The Eye

Avian eyes are big both in relative and absolute terms (Walls, 1942). Diurnal raptors, as well as owls, have relatively larger eyes than other flying birds (Brooke, Hanley, & Laughlin, 1999), and predatory raptors have relatively larger eyes than carrion eating species (Potier et al., 2017a). Some of the largest raptor eyes are more than 30 mm in axial length (e.g., Philippine eagles Pithecophaga jefferyi, Golden eagles Aquila chrysaetos, Secretary birds Sagittarius serpentarius), while the smallest are less than 10 mm (e.g., black-thighed falconets [Microhierax fringillarius] or collared falconets [Microhierax caerulescens]; Ritland, 1982).

Based on the concavity of the sclero-corneal junction, and the ratio between axial length and transverse eye diameter, Walls (1942) has grouped avian eye shapes into three rough categories: flat, globose, and tubular. He described raptor eyes as globose (Figure 1a) or tubular (in some eagle species). The concavity of the sclero-corneal junction is supported by a ring of bones, the scleral ossicles (Figure 1a and Figure 2), which differ in number between groups (13–15 in Catarthidae, 14–17 in Accipitridae, and 14–16 in Falconidae; Curtis & Miller, 1938).

An intraocular structure found in the eyes of all birds, the pecten oculi, projects from the optic disc into the vitreous (Figure 1a). Its size and shape varies between species. In raptors, the pecten is of the pleated type and contains from 10 (turkey vultures [Cathartes aura]; Lord, 1956; golden eagles; Murphy & Dubielzig, 1993) to 18 pleats (red-tailed hawks [Buteo jamaicensis]; Braekevelt, 1991; common buzzards [Buteo buteo]; Gültiken, Yildiz, Onuk, & Karayigit, 2012). It is highly vascularized with large capillaries and melanosome-dense stromal cells (Lord, 1956, Kiama, Bhattacharjee, Maina, & Weyrauch, 1994; Gültiken et al., 2012). As the retina of birds is avascular, the most substantiated function of the pecten is oxygen and nutrient supply to the inner layers of the retina (Meyer, 1977; Pettigrew, Wallman, & Wildsoet, 1990; Blank et al., 2011). Many other tasks, including optical functions such as improvement of visual acuity, motion detection and distance perception wait for experimental validation (reviewed in Meyer, 1977). Some evidence suggests that pecten might act as intraocular shade (Barlow & Ostwald, 1972), thus reducing negative effects of glare on visual acuity that result from the sun’s image falling upon the retina. Finally, as a large retinal region under the base of pecten oculi is devoid of photoreceptors, there should be an elongated blind sector projecting into the dorso-nasal quadrant of the monocular visual field, thus reducing the total visual coverage (Martin, 2009).

Optical System of the Eye

Birds, as all vertebrates, have camera-type eyes. Before reaching the retina light first passes through the ocular media: the cornea, aqueous humor, lens, and vitreous humor (Figure 1a; Land & Nilsson, 2012). Optically, they can be described, in a simplified manner, by the anterior focal length (f) and pupil aperture (A) (Martin, 1983). The anterior focal length is tightly correlated to the axial length of the eye. Thus large raptor eyes have long focal lengths and subsequently large retinal images, potentially leading to more acute vision. The pupil size sets the optical cut-off frequency—the ultimate diffraction limited threshold—for resolving power. The larger the pupil, the lower the diffraction limit and thus the higher the potential spatial resolution of the eye. Even in bright light, many raptors have their pupils widely opened (Miller, 1979), again indicating potentially high visual acuity. However, wide pupils contribute to a low F-number (f/A), and thus a short depth of focus, making the eye more susceptible to optical aberrations (Martin, 1983).

In raptors, the quality of the optical system has only been measured in Congo serpent-eagles (Dryotriorchis spectabilis), in which only a small amount of spherical aberrations was found (Shlaer, 1972). Chromatic aberrations have not been studied in any raptor species. However, Lind, Kelber, and Kröger (2008) found that red-shouldered hawks (Buteo lineatus) and bald eagles (Haliaeetus leucocephalus) have bifocal lenses (with two zones of different refractive powers) potentially helping to reduce longitudinal chromatic aberrations—but only when the pupil is fully open and light strikes a large area of the lens.

Open in new tab

Figure 2. Avian accommodation. Semi-diagrammatic accommodation apparatus of a hawk. Modified from Walls (1942).

Open in new tab

Figure 3. (a) Ocular media transmittance of four raptor species. λ‎_t50 is the wavelength at which 50% of the incoming light is transmitted. (b) Cone spectral sensitivities of common buzzards Buteo buteo. Data from Lind et al. (2013). VS—violet-sensitive cone; SWS—short wavelength-sensitive cone; MWS—medium wavelength-sensitive cone; LWS—long wavelength-sensitive cone; DC—double cone.

