]). Based on the scaling of acuity across vertebrate taxa, the visual acuity of a humpback whale would be estimated to fall between 30 and 60 CPD, leading to a question as to why it was only 3.95 CPD.
Mysticetes, who forage on large schools or blooms of prey, may not need to identify single targets but rather, silhouettes of larger swaths of prey from a distance. This is supported here by the relatively low acuity value calculated and the retained salience of large silhouettes over the AcuityView models of humpback whale visual perception. Further support for this stems from the relatively low acuity values of other mysticetes, ranging among grey whales (Eschrichtius robustus) [63], minke whales (Balaenoptera acutorostrata) [41] and Bryde’s whales (Balaenoptera brydei) [10] between 2.7 and 4.8 CPD. Slightly higher acuity among odontocetes (such as in the orca, Orcinus orca, 5.45 CPD [64]), may relate to how they forage on individual fish and squid and their interactions with relatively smaller conspecifics [6,65]. However, overall lower acuity among cetaceans compared with terrestrial mammals suggest a possible reliance on other sensory capabilities, such as echolocation in odontocetes [66] and olfaction in mysticetes [67]. It is plausible that humpback whales orient to blooms of prey at the surface by olfaction [68] and then localize the blooms using sight to optimize the efficiency of lunge feeding, a foraging strategy with a high metabolic cost [18].
(b) Modelling visual perception
To help assess the capacities of humpback whale spatial vision, we used AcuityView to evaluate the appearance and discriminability of relevant targets based on visual acuity (3.95 CPD). Modelling object appearance over distance, we found that the capacity of humpback whales to resolve the fine-scale spatial information of a visual target was lost over ecologically relevant distances. Specifically, while visual information at the largest spatial scale, such as large object silhouettes, was retained at all distances, the details of the targets, such as the mesh of the net or outlines of individual fish, were lost over sighting distances of roughly 45 to 60 m away. Taken in an ecological context, humpback whales may only need to resolve large objects in their environment to survive, such as a baitball, conspecific, calf or predator (such as an orca, Orcinus orca [69], electronic supplementary material, figure S2). Thus, high-acuity vision may not be required for these animals to survive under natural conditions. Artificial objects in the water column, however, such as nets, may provide a significant challenge to humpback whales, since their more nuanced spatial characteristics may not be resolved.
Further, we aimed to test whether the sighting distances estimated for effective visual acuity fell within the maximum detection ranges of humpback whale vision in their open ocean environment. Between the most conservative and least conservative scenario tested, the extreme bounds of detection distances spanned 31 to 61 m, with this variation primarily driven by tested contrast differences of pelagic targets. High contrast of a target increases detectability against background, as would be provided in this context by frequent specular (mirror-like) reflections from a school of fish [48]. That same school of fish with a lower frequency of specular reflections or alternative prey types such a bloom of krill [48] provides a lower-contrast target that would be more difficult to detect at a distance. The magnitude of variation over this range (twofold difference) and the value of the upper-bound itself represent relatively short distances, however, owing to how these values scale with humpback whale body length (15 m on average). Nearly identical to the effective range of fine-scale visual acuity estimated above (≤60 m), the maximum range over which a humpback whale may even detect a target (61 m) is up to four average humpback body lengths in distance.
(c) Implications
A leading cause of mortality for the humpback whale is entanglements with fishing gear, with up to >50% humpback whales in some populations having entanglement scarring [70]. As mentioned, sensory biologists have leveraged information on animal sensory capabilities to aid in conservation efforts surrounding anthropogenic threats (e.g. [30,71]). The present findings may aid in such an effort by providing a basis for further studies of these occurrences using fieldwork. To that end, the effective range of humpback whale visual performance predicted here centres over 45−60 m and thus, is a distance of 3−4 average humpback whale body lengths away. Thus, with median speeds of 2 ms–1 while swimming [72] and 4 ms–1 during lunges [73], the amount of time before reaching an obstacle could be 30 s at most. Thus, the capacity for the whales to avoid a detected target depends, in part, on their manoeuvrability [72,73], which has yet to be explored within this context and requires further study.
5. Conclusions
Visual physiology is often just one component of a larger sensory repertoire that animals use to interface with their environment. Although the current study is informative about an animal that is otherwise challenging to study in situ, these data must be interpreted conservatively and with the goal of obtaining knowledge to design further experiments both in the lab and the field. The design of further experiments requires careful consideration, but might best be approached by correlating physiological assessments of vision (such as the one presented here) with the visual characteristics of fishing gear reported from entanglement events.
Overall, this study provides new insight into the functional significance of visual performance in humpback whales, suggesting visual capabilities that are suited to their open ocean ecology, but challenged by targets with more nuanced characteristics unless viewed at close range. Our finding of a visual acuity that is much lower than what would be predicted by eye size raises new questions about possible selective forces and constraints on the evolution of spatial vision, particularly among cetaceans. Studies such as these help to close the knowledge gap on large whale sensory ecology, contributing to our fundamental understanding of their natural history and physiological resilience in our changing world.
Ethics
Euthanasia was conducted under NOAA consultation. UNCW stranding response was carried out under a Stranding Agreement from NOAA Southeast, NOAA Marine Mammal Health and Stranding Response Permit 932-1905 and UNCW IACUC protocol A 0809-019.
Data accessibility
All relevant data can be found within the article and the online supporting information.
Supplementary material is available online [74].
Declaration of AI use
We have not used AI-assisted technologies in creating this article.
Authors’ contributions
J.R.B.: conceptualization, data curation, formal analysis, methodology, visualization, writing—original draft, writing— review and editing; V.M.M.: conceptualization, data curation, methodology, supervision, writing—review and editing; S.J.: conceptualization, data curation, formal analysis, methodology, visualization, writing—original draft, writing—review and editing; L.E.S.: conceptualization, methodology, project administration, supervision, writing—original draft, writing—review and editing.
All authors gave final approval for publication and agreed to be held accountable for the work performed therein.
Conflict of interest declaration
We declare we have no competing interests.
Funding
Tissue collection was supported in part by NOAA Prescott Stranding Grant NA10NMF4390250. This work was supported by UNCW’s Center for the Support of Undergraduate Research (CSURF) Paul E. Hosier Research Fellowship awarded to J.R.B.
Acknowledgements
We wish to thank the stranding response team, led by Dr. Vicky Thayer (North Carolina Division of Marine Fisheries), Dr. Craig Harms (North Carolina State University) and Mr. William McLellan (University of North Carolina Wilmington), and including responders and assistance from the NCSU Center for Marine Sciences and Technology, North Carolina Maritime Museum, Carteret Community College, Duke University, US Coast Guard Ft. Macon, Cape Lookout National Seashore, USMAS Cherry Point, and Tow Boat U.S. We also thank Dr. Tiffany Keenan, Ms. Alison Loftis, Dr. Michael Tift, Mr. William McLellan and Dr. Ann Pabst of the UNCW Marine Mammal Stranding Program for their support and guidance. We wish to also thank Dr. Alison Taylor and Ms. Elizabeth Elliott for their support in the Richard M. Diliman Bioimaging Facility at UNCW.
