Module 5: Vision & Navigation

Birds possess sensory capabilities far exceeding our own: tetrachromatic color vision extending into the ultraviolet, and an ability to sense Earth's magnetic field via quantum-mechanical radical-pair chemistry. This module derives the biophysics of these remarkable systems from first principles.

1. Tetrachromatic Avian Vision

The vertebrate retina contains two photoreceptor types: rods (scotopic, achromatic) and cones (photopic, chromatic). Humans possess three cone types (trichromacy) with peak sensitivities at approximately 430 nm (S-cone), 530 nm (M-cone), and 560 nm (L-cone). Most birds possess four cone types (tetrachromacy): an additional ultraviolet or violet-sensitive cone (UVS/VS) peaking between 355–425 nm, plus short (SWS, ~440 nm), medium (MWS, ~508 nm), and long (LWS, ~565 nm) wavelength cones.

1.1 Visual Pigments: Opsin + Retinal

Each cone type expresses a specific opsin protein covalently bonded to 11-cis-retinal via a Schiff base to a lysine residue (K296 in bovine rhodopsin numbering). Photon absorption isomerizes 11-cis-retinal to all-trans-retinal, triggering a conformational cascade:

\[ \text{Opsin} \cdot \text{11-}cis\text{-retinal} \xrightarrow{h\nu} \text{Batho} \to \text{Lumi} \to \text{Meta I} \to \text{Meta II} \to \text{Opsin} + \text{all-}trans\text{-retinal} \]

Meta II (R*) activates transducin (G-protein), which activates phosphodiesterase (PDE), hydrolyzing cGMP. The drop in [cGMP] closes cyclic-nucleotide gated (CNG) channels, hyperpolarizing the cell and reducing glutamate release onto bipolar cells. The quantum efficiency of photoisomerization is remarkably high: $\phi \approx 0.67$.

The peak absorption wavelength $\lambda_{\max}$ is tuned by amino-acid residues in the retinal binding pocket (spectral tuning). A single amino acid substitution at position 86 (Phe→Tyr) in passerine UVS opsins shifts sensitivity from UV (∼365 nm) to violet (∼410 nm) — this is the VS/UVS dichotomy among birds.

1.2 Oil Droplet Filters

Uniquely avian (and shared with reptiles and some fish), each cone contains a spherical oil droplet (∼1–3 μm diameter) positioned between the inner and outer segment. These droplets are loaded with carotenoid pigments that act as long-pass cutoff filters, absorbing short wavelengths and transmitting only wavelengths above a cutoff $\lambda_c$.

The effect on effective spectral sensitivity is a convolution of the opsin absorption spectrum$A(\lambda)$ with the droplet transmittance $T(\lambda)$:

\[ S_{\text{eff}}(\lambda) = A(\lambda) \cdot T(\lambda), \quad T(\lambda) = \frac{1}{1 + e^{-k(\lambda - \lambda_c)}} \]

This sigmoidal transmittance (with steepness parameter $k \approx 0.1$ nm$^{-1}$) narrows each cone's effective spectral bandwidth, reducing overlap between adjacent cone classes. Reduced overlap means the visual system can discriminate colors that would appear identical without the filters — the droplets paradoxically increase color discrimination despite reducing total light capture.

Droplet Type	Color	Carotenoid	Cutoff (nm)	Cone type
T (transparent)	Clear	None	<400	UVS/VS
C (clear)	Pale yellow	Trace	~420	SWS
Y (yellow)	Yellow	Zeaxanthin	~480	MWS
R (red)	Orange-red	Astaxanthin	~540	LWS

1.3 UV Vision: Ecological Significance

Many bird species reflect UV light from plumage structures that are completely invisible to the human eye. Blue tits (Cyanistes caeruleus) have UV-reflective crown patches visible to conspecifics for mate assessment. Kestrels (Falco tinnunculus) detect UV-reflective vole urine trails. Many berries and fruits have UV-absorbing or reflecting patterns guiding avian foragers.

The human cornea absorbs UV below ∼400 nm, and the crystalline lens absorbs below ∼320 nm. Birds lack this UV-blocking lens absorption — their lenses are UV-transparent, permitting photons down to ∼320 nm to reach the UVS cones. This represents a significant evolutionary expansion of the perceptual color space.