The main function of the lens and cornea is to focus the image on the retina. Many birds have two accommodation mechanisms: lenticular and corneal accommodation (Glasser & Howland, 1996; Land & Nilsson, 2012). Brücke’s muscles squeeze the soft and flexible lens through the pupil, supported by a muscular iris. Crampton’s muscles pull the cornea and reduce its radius of curvature to increase refractive power (Figure 2). Glasser, Pardue, Andison, and Sivak (1997) studied accommodation in several species of raptors and found that red-tailed hawks can accommodate up to 28 Dioptres (D) while other studied species accommodated between 4 D (sharp-shinned hawk [Accipiter striatus]) and 16 D (American kestrel [Falco sparverius]). Moreover, while smaller species with high corneal curvatures did not show corneal accommodation, turkey vultures and three species of eagles (African fish-eagles [Haliaeetus vocifer], bald and golden eagles) showed between 3 and 6 D of corneal accommodation (Glasser et al., 1997). These raptors were able to accommodate on objects viewed monocularly while the focus of the other eye remained unchanged. However, when focusing objects presented frontally, in the binocular visual field, both eyes accommodated symmetrically. In addition, Murphy, Howland, and Howland (1995) have shown that Swainson’s hawks (Buteo swainsoni), Cooper’s hawks (Accipiter cooperii), and American kestrels do not have lower-field myopia, which is common in ground-feeding birds.

Apart from their refractive function, the ocular media also alter the spectrum of light reaching the retina, mainly by absorbing light of shorter wavelengths. The ocular media transmittance is similar in common buzzards, European sparrowhawks (Accipiter nisus) and common kestrels (Falco tinnunculus), in which the spectral position of 50% transmittance (λ‎_t50) is found at 375±5 nm (λ‎_t50 is a commonly used indicator of UV transparency; Figure 3a; Lind, Mitkus, Olsson, & Kelber, 2013). The ocular media in red kites (Milvus milvus) transmit even less UV light, with the λ‎_t50 at 394 nm (Lind et al., 2013). Thus, compared to the majority of birds, which have strongly UV-transmissive ocular media, much less ultraviolet light reaches the retina of a raptor (Lind, Mitkus, Olsson, & Kelber, 2014). This obviously influences the spectral sensitivity of the photoreceptors in the retina (Figure 3b).

The Retina

Open in new tab

Figure 4. Histological (a, b) and Optical Coherence Tomography (OCT) cross-sections (c, d) of the peregrine falcon (Falco peregrinus) (a, b) and Harris’s hawk (Parabuteo unicinctus) (c, d) retina. (a, c)—central fovea, (b, d)—temporal fovea. While resolution in histological cross-sections is much higher, OCT imaging is a non-invasive technique, thus devoid of possible histological processing artifacts. However, without knowing the focal length of the eye, the lateral scale in the image cannot be obtained (note only vertical scale for c and d), therefore foveal shape in such images should not be misinterpreted. Note that oil droplets have lost their pigmentation during histological processing in (a) and (b) and that pigment epithelium layer is lost in (b). Sections in (a) and (b) are 2 μ‎m thick; stained with Azur II—Methylene Blue. RNFL—retinal nerve fibre layer; GCL—ganglion cell layer; IPL—inner plexiform layer; INL—inner nuclear layer; OPL—outer plexiform layer; ONL—outer nuclear layer; ELM—external limiting membrane; PIS—photoreceptor inner segments; POS—photoreceptor outer segments; RPE—retinal pigment epithelium.

The raptor retina has the inverted design typical for all vertebrates (Land & Nilsson, 2012). Before reaching the photoreceptors, which are located outermost in the retina, light passes several distinct neuronal layers (naming and abbreviations in Figure 4). The general arrangement of layers is constant over the retina (except in the fovea; see “Retinal Specializations”), but the retina tends to be thinner in the periphery. In raptors, the central retina is 400–500 μ‎m thick (Potier et al., 2017a) and thus thicker than in other birds (200–350 μ‎m; e.g., Rochon-Duvigneaud, 1943), in humans, and in other mammals (230 and 160–250 μ‎m respectively; Buttery, Hinrichsen, Weller, & Haight, 1991).

Photoreceptors

Open in new tab

Figure 5. Schematic representation of photoreceptor complement in the retina of a diurnal raptor. Because no microspectrophotometric data exist for receptors of any diurnal raptor species, the spectral properties of cones, cone pigments and oil droplets were predicted according to generalisations in other birds (Hart & Vorobyev, 2005) based on the peak wavelength (λ‎ max = 405 nm) of the sws1 pigment (Ödeen & Håstad, 2003) and ocular media transmittance (Lind et al., 2013) of common buzzards Buteo buteo, The λ‎ cut values of T-type (VS cone) and P-type (double cone) oil droplets are taken from Hunt et al. (2009) as generalized values of a passerine bird. The VS cone sensitivity is shifted by the ocular media transmittance, not the absorbance of the T-type oil droplet. The spectral properties of the rod photoreceptor are very conservative among more than 20 bird species from different orders, for which measurements exist; therefore it is safe to assume that diurnal raptor rods are similar. The color of the oil droplets is indicative of its type (T, C, Y, R, P, and A) and is associated with a specific type of photoreceptor. There is no data on the spectral parameters of the A-type oil droplet in the accessory member of the double cone. VS—violet-sensitive cone; SWS—short wavelength-sensitive cone; MWS—medium wavelength-sensitive cone; LWS—long wavelength-sensitive cone; DC—double cone; Rh1—rhodopsin type 1; sws1—short wavelength-sensitive pigment type 1; sws2—short wavelength-sensitive pigment type 2; Rh2—rhodopsin type 2; m/lws—medium/long wavelength-sensitive pigment; λ‎ max—wavelength of the peak sensitivity; λ‎ mid—wavelength of half-maximum absorbance; λ‎ cut—wavelength below which no significant amount of light is transmitted by the oil droplet (diagrams of photoreceptors are modified from illustrations provided by Dr. Peter Olsson).

