Magnetic compass reception in birds appears to be a light dependent process based on antagonistically interacting spectral mechanisms with at least one short-wavelength and one long-wavelength mechanism (Deutschlander et al., 1999b) or multiple states of a short-wavelength sensitive mechanism (Ritz et al., 2000, 2004). The avian magnetic compass differs from the technical compass used by humans in that the avian compass is an inclination compass (Wiltschko and Wiltschko, 1972). Rather than distinguishing North and South, the compass distinguishes the direction of the pole from the direction of the equator; the North Pole and the South Pole are not distinguishable (Fig. 1). Thus, for a migratory songbird, fall migration is typically equatorward and spring migration is typically poleward. The system works equally well for the northern and southern hemispheres. Birds that transit the equator must somehow switch from an equatorward strategy to a poleward strategy. Experience with the horizontal magnetic field appears to trigger this response (Beason, 1992; Wiltschko and Wiltschko, 1992).
Evidence for the use of a compass that responds to changes in the ambient magnetic field comes from experiments with homing pigeons, migratory songbirds, and from radar studies of migrating birds (reviewed by Wiltschko and Wiltschko, 1995
). Although the biophysical mechanism of the receptor for the magnetic compass has not been identified conclusively in any species, the most convincing evidence indicates that the magnetic compass receptor involves photopigments.
Electrophysiological recordings from the optic tectum of the pigeon (Semm and Demaine, 1986) and the Bobolink (Dolichonyx oryzivorus; Beason and Semm, 1987) revealed the presence of neurons that respond to changes in the direction of the ambient magnetic field, but only in the presence of light. These data could be interpreted to mean either that there is a light-dependent magnetic receptor associated with the visual system or that there is magnetic input to the visual system that is mediated by stimulation of the retinal photoreceptors. Because most of the input to the optic tectum is from the retinal ganglion cells, it is unlikely that the magnetic responses are only gated by input from the visual system; a more parsimonious explanation is that the responses are from magnetic receptors. The responses of the visual system to magnetic stimulation resemble its responses to light. Some units responded only to specific orientations of the magnetic field. Unfortunately, there was no analysis of wavelength sensitivity in these experiments, only the presence or absence of light. The responses were extinguished in total darkness.
The avian magnetic compass must integrate information from the magnetic receptors and the vestibular system in order to determine the direction in which the magnetic field dips below the horizontal. The Nucleus of the Basal Optic Root (nBOR) of the pigeon appears to be involved in this process (Semm et al., 1984; Semm and Demaine, 1986); it receives input from the vestibular system and the visual system. Its responses to changes in the magnetic field depend on both the orientation of the magnetic field and the orientation of the bird's head to the horizon. Single neurons responded to changes in orientation of the magnetic field but did not respond to changes in the intensity of the magnetic field. The neurons that responded to the magnetic field also responded to directional movements of light, especially those cells that responded to axial movements of light. Cells that did not respond to the movement of light did not respond to changes in the magnetic field. The coupling of these specific responses might be an indication of how the magnetic sensitivity of the avian visual system is mediated.
Leask (1977) proposed an optical pumping resonance model to account for light sensitivity of the avian magnetic receptor system, based on the triplet state of a visual pigment such as rhodopsin. Schulten (1982), Phillips (Phillips and Borland, 1994), and Ritz (Ritz et al., 2000, 2002) subsequently modified Leask's (1977) original model so that it involves two pigments or one pigment in two states. These models postulate a wavelength (color) sensitivity in which orientation would be accurate when the animal is illuminated by one category of light and disoriented or reoriented with other categories.
Behavioral experiments also support the concept of a light-dependent, wavelength-sensitive magnetic compass system in birds. Homing pigeons transported in total darkness were more poorly oriented at the release site than pigeons transported with illumination inside their boxes (Wiltschko and Wiltschko, 1981). Neither group of pigeons could not see outside the boxes and, thus, did not have access to visual cues that might be provided by the landscape or sky. Bobolinks tested in orientation funnels in a dark planetarium were inactive. Playback of nocturnal call notes did not stimulate them to show migratory hopping (R.C.B., unpublished data). Likewise, pigeons released with their eyes covered by opaque lens flapped to the ground (C. Walcott, personal communication). One explanation is that the birds lacked a functional frame of reference and chose not to attempt to migrate or fly home.
