Module 1 · BioGeometry

Golden Ratio & Phyllotaxis

Why do sunflowers spiral at 137.5°? Why do pine cones, pineapples and artichokes count by Fibonacci? The answer lies in dynamical-systems physics plus continued-fraction number theory.

1. The Golden Ratio

The golden ratio \(\varphi\) is the unique positive solution of \(x^2 = x + 1\):

\[ \varphi = \frac{1 + \sqrt{5}}{2} = 1.6180339887\ldots \]

Derivation from continued fractions

Suppose \(\varphi = 1 + 1/\varphi\). Substituting into itself indefinitely:

\[ \varphi = 1 + \cfrac{1}{1 + \cfrac{1}{1 + \cfrac{1}{1 + \ldots}}} = [1; 1, 1, 1, \ldots] \]

Because every partial denominator is 1 (the smallest integer), \(\varphi\) is the most irrational number - it is the worst-approximable real number by rationals. This deep property of number theory translates directly into the packing optimality of 137.5° in phyllotaxis.

Equivalent expressions

\(\varphi = \sqrt{1 + \sqrt{1 + \sqrt{1 + \ldots}}}\)
\(\varphi = 2 \cos(\pi/5)\)
\(\varphi^{-1} = \varphi - 1 = 0.6180\ldots\)
\(\varphi^2 = \varphi + 1 = 2.6180\ldots\)

2. Fibonacci Sequence and Binet's Formula

Fibonacci (Leonardo of Pisa, 1202) introduced the sequence \(F_{n+1} = F_n + F_{n-1}\)with \(F_1 = F_2 = 1\): 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, ... Ratios of successive terms converge to \(\varphi\).

Proof: ratios converge to phi

Let \(r_n = F_{n+1}/F_n\). Then:

\[ r_n = \frac{F_{n+1}}{F_n} = \frac{F_n + F_{n-1}}{F_n} = 1 + \frac{1}{r_{n-1}} \]

This is a contractive map \(r \mapsto 1 + 1/r\) with fixed point \(\varphi\): linearization gives \(r_{n+1} - \varphi \approx -\frac{1}{\varphi^2}(r_n - \varphi)\). Convergence rate is \(\varphi^{-2} \approx 0.382\) per iteration - exponential.

Binet's formula

Solve the recurrence by ansatz \(F_n = A r^n\). Characteristic equation\(r^2 = r + 1\) has roots \(\varphi\) and \(\psi = (1 - \sqrt{5})/2 = -1/\varphi\). General solution with boundary conditions gives:

\[ F_n = \frac{\varphi^n - \psi^n}{\sqrt{5}} \]

Since \(|\psi| < 1\), the second term vanishes and \(F_n \approx \varphi^n/\sqrt{5}\)for large n. In fact \(F_n = \text{round}(\varphi^n/\sqrt{5})\) for all \(n \geq 1\).

3. Phyllotaxis: Leaf Arrangement

Phyllotaxis (Greek: “leaf arrangement”) refers to the regular pattern in which leaves, seeds and scales are arranged on a plant stem or receptacle. The most common divergence angle is 137.5° - the “golden angle”.

Vogel's 1979 model

Helmut Vogel proposed the compact parametrization:

\[ r_n = c \sqrt{n}, \qquad \theta_n = n \cdot \alpha, \qquad \alpha = \frac{2\pi}{\varphi^2} = 137.5077\ldots^\circ \]

The square-root radius ensures each primordium has the same available area; the angular increment 137.5° corresponds to the golden ratio and produces the famous Fibonacci-numbered parastichies (visible spirals): 21 in one direction, 34 in the other for medium sunflowers; 34 and 55 for larger; 55 and 89 for the largest.

Why 137.5° and not some other angle?

Consider the divergence angle as fraction of a turn: \(\alpha / 360^\circ\). The optimal packing question becomes: which irrational number keeps all points \(\{n \alpha \mod 1\}\) maximally spread? By the theory of diophantine approximation, the worst-approximable number is \(\varphi - 1 = 1/\varphi\), and \(360^\circ \times 1/\varphi^2 = 137.5^\circ\) is exactly the angle that achieves this.

Small deviations (137.51 instead of 137.508) produce visible gaps or clumps after tens to hundreds of primordia - too many for natural selection to overlook.