Open in new tab

Figure 6. Photoreceptor layer in the temporal retina of the red kite (Milvus milvus). In the tangential section (a) single cones, double cones, and rods can be easily identified based on specific morphology. In the cross-section through the photoreceptor inner segments (b) only double cones can be distinguished from other photoreceptors. Note that oil droplets have lost their pigmentation during histological processing. Sections are 2 μ‎m thick; stained with Azur II—Methylene Blue. PC—principle member of the double cone; AC—accessory member of the double cone; R—rod; OD—oil droplet; OSC—outer segment of the single cone; ISC—inner segment of the single cone; DC—double cone.

Most diurnal raptors have the full complement of photoreceptors found in birds: one type of rod mediating vision in dim light, one type of double cone, and four types of single cone (Figure 5). All cones are equipped with oil droplets, small spherical organelles situated between the inner and outer segments filled with lipids and carotenoids (Toomey et al., 2015). Each single cone type combines one opsin-based visual pigment and one type of oil droplet: a transparent oil droplet is found in the violet-sensitive cone expressing an SWS1 opsin, which in raptors is probably maximally sensitive in the violet (405–420 nm), not the UV region of the spectrum (Ödeen & Håstad, 2003; Ödeen & Håstad, 2013). In two species of Accipitriformes, cinereous vultures (Aegypius monachus) and black-winged kites (Elanus caeruleus), Wu et al. (2016) did not detect mRNA coding for an SWS1 opsin. While the loss of SWS1 opsin was also observed in owls (Wu et al., 2016) and mammals living in dim light environments (Jacobs, 2013), the reasons for this characteristic are not fully clear yet. The sensitivity of the violet-sensitive cone is strongly influenced by the UV-absorbing ocular media (Figure 3). The other three single cone types have colored oil droplets that narrow the spectral sensitivity of the cones and shift them toward longer wavelengths (Figure 5; Toomey et al., 2015), but detailed studies of raptor oil droplets have not yet been performed.

By citing older studies, Walls and Judd (1933) mention that Eurasian hobbies (Falco subbuteo) and American kestrels, as well as common buzzards, have four types of oil droplets throughout the retina but lack the red type in the fovea. Gondo and Ando (1995) report four types of oil droplets in the retina of Japanese lesser sparrowhawks (Accipiter gularis), without specific information on their distribution.

In addition to single cones, raptors, like other birds, have double cones (Figures 5 and 6), which have broader spectral sensitivity (Figure 3). In Bonelli’s eagles (Aquila fasciata) and red kites the principal member of double cones has a yellow oil droplet, while the accessory member has a small colorless droplet (Gallego, Baron, & Gayoso, 1975). Double cones are thought to mediate achromatic vision, for instance for motion detection, but are absent from the foveae of at least some raptors (Mitkus, Olsson, Toomey, Corbo, & Kelber, 2017), probably to allow for high cone density and thus high resolution (Figure 6). The single cone packing density in the fovea of some Accipitriform raptors is close to the maximum predicted by the limitations of waveguide optics (Reymond, 1985, 1987; Land & Nilsson, 2012). This allows for high resolution (see “Anatomical Spatial Resolution”) but limits sensitivity and may be a main reason why most Accipitriforms show no crepuscular or nocturnal activity.

Anatomical Spatial Resolution

Spatial resolution describes the ability of an eye to resolve spatial detail in the visual scene. When the image is formed on the retina by the optical system of the eye, it is sampled by the photoreceptor array. The finer the photoreceptor mosaic, the more detail that can be captured from the retinal image. The narrow photoreceptors in some raptor species attain densities of more than 450,000 cells/mm² (wedge-tailed eagles [Aquila audax]; Reymond, 1985). However, due to the wave nature of light, photoreceptors cannot function optimally when their diameters approach the wavelength of light. Indeed, the narrowest photoreceptors in the animal kingdom are about one micrometre wide (Land & Nilsson, 2012) and cones in wedge-tailed eagles (1.6 μ‎m; Reymond, 1985), common buzzards, common kestrels (1.7 μ‎m; Oehme, 1964), and brown falcons (Falco berigora)(1.8 μ‎m; Reymond, 1987) are among the thinnest reported in birds to date. To reduce the deleterious effects of optical “cross-talk” between the narrow neighboring photoreceptors the pigment of retinal pigment epithelium cells surrounds and separates photoreceptor outer segments (Cronin, Johnsen, Marshall, & Warrant, 2014; Figure 6a). Narrow photoreceptors unavoidably collect less light, thus the long cone outer segments in the high-acuity regions of raptors may be a way to overcome this problem (Walls, 1942; Oehme, 1964).