There appears to be some species-specific responses of migratory orientation by songbirds under narrow-band illumination. All species tested under short wavelengths of light (blue to humans: 425–450 nm) at low-intensity showed normal orientation. Species differences appeared under intermediate wavelengths (green-yellow: 500–575 nm). European Robins (Erithacus rubecula; Wiltschko and Wiltschko, 1999, 2001, 2002; Wiltschko et al., 2001) and Australian Silvereyes (Zosterops lateralis; Wiltschko et al., 1993; Munro et al., 1997) showed normal orientation but Bobolinks were disoriented (Beason and Swali, 2001; Beason, unpublished data). At longer wavelengths there were also species-specific responses. Under yellow-orange light (550–585 nm) Bobolinks showed significant rotation in their headings and became axially bimodal in their responses. European Robins were disoriented under orange (590 nm) and red (635 nm) light. Bobolinks were also disoriented under red (600 nm) illumination. Homing pigeons illuminated with colored lights showed responses similar to the European Robin. They were normally oriented at the release site when transported with white or green light and disoriented when transported under red illumination (Wiltschko and Wiltschko, 1998). If robins were pre-exposed to low intensity red light, they were not disoriented but were able to orient normally under red illumination (Wiltschko et al., 2004a). These results indicate that some type of adaptation is taking place within the magnetic receptors. These observations do not fit any of the current models, which all predict disorientation under long wavelength illumination, and merit further research. Juvenile robins showed appropriate orientation under 560.6 nm illumination but not with 567.5 nm; they were significantly oriented under 617 nm light but not in the migratory direction (Muheim et al., 2002). Under brighter light intensities robins sometimes exhibited bimodal fixed orientation unrelated to the migratory direction (Wiltschko and Wiltschko, 2001, 2002).
When the yellow (590 nm) was simultaneously presented with blue (424 nm) or green (565 nm), robins were oriented but not necessarily in a migratory direction. Regardless of season, the combination of yellow and blue resulted in birds orienting in a southerly direction while the combination of yellow and green produced northerly orientation (Wiltachko et al., 2004b). In all cases adding yellow light resulted in a fixed direction response, as did brighter light intensities. Although illuminating a bird with monochromatic or dichromatic light is artificial, behavioral responses to these situations provide some insight into how the receptor system operates. What is unknown is whether there is a single wavelength sensitive receptor or an interaction of multiple receptors.
In the European Robin (Erithacus rubecula), the Silvereye (Zosterops lateralis), and, perhaps the pigeon, the ability to utilize the Geomagnetic field for migratory orientation is strongly lateralized, with a marked dominance of the right eye for magnetoreception (Wiltschko et al., 2002, 2004a).
From these physiological and behavioral responses of birds, we can conclude that photopigments are involved with the avian magnetic compass system. The wavelength specific effects indicate the effect is on the receptor rather than a generalized motivational response. The differences among species to similar wavelengths of light might be due to differences in the visual pigments used to transduce the magnetic field. The location and structure of the wavelength-sensitive magnetoreceptor remains unresolved. It could involve retinal photoreceptors, extraretinal photoreceptors in the pineal or elsewhere (Semm et al., 1980), photopigments such as cryptochromes (Cashmore et al., 1999) within the retina (Möller et al., 2004; Mouritsen et al., 2004) or brain (Wilson, 1991), or it might be mediated through a non-visual pathway. Mechanisms have been put forth for each of these ideas. Phillips and coworkers (see Deutschlander et al., 1999b for review) proposed that detection of the magnetic field is through energy transfer between two dissimilar visual pigments. For this system to work most effectively the pigments must be contained within the same cell. The wavelength-sensitive magnetic receptor in the eastern red-spotted newt (Notophthalmus viridescens) is located in the pineal organ, which contains functioning photoreceptors (Deutschlander et al., 1999a). Pineal photoreceptors of the newt contain two visual pigments or one pigment in two states, each sensitive to different wavelengths of light (Fig. 2). The avian pineal does not appear to be the site of the avian magnetic compass because pinealectomized pigeons orient as well as unmanipulated birds (Maffei et al., 1983) and pinealectomized Pied Flycatchers (Ficedula hypoleuca) orient when supplied with melatonin (Schneider et al., 1994). In the newt the orientation of the external magnetic field is hypothesized to affect the efficiency of energy transfer and this efficiency is used by the receptor to determine the direction of the magnetic field. The means by which the receptor determines the efficiency of the energy transfer is unclear. Ritz (Ritz et al., 2000, 2004) and Schulten (1982) have developed theoretical radical-pair models in which the balance between two states of a photopigment, such as a crypotochrome, is influenced by the external magnetic field (Fig. 3). Because of the sensitivity of these pigments to short wavelengths, sensitivity to magnetic fields is predicted to disappear at long wavelengths. The responses of Garden Warblers to RF treatment is consistent with the use of a radical pair reaction as the source of magnetic compass information in this species (Ritz et al., 2004).
The avian double cone has also been proposed as the light-dependent, wavelength-sensitive receptor (Beason and Swali, 2001
). It is composed of two photoreceptors, each with a different pigment and oil droplet or filter (Fig. 4
). Consequently, each receptor has a different spectral sensitivity, similar to the model proposed by Phillips. However, because each pigment is in a different cell, it would be difficult to transfer energy between pigments. Exactly how the double cone could function as a magnetic receptor is unclear but the wavelength sensitivities of the Bobolink correspond to the sensitivities of the two receptors.