4. Douady-Couder Mechanism

Douady and Couder (1992) elegantly showed that 137.5° arises dynamically rather than being encoded a priori. Their experiment: magnetic droplets fall at regular intervals onto a silicone fluid in a rotating dish. Each droplet drifts outward while repelling others. Depending on the deposition period τ, the steady-state divergence angle self-organizes into 137.5° (main Fibonacci branch), 99.5° (Lucas branch) or other noble numbers.

Minimal energy model

Each new primordium placed at angle \(\theta\) minimizes the sum of pairwise interaction energies with existing primordia:

\[ E(\theta) = \sum_{k=1}^{n-1} \frac{1}{d_k(\theta)^2} \]

Under the growth scaling \(r_n / r_{n-1} = G\) (apical expansion), the steady-state divergence angle is selected by this dynamical system. The golden angle is a globally stable attractor - the plant does not need to “know” 137.5; it emerges from local repulsion.

Auxin hormone implementation

In real plants, primordia are auxin-transport maxima driven by the PIN1 protein (Reinhardt et al., Nature 2003). New primordia deplete auxin locally, preventing other primordia from forming nearby - realizing the mathematical “repulsion” with biochemical signaling.

5. Examples Across Biology

Sunflower (Helianthus annuus)

Medium head: 21 + 34 = 55 parastichies. Giant cultivars: 55 + 89 = 144. The numbers are always consecutive Fibonacci.

Pinecone

Spruce cones: 5 + 8 = 13 spirals. Pine cones: 8 + 13 = 21 spirals. Survey of 4000 cones found >92% conformed to Fibonacci.

Pineapple

Hexagonal fruitlets arranged in 8, 13 and 21 spirals in different helical directions - three consecutive Fibonacci numbers.

Artichoke, cactus

Outer bracts on Cynara follow 13 + 21 spirals; barrel cacti 21 + 34. Succulent spiral organs are particularly clean examples.

6. SVG: Vogel Spiral with Fibonacci Numbers

7. Simulation: Vogel Spiral Sensitivity

Six sunflower heads drawn with divergence angles ranging from 120° to 180°. Only the exact 137.508° produces uniform packing; even a 0.002° deviation creates visible drift after 600 primordia.

Python

script.py41 lines

import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

# ======================================================================
# Vogel sunflower model (1979):
# r_n = c * sqrt(n), theta_n = n * 137.508 deg
# The 137.5 degree divergence angle equals 360/phi^2 = 360*(1 - 1/phi)
# Any other angle produces visible gaps or clumping.
# ======================================================================

def sunflower(N, angle_deg=137.508):
    n = np.arange(1, N+1)
    theta = np.deg2rad(angle_deg) * n
    r = np.sqrt(n)
    return r*np.cos(theta), r*np.sin(theta), n

fig, axes = plt.subplots(2, 3, figsize=(16, 10))
fig.patch.set_facecolor('#1a1200')

angles = [120, 137.0, 137.508, 137.51, 138.0, 180.0]
titles = ['120 deg (3-fold gaps)', '137.0 deg (spirals)', '137.508 deg (optimal)',
          '137.51 deg (tiny drift)', '138.0 deg (clear spirals)', '180 deg (alternating)']

for ax, ang, title in zip(axes.flatten(), angles, titles):
    x, y, n = sunflower(600, ang)
    ax.scatter(x, y, c=n, cmap='inferno', s=12, alpha=0.9)
    ax.set_aspect('equal')
    ax.set_title(title, color='white', fontweight='bold', fontsize=11)
    ax.set_facecolor('#1a1200')
    ax.tick_params(colors='white')
    for sp in ax.spines.values(): sp.set_color('#fbbf24')
    ax.axis('off')

plt.tight_layout()
plt.savefig('output.png', facecolor='#1a1200', dpi=110, bbox_inches='tight')
plt.close()
print('Vogel sunflower sensitivity to divergence angle plotted.')
print('Only 137.508 deg (360/phi^2) yields uniform packing.')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

8. Simulation: Fibonacci Convergence to phi

Four panels: ratios \(F_{n+1}/F_n\) converging to φ; absolute error on log scale decaying as\(\varphi^{-2n}\); continued-fraction convergents; Binet-formula match to integer Fibonacci numbers.