For maximal spatial resolution, the high acuity region of the retina should contain only those photoreceptors, which contribute to high-resolution vision. As rods are not active in bright light, which is required to resolve fine spatial detail, they are absent from the high acuity regions of some birds (Bruhn & Cepko, 1996; Querubin, Lee, Provis, & O’Brien, 2009; Coimbra, Collin, & Hart, 2015), including some diurnal raptors (Oehme, 1964; Mitkus et al., 2017). Even double cones, which have larger diameters than single cones, are absent from the foveae of some raptor species (Mitkus et al., 2017).

Furthermore, to maximise information transfer from the photoreceptors to the brain, photoreceptor signals should not be pooled in the RGCs. Low convergence ratios have been shown in high-acuity zones of several birds (Oehme, 1964; Fite & Rosenfield-Wessels, 1975; Querubin et al., 2009; Coimbra et al., 2015). A low convergence ratio unavoidably increases retinal neuron density, retinal thickness, and the number of ganglion cell axons, which leave the eye. Thus, with a low convergence ratio in the entire retina, RGC axons would cover the inner retina in a thick layer, and the optic nerve would need to be as wide as the whole eye (Land & Nilsson, 2012), rendering such eye highly improbable. Furthermore, because the retina is an energetically costly tissue (Laughlin, de Ruyter van Steveninck, & Anderson, 1998), high visual acuity in the entire retina would make the eye a very expensive organ. Therefore, the regions of highest retinal neuron density, and thus sharpest vision are small even in raptor eyes (e.g., 1.5° in wedge-tailed eagles; Reymond, 1985).

Retinal Specializations

The distribution of photoreceptors and other retinal interneurons varies across the retina. A concentric region of increased retinal neuron density is called an area, whereas an elongated, bandlike region is referred to as a visual streak. Both area and visual streak may contain a fovea—a region where inner retinal layers are fully or partially displaced, resulting in an invagination on the inner retinal surface. In the deepest foveae even photoreceptor nuclei are displaced, whereas in the shallowest foveae no or little reduction in RGC number occurs and only a small depression is present (Figure 4).

Open in new tab

Figure 7. Retinal ganglion cell density maps of the black-chested buzzard-eagle (Geranoaetus melanoleucus) (a) and Andean condor (Vultur gryphus) (b). Numbers represent cell density boundaries in ×1000 cells/mm². Black dots in the highest cell density regions indicate the position of the foveae. Note both central and temporal foveae in the black-chested buzzard-eagle, but only central fovea in the Andean condor. N—nasal; V—ventral. Modified from Inzunza et al. (1991).

Most raptor species studied to date have a central and a temporal area. RGC topography maps exist only for six diurnal raptor species: American kestrel, black-chested buzzard-eagle (Geranoaetus melanoleucus), chimango caracara (Phalcoboenus chimango), Andean condor (Vultur gryphus), black vulture (Coragyps atratus), and turkey vulture (Figure 7; Inzunza, Bravo, Smith, & Angel, 1991; Lisney et al., 2013). All these species have a central and a temporal area within a weakly pronounced visual streak. Most predatory raptors have a deep central fovea and a shallow temporal fovea, while the carrion-eating raptors have only the deep central fovea (Inzunza et al., 1991; Ruggeri et al., 2010; Potier et al., 2017a). Many other bird species also possess a fovea, and two foveae have been reported in other pursuit-hunting species such as swallows, martins, kingfishers, and terns (Wood, 1917; Walls, 1942; Rochon-Duvigneaud, 1943; Meyer, 1977).

The peculiar anatomy of the deep raptor fovea has inspired scientists to search for a functional explanation of this structure. Walls (1937, 1940) was the first to suggest that the fovea might have an optical magnifying effect and thus improve spatial resolution beyond the limits set by photoreceptor and RGC density and the focal length of the eye. Based on refractive index measurements of the retina and vitreous humour (Valentin, 1879), Walls reasoned that the curved slopes of the fovea should refract rays of light outward, thus magnifying the retinal image locally. Inspired by findings of Fox, Lehmkuhle, and Westendorf (1976) that American kestrels have a 2.6 times higher visual acuity than humans (which in other studies has been shown to be similar to humans; see Hirsch, 1982; Gaffney & Hodos, 2003), Snyder and Miller (1978) revitalized this idea and proposed that the bottom-most part of the deep fovea and the ocular lens could create a telephoto system, thus explaining the unusually high visual acuity of this small raptor. However, Reymond, (1985, 1987) showed that anatomical and behavioral limits of spatial resolution almost perfectly match in brown falcons and wedge-tailed eagles, without any need for magnifying effect of the fovea.

Pumphrey (1948) proposed an alternative hypothesis. He calculated that foveal magnification should be uneven and thus should distort the image, making the magnifying effects negligible for improvement of spatial resolution. Instead, he argued that the magnification and distortion on the foveal slopes could improve detection and fixation on small moving objects. More than half a century later, Pumphrey’s calculation was backed up by advanced optical modeling. Even though Frey et al. (2017) modeled light paths in a shallow human fovea, they concluded that optical effects of the fovea are in general small and that substantial foveal magnification, as proposed by Snyder and Miller (1978), could be excluded. However, experimental evidence to support Pumphrey’s hypothesis has yet to be provided.