Figure 1: Avian vs. Human Cone Spectral Sensitivities (with Oil Droplet Filtering)

2. Magnetoreception

Birds can detect the direction and inclination of Earth's magnetic field ($|\mathbf{B}| \approx 25\text{--}65\ \mu\text{T}$). Two distinct biophysical mechanisms have been proposed, with strong evidence for at least one being light-dependent and quantum mechanical in origin.

2.1 Cryptochrome Radical Pair Mechanism

The radical pair mechanism (RPM) was proposed by Schulten et al. (1978) and later identified with cryptochrome proteins (particularly CRY4, expressed in double cones of the right eye in many migratory species).

Step-by-Step Mechanism:

1.Blue light (350–500 nm) is absorbed by the FAD (flavin adenine dinucleotide) cofactor within the CRY4 protein, promoting it to the singlet excited state FAD*.
2.Ultrafast electron transfer (<1 ps) from a nearby Trp (tryptophan) residue to FAD* produces the radical pair $[\text{FAD}^{\bullet-} \cdots \text{TrpH}^{\bullet+}]$. This is initially in the singlet spin state $|S\rangle$.
3.Hyperfine coupling (HFC) between the electron spin and nearby nuclear spins (principally${}^{14}\text{N}$ and ${}^1\text{H}$) drives coherent interconversion between singlet $|S\rangle$ and triplet $|T_0\rangle$ states.
4.The external magnetic field modifies the rate of S–T interconversion by Zeeman splitting of the triplet sublevels $|T_{+1}\rangle, |T_0\rangle, |T_{-1}\rangle$, changing the singlet yield $\Phi_S$ in a direction-dependent manner.
5.Singlet and triplet radical pairs have different chemical fates (different protein conformations / signaling outputs), so $\Phi_S(\theta, \phi)$ provides directional magnetic information to the neural circuitry.

Larmor Precession Frequency:

The fundamental frequency at which electron spins precess about an external field $\mathbf{B}$is the Larmor frequency:

\[ \omega_L = \gamma_e \, B \]

where $\gamma_e = \frac{e}{m_e} = 1.760 \times 10^{11}\ \text{rad s}^{-1} \text{T}^{-1}$ is the electron gyromagnetic ratio (in the free-electron approximation).

For Earth's field $B \approx 50\ \mu\text{T}$:

\[ \omega_L = (1.760 \times 10^{11}\ \text{rad s}^{-1}\text{T}^{-1}) \times (50 \times 10^{-6}\ \text{T}) \approx 8.80 \times 10^{6}\ \text{rad s}^{-1} \]\[ f_L = \frac{\omega_L}{2\pi} \approx 1.40\ \text{MHz} \]

This MHz-scale precession is on the same timescale as radical pair lifetimes (typical $\tau \sim 1$–10 μs), meaning the magnetic field has sufficient time to modulate the S–T interconversion before the radical pair recombines. This is essential for the compass to work.

Singlet Yield as a Function of Field Angle:

Within a simplified model (one electron spin interacting with one nuclear spin via HFC tensor$\mathbf{A}$), the singlet yield depends on the angle $\theta$ between$\mathbf{B}$ and the principal axis of $\mathbf{A}$:

\[ \Phi_S(\theta) = \Phi_S^{(0)} + \Delta\Phi \cdot f(\theta, B/A) \]

For an axially symmetric HFC tensor ($A_\parallel, A_\perp$), the anisotropy$\Delta\Phi(\theta)$ varies as $\cos^2\theta$ in the low-field limit and saturates at high fields.

Crucially, this is an inclination compass: it senses the angle between $\mathbf{B}$ and the body axis, not the polarity. Experiments with oscillating radiofrequency fields at the Larmor frequency disrupted magnetic compass orientation in European robins (Erithacus rubecula), providing the strongest behavioral evidence for the RPM.

2.2 Magnetite-Based Mechanism

Single-domain magnetite (Fe$_3$O$_4$) crystals have been identified in the beak tissue of homing pigeons and other species. These superparamagnetic particles (diameter ∼50 nm) can rotate in response to magnetic fields, potentially deforming mechanosensitive ion channels in associated nerve endings (trigeminal nerve).