Visual Pathways

The axons of RGCs reach the brain via the optic nerve, after a complete nerve crossing in birds, and deliver information to three parallel visual systems: tectofugal pathway, thalamofugal pathway, and accessory optic system (reviewed in Güntürkün, 2000 or Wylie, Gutiérrez-Ibáñez, & Iwaniuk, 2015). The majority of RGCs in birds contribute to the tectofugal pathway, which is involved in pattern, brightness, color, and motion analysis, among other visual tasks. In American kestrels the entire retina is represented on the contralateral optic tectum, and the retinal mapping is similar to that found in chickens and pigeons (Frost, Wise, Morgan, & Bird, 1990). However, despite occupying a small part of the total retinal area, the central and temporal foveal regions of American kestrels occupy a disproportionally large sector of the optic tectum (Frost et al., 1990; Inzunza & Bravo, 1993). Other retinal regions are represented in smaller tectal parts.

In contrast to the complete retinal representation in the tectofugal pathway (Frost et al., 1990), only the temporal, but not the central, retinal regions of American kestrels and black vultures have an extended representation in the visual Wulst (Pettigrew, 1978), which is the major termination of the thalamofugal pathway. The visual Wulst, which is considered analogous to the primary visual cortex of mammals, is involved in motion analysis, spatial orientation, and binocular vision (Wylie et al., 2015). In American kestrels the visual Wulst is well developed, contains many binocular cells and has the same retinal mapping as in Barn owls Tyto alba (Pettigrew, 1978). The binocular cells are sensitive to retinal disparity and are involved in stereopsis, which has been behaviorally confirmed in American kestrels (Fox, Lehmkuhle, & Bush, 1977) and barn owls (van der Willigen, 2011) but also in pigeons (McFadden, 1987). Differently from pigeons (Remy & Güntürkün, 1991), the frontal specialization of the thalamofugal pathway in kestrels may be related to the need of precise object analysis in the binocular visual field at the last moments of prey capture (Güntürkün, 2000), even though information on local motion (produced by an object moving within the visual field) is also analyzed in the tectofugal pathway (Frost, 2010).

When a bird moves through the environment, patterns of self-induced motion are generated across all visual field (global motion). The optic flow provides information about the 3D structure of the surroundings as well as the bird’s movement through space, and is processed in a third visual pathway, the accessory optic system (AOS; Frost, 2010). Davies and Green (1990) have shown, that Harris’s hawks (Parabuteo unicinctus), unlike pigeons, use self-induced rates of motion expansion to time feet extension for landing. However, not much more is known about the AOS in raptors.

The optic nerve of birds not only contains axons leaving the retina but also axons reaching the retina from the brain. The isthmo-optic nucleus (ION) receives input from the optic tectum and projects to the retina. Many functions have been proposed for this centrifugal pathway, including shift of visual attention, feeding/pecking, and detection of aerial predators (reviewed in Wylie et al., 2015). The neuronal complexity and relative size of the ION differs largely between species, but the ION of diurnal raptors, as well as owls, is smallest among all bird groups investigated (Shortess & Klose, 1975; Weidner, Repérant, Desroches, Miceli, & Vesselkin, 1987; Gutiérrez-Ibáñez et al., 2012).

Visual Fields

Visual fields describe the space around an animal, from which visual information can be extracted. Comparative studies across a wide range of bird species have been reviewed (Martin, 2017). Visual fields vary markedly between species, and also subtle, but functionally significant differences occur between closely related species, indicating that the configuration of visual fields is as subject to natural selection as are optics and retinal topography.

In birds, contrary to humans, each eye looks outward at a different scene and the overlap of both eyes’ visual fields (the binocular region) is relatively small (Figure 1b). The binocular overlap in birds is smaller than 60° (it is 120° in humans), often between 20–30°, and it is as narrow as 5–10° in some species (Martin, 2017). Because of the relatively small region of binocular overlap in front of the head the total visual field is extensive. Many birds have just a narrow blind region behind the head, and some species have comprehensive visual coverage both around and above the head (Martin, 2007). Birds, and especially raptors, have smaller eye movements than humans (O’Rourke, Hall, Pitlik, & Fernandez-Juricic, 2010), however, differences do exist between species. It appears that predators that search for fast-moving prey have larger eye movements (up to 12° on the horizontal axis in Cooper’s hawks and 10° in red-tailed hawks) that may reduce retinal blur by tracking the prey without a need to move the head, which is costly during flight (Tucker, 2000). On the contrary, species that do not engage in prey pursuit have smaller eye movements (less than 2° in American kestrels; see O’Rourke et al., 2010). These differences in eye movements may also be associated with different visual fields configurations.

Variation in visual fields of birds results from two antagonistic key drivers (Martin, 2017). The primary driver is the control of bill (or feet) position and the timing of bill (or feet) arrival at a target, the secondary driver is the task of detecting predators. The control of bill or feet position is based upon information derived from the optic flow-field in the binocular region that approximately encompasses the direction in which the bill projects or into which talons are positioned just prior to prey capture (Martin, 2009). This may explain the relatively large binocular field found in some pursuit-hunting raptors compared to those that take more sedentary prey. For example, a 47° binocular field is present in Harris’s hawks (Potier et al., 2016a), which take fast moving prey on the ground. However, only a 20° binocular field is found in short-toed snake eagles (Circaetus gallicus), which mainly hunt reptiles on the ground (Martin & Katzir, 1999) and similar field widths are found in vultures, which eat carrion (Martin, Portugal, & Murn, 2012; also see Table 1).