The magnetic torque on a single-domain particle of volume $V$ and magnetization$M_s$ at angle $\alpha$ to the field:

\[ \tau = \mu_0 M_s V B \sin\alpha \]

For Fe$_3$O$_4$: $M_s \approx 4.8 \times 10^5\ \text{A m}^{-1}$,$V = \frac{4}{3}\pi(25\,\text{nm})^3 \approx 6.5 \times 10^{-23}\ \text{m}^3$,$B = 50\ \mu\text{T}$:$\tau \approx 1.5\ \text{fN}\cdot\text{m}$ per particle — potentially detectable by mechanosensory channels with pN-scale thresholds if thousands of particles act cooperatively. Recent work suggests magnetite may serve as an intensity sensor rather than a directional compass.

Figure 2: Cryptochrome Radical Pair Mechanism

3. Celestial Navigation

3.1 Star Compass

Migratory birds (e.g., Garden warbler, Indigo bunting) learn to recognize the center of stellar rotation — the celestial pole (near Polaris in the northern hemisphere) — as the directional reference for north during a critical imprinting period as juveniles.

Planetarium experiments by Emlen (1969) with Indigo buntings demonstrated that birds exposed to an artificial sky rotating around a star other than Polaris used that star as their north reference. The compass is learned, not innate, and is calibrated by the pattern of stellar rotation, not individual star patterns.

3.2 Sun Compass

Many birds can use the sun as a directional cue, but the sun's azimuth changes throughout the day (∼15°/hour). Birds compensate using an internal circadian clock, adjusting their interpretation of sun position by approximately the correct rate of apparent solar motion.

Clock-shift experiments: birds kept under artificial light regimes 6 hours fast or slow showed predictable 90° deviations in departure direction under clear skies, confirming the time-compensated sun compass. The clock driving this is the circadian oscillator with ∼24-hour period, entrained by light/dark cycles.

3.3 Polarized Light Detection

Skylight is polarized in a pattern determined by the position of the sun (Rayleigh scattering). The e-vector of polarized light is perpendicular to the sun-zenith-observer plane. Even under overcast conditions, a detectable polarization pattern can persist, providing compass information.

Double cones in the avian retina contain aligned molecules that may detect polarization. The polarization gradient at sunset provides a particularly robust cue used to calibrate other compasses. Polarization information processed through the avian area centralis (equivalent of fovea) may be projected to the visual Wulst for integration with other compass modalities.

4. Python: Radical Pair Singlet Yield & Spectral Sensitivity Models

The following code (1) computes the singlet yield $\Phi_S(\theta)$ as a function of the angle between the magnetic field and the HFC principal axis, illustrating the compass anisotropy, and (2) plots avian vs. human cone spectral sensitivities with oil droplet filtering applied.

Python

script.py148 lines

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec

fig = plt.figure(figsize=(14, 10))
gs = gridspec.GridSpec(2, 2, figure=fig, hspace=0.45, wspace=0.38)

# ============================================================
# Panel A: Radical pair singlet yield vs B-field angle theta
# ============================================================
ax1 = fig.add_subplot(gs[0, :])

theta_deg = np.linspace(0, 180, 360)
theta_rad = np.radians(theta_deg)

# Simplified model: singlet yield modulated by angle between B and HFC axis
# S_yield = S0 + dS * (a + b*cos(2*theta)) -- typical angular dependence
# For Earth field (weak-field limit) with axial HFC tensor
# Anisotropy amplitude depends on ratio x = omega_L / a_hfc
gamma_e = 1.760e11   # rad/s/T
B_earth  = 50e-6     # 50 uT

omega_L = gamma_e * B_earth
f_L     = omega_L / (2 * np.pi)

a_hfc_rad  = 2 * np.pi * 0.5e6  # 0.5 MHz HFC (typical for Trp radical)

x = omega_L / a_hfc_rad  # dimensionless field strength

# Analytical formula for singlet yield in RP model (axial tensor, one nuclear spin I=1/2)
# Phi_S(theta) = S_base + Delta_S * (1 - 3*cos^2(theta)) * g(x)
# g(x) is a damping function that peaks near x~1
def g_factor(x_val):
    return x_val / (1 + x_val**2)

S_base  = 0.50
Delta_S = 0.08 * g_factor(x)

phi_S_B50   = S_base + Delta_S * (1 - 3 * np.cos(theta_rad)**2)

# Also compute for B = 0 (isotropic, no anisotropy)
phi_S_B0    = np.full_like(theta_rad, S_base)

# For B = 500 uT (strong field limit, anisotropy reverses sign)
x2     = gamma_e * 500e-6 / a_hfc_rad
Delta_S_hi = 0.08 * g_factor(x2)
phi_S_B500 = S_base + Delta_S_hi * (1 - 3 * np.cos(theta_rad)**2)

ax1.axhline(S_base, color='gray', linewidth=1, linestyle=':', label='No field (isotropic)')
ax1.plot(theta_deg, phi_S_B50,  color='#38bdf8', linewidth=2.5, label=f'B = 50 uT (Earth)   [omega_L = {f_L/1e6:.2f} MHz]')
ax1.plot(theta_deg, phi_S_B500, color='#f97316', linewidth=2,   linestyle='--', label='B = 500 uT (10x Earth field)')
ax1.axvline(0,   color='#a855f7', alpha=0.4, linewidth=1, linestyle='--')
ax1.axvline(90,  color='#a855f7', alpha=0.4, linewidth=1, linestyle='--')
ax1.axvline(180, color='#a855f7', alpha=0.4, linewidth=1, linestyle='--')
ax1.fill_between(theta_deg, phi_S_B0, phi_S_B50, alpha=0.15, color='#38bdf8')
ax1.set_xlim(0, 180)
ax1.set_xlabel('Angle theta between B-field and HFC principal axis (degrees)', fontsize=11)
ax1.set_ylabel('Singlet Yield Phi_S', fontsize=11)
ax1.set_title('Cryptochrome Radical Pair Compass Anisotropy\n'
              r'$Phi_S(	heta) = Phi_S^{(0)} + DeltaPhi cdot [1 - 3cos^2	heta] cdot g(omega_L/a_{hfc})$', fontsize=12)
ax1.legend(fontsize=9)
ax1.set_xticks([0, 30, 60, 90, 120, 150, 180])
ax1.grid(True, alpha=0.3)
ax1.annotate(f'omega_L/a_hfc = {x:.3f}
omega_L = gamma_e * B = {omega_L/1e6:.2f} Mrad/s\nf_L = {f_L/1e6:.2f} MHz',
             xy=(90, S_base + Delta_S * (1-3*np.cos(np.pi/2)**2)),
             xytext=(130, S_base + 0.06),
             fontsize=8.5, color='#38bdf8',
             arrowprops=dict(arrowstyle='->', color='#38bdf8', lw=1.2))

# ============================================================
# Panel B: Avian vs Human cone sensitivities (Gaussian model)
# ============================================================
ax2 = fig.add_subplot(gs[1, 0])

lam = np.linspace(300, 750, 900)

def gaussian_cone(peak, width=40):
    return np.exp(-0.5 * ((lam - peak) / width)**2)

def oil_droplet_filter(cutoff, k=0.12):
    return 1.0 / (1.0 + np.exp(-k * (lam - cutoff)))

# Avian cones: peak, oil droplet cutoff, color
avian_cones = [
    (365,  320, '#a855f7', 'UVS (365nm)'),
    (440,  410, '#38bdf8', 'SWS (440nm)'),
    (508,  475, '#4ade80', 'MWS (508nm)'),
    (568,  540, '#f97316', 'LWS (568nm)'),
]

for peak, cutoff, col, label in avian_cones:
    raw = gaussian_cone(peak)
    filtered = raw * oil_droplet_filter(cutoff)
    filtered /= filtered.max()
    ax2.plot(lam, filtered, color=col, linewidth=2.2, label=label)
    ax2.fill_between(lam, 0, filtered, alpha=0.1, color=col)

# Human cones (no oil droplets, broader, no UV)
human_cones = [
    (430, '#818cf8', 'S (430nm)', 35),
    (530, '#86efac', 'M (530nm)', 45),
    (560, '#fca5a5', 'L (560nm)', 45),
]
for peak, col, label, width in human_cones:
    raw = gaussian_cone(peak, width)
    raw /= raw.max()
    ax2.plot(lam, raw, color=col, linewidth=1.5, linestyle='--', label=label, alpha=0.75)

ax2.set_xlim(300, 750)
ax2.set_ylim(0, 1.1)
ax2.set_xlabel('Wavelength (nm)', fontsize=10)
ax2.set_ylabel('Relative Sensitivity', fontsize=10)
ax2.set_title('Cone Spectral Sensitivities\n(Avian solid, Human dashed)', fontsize=10)
ax2.legend(fontsize=7.5, loc='upper right', ncol=2)
ax2.axvspan(300, 400, alpha=0.07, color='purple', label='UV')
ax2.grid(True, alpha=0.25)
ax2.text(340, 1.02, 'UV', color='#a855f7', fontsize=8, ha='center')
ax2.text(550, 1.02, 'Visible', color='#e2e8f0', fontsize=8, ha='center')