Open in new tab

Figure 8. Visual fields in Harris’s hawks (Parabuteo unicinctus) (a, c, e) and black kites (Milvus migrans) (b, d, f). The blind sector above the head is significantly larger in the predatory Harris’s hawk as compared with the carrion-eating black kite. (c, d) Orthographic projection of the boundaries of the retinal fields. The latitude and longitude coordinate system has the equator aligned vertically in the median sagittal plane. The bird’s head is imagined to be at the centre of the globe (grid is at 20° intervals in latitude and 10° in longitude). (e, f) Horizontal sections through the horizontal plane (90–270°). Modified from Potier et al. (2016a).

In the binocular field, flow-field information from the environment relatively close to the bird is extracted, which depends upon relatively low spatial resolution. Frontal vision is the area into which the temporal fovea projects. This fovea has lower spatial resolution than the central fovea, which is directed toward the lateral field (Wood, 1917; Inzunza et al., 1991; Kane & Zamani, 2014). The lateral regions of the visual field, besides having a function for flight control, primarily serve the detection of predators, prey or conspecifics, and it is advantageous for these behavioral tasks if the blind region is small.

Even closely related (congeneric) species of raptors may differ in their visual field configuration, which can be explained functionally by differences in foraging ecology. For example, in Cooper’s hawks, the maximum binocular overlap is placed significantly higher above the plane of the bill compared with red-tailed hawks (O’Rourke et al., 2010), possibly due to the type of prey they capture. Cooper’s hawks capture mainly flying birds by attacking them from below, often needing precise talon positioning above the plane of the bill, while red-tailed hawks mostly capture mammals on the ground after an approach from above.

Other differences in visual field configuration can be found between predators and carrion eaters (Figure 8). It has been shown that carrion-eating black kites (Milvus migrans) have a narrower blind area above the head compared to predatory Harris’s hawks (Potier et al., 2016a). The larger visual field of black kites may improve carrion detection using social information from conspecifics.

Visual fields differ in many parameters; the one listed here, maximum binocular overlap, is just one parameter, although it is an important one for understanding the form and functions of visual fields. Each of the papers cited should be consulted for diagrams of the visual fields and the values of other parameters, such as the vertical height and position of the binocular region, the locations and width of blind areas above and behind the head, the widths of the field in a single eye. It is clear that there is variation in maximum binocular field widths between the eagles, vultures, and hawks, but also among the hawks. It is worth noting that the broadest binocular fields do not occur in raptors as is commonly supposed. Statistical analysis of all available visual field data (see Martin, 2017, Appendix 2) has shown that the binocular fields of passerine birds are significantly broader than those of diurnal raptors and are indeed broader than those of all non-passerine species (Troscianko, von Bayern, Chappell, Rutz, & Martin, 2012).

Open in new tab

Figure 9. Sun shades in raptors. A characteristic feature of the visual fields of diurnal raptors is the presence of broad blind areas above the head. These prevent the sun from being imaged upon the retina. (a) The shading effect of these ridges is show in this photograph of a short-toed snake eagle (Circaetus gallicus). The bird has been photographed from beneath when it was sitting under a bright overhead sun. While the feathers above the bill are brightly illuminated by the sun, the eyes are well shaded by the prominent ridges above the eyes. (b) The photograph of the ridge above the left eye of a Griffon vulture Gyps fulvus shows a substantial structure that protrudes prominently from the skin of the skull; the ridges are not a feature of the skull itself.

On the other hand, the wider blind area above the head may improve prey detection in Harris’s hawks by reducing the dazzling effect of sunlight. Generally, visual field characteristics above the head are driven by the need to avoid imaging the sun on the retina, specifically in species with large eyes adapted for high resolution. Indeed, imaging the sun results in a general decrease of resolution because of glare. This likely is the main reason why some birds, especially hunting raptor species (e.g., Harris’s hawks, short-toed snake eagles; Potier et al., 2016a; Martin & Katzir, 2000), have prominent optic adnexa (eyelashes and ridges above the eyes; see Figure 9; Martin, 2007). These adnexa lead to large blind areas above the head but serve as sunshades for the eyes (Martin & Katzir, 2000). In smaller-eyed species, which do not have that high resolution, light from the sun may not degrade the image as much. These species may thus benefit from a more complete visual coverage of the world above them, for example, to see larger aerial predators.

Finally, visual field configurations may have a direct impact on collision susceptibility (Martin, 2017). This is particularly the case in the open airspace, in which many large raptors forage by flying high above the ground and looking downward. Because natural large obstacles are usually absent from open airspace, raptors may not have been selected to detect objects in front of them when flying in the open and searching below. Raptors, especially large vultures (e.g., Griffon vultures [Gyps fulvus]) suffer strongly from man-made devices such as power lines or wind turbines (De Lucas, Janss, Whitfield, & Ferrer, 2008). The structure of their visual fields may provide one explanation for the high collision rate, as Griffon vultures have a blind sector in the direction of their travel when they forage (Martin et al., 2012).

Visual Acuity

Visual acuity can be estimated both from anatomy (see “Anatomical Spatial Resolution”) and determined in behavioral experiments. While estimation from optics and cell densities provides the upper limit of spatial resolving power, behavioral analyses measure the functional visual acuity (i.e., visual acuity used by the bird in a behavioral context).

Controlled behavioral studies allow for a comparison of visual acuity between species. The most commonly used method is a two-alternative forced choice discrimination task. An animal is trained to make choices between two different simultaneously presented stimuli, usually a grating of black-and-white stripes and a uniform gray field of the same mean luminance. The spatial frequency of the grating (in cycles, or pairs of black-and-white stripes, per degree of visual angle, cyc/deg), which cannot be reliably resolved and discriminated from the uniform gray gives the visual acuity threshold.

Spatial resolution has been studied in eleven diurnal raptor species—nine of which with behavioral experiments (Table 2). While the behaviorally determined visual acuities of some raptors, such as Gyps vultures or wedge-tailed eagles, are the highest reported for any animal to date (104–142 cyc/deg; Fischer, 1969; Reymond, 1985), in others, such as Chimango caracaras (Potier, Bonadonna, Kelber, & Duriez, 2016b), it is more than four times lower. Such variation reflects the correlation of acuity with eye size (Brooke et al., 1999; Kiltie, 2000), and different needs for sharp vision in these species. Among raptors, the highest acuity was found in large species that search for large food items (relatively large mammals for the eagle and large carrion for the Gyps vultures). On the contrary, smaller species that search for small insects from a shorter distance, such as Chimango caracaras, have lower visual acuity (Potier et al., 2016b). However, many behavioral experiments were done on only one individual per species, and inter-individual differences may be high, as found in some studies (Table 2).

Some raptors have high spatial resolution, but this comes with important limitations. First, raptors have a rather limited sensitivity for ultraviolet light. Increasing UV sensitivity causes image blur, resulting from longitudinal chromatic aberration. Second, there is a fundamental trade-off between spatial resolution and sensitivity. Therefore, in raptors, as light levels fall, resolution drops rapidly (Figure 10), as found in brown falcons (Reymond, 1987), wedge-tailed eagles (Reymond, 1985), and some vultures (Fischer, 1969). Finally, the maximal visual acuity is always determined and only valid for high-contrast stimuli.

Contrast Sensitivity

Open in new tab

Figure 10. (a) Spatial resolution as a function of stimulus luminance. Data from: Griffon vulture Gyps fulvus and Egyptian vulture (Neophron percnopterus) (Fischer, 1969), wedge-tailed eagle Aquila audax (Reymond, 1985), Brown falcon Falco berigora (Reymond, 1987). (b) Contrast sensitivity function for stationary stimuli of the two diurnal raptor species tested to date: American kestrel (Falco sparverius) (Hirsch, 1982) and wedge-tailed eagle (Reymond & Wolfe, 1981). Contrast sensitivity is the inverse of contrast threshold, and contrast is given as the difference between the luminances of two stimuli divided by their sum (see, e.g., Lind et al., 2012). Barn owl (Tyto alba) (Harmening, Nikolay, Orlowski, & Wagner, 2009) and human (Bisti & Maffei, 1974) data are presented for comparison.

In many natural situations, the object of interest, whether it is a prey, a predator or an ornament in the plumage of a potential mate, does not have 100% contrast to its background. Therefore, the maximum visual acuity described in the previous section can rarely be utilized. Due to the fundamental limits of the wave nature of light, contrast of high spatial frequencies is reduced even in an ideal optical system (Land & Nilsson, 2012). Thus, in order to understand visual abilities of an animal better, visual acuity for different contrasts is of paramount interest. The relationship between resolvable spatial detail and its contrast is described by the contrast sensitivity function (CSF; Ghim & Hodos, 2006).

To date, the CSF (Figure 10b) has been determined only for two raptor species, American kestrels (Hirsch, 1982; Ghim & Hodos, 2006) and wedge-tailed eagles (Reymond & Wolfe, 1981). Surprisingly (but similar to other birds) these raptors have lower contrast sensitivities than many mammals studied to date (see summary in Lind, Sunesson, Mitkus, & Kelber, 2012). For example, humans, who have the best-known contrast sensitivity, can differentiate two stimuli with luminance contrast as low as 0.4% (contrast sensitivity of 230; Bisti & Maffei, 1974), while the limit for American kestrels and wedge-tailed eagles is at 3.2% and 7.1% (contrast sensitivity of 31 and 14, respectively). It is important to note that the best contrast sensitivity is never found for the highest spatial frequency that an animal can resolve but for a much coarser pattern. Objects or ornaments of low contrast can only be resolved if they are of particular angular size, as indicated in Figure 10b. Finally, much like maximum visual acuity, contrast sensitivity deteriorates as light levels decrease.