# ============================================================
# Panel C: Larmor frequency vs B field
# ============================================================
ax3 = fig.add_subplot(gs[1, 1])

B_range = np.linspace(5, 500, 500) * 1e-6  # 5 to 500 uT
omega_range = gamma_e * B_range
f_range     = omega_range / (2 * np.pi) / 1e6  # MHz

ax3.plot(B_range * 1e6, f_range, color='#38bdf8', linewidth=2.5)
ax3.axvline(50, color='#f59e0b', linewidth=2, linestyle='--', label='Earth field (50 uT)')
ax3.axhline(f_L / 1e6, color='#f59e0b', linewidth=1, linestyle=':', alpha=0.6)
ax3.scatter([50], [f_L/1e6], color='#fbbf24', zorder=5, s=80)
ax3.annotate(f'f_L = {f_L/1e6:.2f} MHz\nat 50 uT', xy=(50, f_L/1e6),
             xytext=(150, f_L/1e6 + 6), fontsize=9, color='#fbbf24',
             arrowprops=dict(arrowstyle='->', color='#fbbf24'))
ax3.axhspan(0.1, 10, alpha=0.08, color='#4ade80')
ax3.text(400, 5, 'RP lifetime\nwindow', color='#4ade80', fontsize=8, ha='center')
ax3.set_xlabel('Magnetic Field Strength (uT)', fontsize=10)
ax3.set_ylabel('Larmor Frequency (MHz)', fontsize=10)
ax3.set_title('Electron Larmor Frequency\nomega_L = gamma_e * B', fontsize=10)
ax3.legend(fontsize=9)
ax3.grid(True, alpha=0.3)
ax3.set_xlim(5, 500)

fig.suptitle('Avian Magnetoreception & Vision: Biophysical Models', fontsize=14, fontweight='bold', y=1.01)
plt.savefig('output.png', dpi=130, bbox_inches='tight', facecolor='#0f172a')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

5. Compass Hierarchy & Neural Integration

Migratory birds possess multiple redundant compasses (magnetic, stellar, solar, polarized light) that are hierarchically calibrated: celestial cues (stellarrotation > sunset polarization > sun azimuth) take precedence over the magnetic compass in most species tested. The magnetic compass is recalibrated daily at sunrise/sunset by celestial information.

Magnetic Compass

Inclination type. CRY4 radical pairs in right-eye double cones. Processed in cluster N (nocturnal migrants) or visual Wulst.

Stellar Compass

Learned during juvenile sensitive period. Pattern of rotation, not star identity. Emlen funnel experiments confirm.

Sun Compass

Time-compensated by circadian clock. Disrupted by clock-shifting. Used mainly by diurnal migrants; calibrates magnetic compass.

Map vs. Compass: True Navigation

A compass tells direction; a map tells position. True navigation (returning to a specific location from an unknown displacement) requires both. Homing pigeons (Columba livia) use a bi-coordinate map based on olfactory gradients and/or geomagnetic field gradients (total intensity and inclination angle each vary with latitude/longitude, providing a unique geomagnetic address at most locations). The map sense is likely separate from and more complex than the compass sense.

References

Hart, N. S. & Hunt, D. M. (2007). Avian visual pigments: characteristics, spectral tuning, and evolution. American Naturalist, 169, S7–S26.
Huth, T. & Ritz, T. (2012). The mechanism of the avian magnetic compass. Physics of Life Reviews, 9, 55–64.
Ritz, T., Adem, S. & Schulten, K. (2000). A model for photoreceptor-based magnetoreception in birds. Biophysical Journal, 78, 707–718.
Mouritsen, H. (2018). Long-distance navigation and magnetoreception in migratory animals. Nature, 558, 50–59.
Goldsmith, T. H. (2006). What birds see. Scientific American, 295, 68–75.
Wiltschko, R. & Wiltschko, W. (2014). Sensing magnetic directions in birds: radical pair processes involving cryptochrome. Biosensors, 4, 221–242.
Emlen, S. T. (1975). The stellar-orientation system of a migratory bird. Scientific American, 233, 102–111.
Gill, F. B. (2007). Ornithology, 3rd ed. W. H. Freeman.

←Metabolic Energetics Egg Biochemistry→