Absolute Sensitivity and Adaptation to Diurnal and Nocturnal Habits

Most diurnal raptors are typically active in daylight and probably for good reason: the best visual acuity of their eyes can only be achieved in high light levels. The absolute sensitivity has never been determined in diurnal raptors; however, the visual acuity in different light levels is known for several species (Figure 10a). In wedge-tailed eagles, the acuity decreases four-fold as light level drops four orders of magnitude (138 cyc/deg at 2000 cd/m² and 34 cyc/deg at 0.2 cd/m²; Reymond, 1985), whereas in brown falcons the change is only 1.5-fold (73 cyc/deg at 2000 cd/m² and 47 cyc/deg at 0.2 cd/m²; Reymond, 1987). Brown falcons not only have a lower F-number, but they also have a shorter focal length and larger receptors in the fovea (Reymond, 1985, 1987) that enable them to capture more photons due to the larger receptor acceptance angle. Indeed, brown falcons are sometimes foraging at dusk, much like many other falconiforms (Reymond, 1987).

A few species of falcons (pygmy-falcons [Polihierax spp.] and forest-falcons [Micrastur spp.]), which have relatively large eyes, are adapted to hunt under the canopy of tropical forests where light levels are low (White, Olsen, & Kiff, 1994). Red-throated caracaras Daptrius americanus are active before sunrise and after sunset, hunting prey by walking on the ground (White et al., 1994). Several falcons use open landscape to regularly hunt on the wing migrant passerines or nocturnal seabirds, insects, or bats after sunset: Eleonora falcons (Falco eleonorae), sooty falcons (Falco concolor), four species of hobby (Falco subbuteo, Falco cuvierii, Falco severus, Falco longipennis), Dickinson kestrels (Falco dickinsoni), bat falcons (Falco rufigularis), merlins (Falco columbarius) peregrine falcons (Falco peregrinus), and Lanner falcons (Falco biarmicus) (Fenton et al., 1994; Hanmer & Blackwood, 1982; Lee & Kuo, 2001; Stanton, 2016; White et al., 1994; Yosef, 1991). Falcons living in artificially illuminated cities increase their foraging time by being active at night. Common kestrels hunt bats in twilight by a sit-and-wait ambush strategy at the exit of bat roosts (Mikula, Hromada, & Tryjanowski, 2013) and lesser kestrels (Falco naumanni) hunt insects in flight next to illuminated cathedrals (Negro, Bustamante, Melguizo, Ruiz, & Grande, 2000). Peregrine falcons have also been observed hunting at night in several cities (DeCandido & Allen, 2006; Rejt, 2004).

Far fewer species of accipitriforms have been reported to be active at night or dusk. Red-tailed hawks, African goshawks (Accipiter tachiro), and Wahlberg’s eagles (Aquila wahlbergi), can hunt bats just before sunset (Fenton et al., 1994; Lee & Kuo, 2001). Some vultures have been observed feeding on carrion at night but most likely arrived at the carcass before dusk (Mundy, Butchard, Ledger, & Piper, 1992). The two most nocturnal accipitriform species are bat hawks (Macheiramphus alcinus), which mainly hunt bats at dusk but also during the night under moonlit conditions (Black, Howard, & Stjernstedt, 1979), and letter-winged kites (Elanus scriptus), which hunt nocturnal desert rodents at night (Pavey, Gorman, & Heywood, 2008; Stanton, 2016). The letter-winged kites have lower F-number than the congeneric black-shouldered kites (Elanus notatus) (0.98 and 1.11, respectively; Pettigrew, 1982). Some authors have argued a convergent evolution between Elanus kites and owls, both on behavioral and ecological terms (Negro et al., 2006).

Besides feeding behavior, some raptors using powered flapping flight can migrate at night, especially when crossing the sea: ospreys (Pandion haliaetus), three species of falcons, seven species of Accipitridae including honey buzzards (Pernis spp.) and northern harriers (Circus cyaenus) (DeCandido, Bierregaard, Martell, & Bildstein, 2006; Russell, 1991). Turkey vultures have occasionally been reported flying at night but apparently not for migratory reasons (Tabor & McAllister, 1988).

While general mechanisms enabling an animal to have more sensitive vision are well understood (higher photoreceptor optical sensitivity, higher degree of spatial and temporal summation; Land & Nilsson, 2012), they have not been investigated in diurnal raptors in detail. The evidence of nocturnal and crepuscular behaviors in some raptor species indicates that their visual sensitivity is sufficient to perform certain tasks. However, not all dim light habits necessarily indicate a high sensitivity of the visual system. The activity pattern can be shaped by other factors, such as type of prey, hunting technique, habitat complexity, and use of other sensory systems in foraging.

Color Vision

With four single cone types, raptors should have similar color vision to other birds. However, there have been no behavioral studies of color discrimination in raptors. Although it is likely that color contrast helps raptors to detect and discriminate conspecifics, predators, and prey, color vision systems tend to have lower resolution in general (Lind & Kelber, 2011). The low transmittance of the ocular media for ultraviolet light and thus low sensitivity for light of short wavelengths restricts the visual abilities of raptors. Color signals in the plumage of small birds are often invisible for raptors and so are the urine traces left by small rodents, which have earlier been suggested to be useful cues for some raptor species (Lind et al., 2013).

Key Points

Key Future Questions