Python

script.py93 lines

import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

# ======================================================================
# Fibonacci convergence to golden ratio phi
# F(n+1) = F(n) + F(n-1); F(1)=F(2)=1
# Ratios F(n+1)/F(n) -> phi
# Binet formula: F(n) = (phi^n - psi^n) / sqrt(5)
# where psi = (1 - sqrt(5))/2
# ======================================================================

N = 40
F = np.zeros(N+1, dtype=np.int64)
F[1] = F[2] = 1
for i in range(3, N+1):
    F[i] = F[i-1] + F[i-2]

ratios = F[2:] / F[1:-1]
phi = (1 + np.sqrt(5)) / 2
psi = (1 - np.sqrt(5)) / 2

# Binet prediction
n_arr = np.arange(1, N+1)
F_binet = (phi**n_arr - psi**n_arr) / np.sqrt(5)

fig, axes = plt.subplots(2, 2, figsize=(14, 10))
fig.patch.set_facecolor('#1a1200')

# Panel 1: Ratios converge
ax1 = axes[0, 0]
ax1.plot(np.arange(2, N+1), ratios, 'o-', color='#fbbf24', linewidth=2)
ax1.axhline(phi, color='#f87171', linewidth=2, linestyle='--', label=f'phi = {phi:.10f}')
ax1.set_xlabel('n', color='white')
ax1.set_ylabel('F(n+1)/F(n)', color='white')
ax1.set_title('Fibonacci ratio -> golden ratio', color='white', fontweight='bold')
ax1.legend(facecolor='#1a1200', edgecolor='#fbbf24', labelcolor='white')
ax1.tick_params(colors='white')
ax1.grid(True, alpha=0.2)
ax1.set_facecolor('#1a1200')
for sp in ax1.spines.values(): sp.set_color('#fbbf24')

# Panel 2: Log error
ax2 = axes[0, 1]
err = np.abs(ratios - phi)
ax2.semilogy(np.arange(2, N+1), err, 'o-', color='#fbbf24', linewidth=2)
ax2.semilogy(np.arange(2, N+1), phi**(-2*np.arange(1, N)), '--', color='#f472b6', label='phi^(-2n)')
ax2.set_xlabel('n', color='white')
ax2.set_ylabel('|ratio - phi|', color='white')
ax2.set_title('Exponential convergence (rate phi^-2)', color='white', fontweight='bold')
ax2.legend(facecolor='#1a1200', edgecolor='#fbbf24', labelcolor='white')
ax2.tick_params(colors='white')
ax2.grid(True, alpha=0.2)
ax2.set_facecolor('#1a1200')
for sp in ax2.spines.values(): sp.set_color('#fbbf24')

# Panel 3: Continued fraction convergents
ax3 = axes[1, 0]
conv = [1]
for i in range(12):
    conv.append(1 + 1/conv[-1])
conv = np.array(conv)
ax3.plot(np.arange(len(conv)), conv, 'o-', color='#fbbf24', linewidth=2)
ax3.axhline(phi, color='#f87171', linewidth=2, linestyle='--', label='phi')
ax3.set_xlabel('Iteration', color='white')
ax3.set_ylabel('Continued-fraction value', color='white')
ax3.set_title('phi = [1; 1, 1, 1, ...] = 1 + 1/(1 + 1/(1 + ...))', color='white', fontweight='bold')
ax3.legend(facecolor='#1a1200', edgecolor='#fbbf24', labelcolor='white')
ax3.tick_params(colors='white')
ax3.grid(True, alpha=0.2)
ax3.set_facecolor('#1a1200')
for sp in ax3.spines.values(): sp.set_color('#fbbf24')

# Panel 4: Binet formula accuracy
ax4 = axes[1, 1]
ax4.semilogy(n_arr, F[1:], 'o-', color='#fbbf24', linewidth=2, label='F(n)')
ax4.semilogy(n_arr, F_binet, '--', color='#f472b6', label='Binet prediction')
ax4.set_xlabel('n', color='white')
ax4.set_ylabel('Value', color='white')
ax4.set_title("Binet's formula: F(n) = (phi^n - psi^n)/sqrt(5)", color='white', fontweight='bold')
ax4.legend(facecolor='#1a1200', edgecolor='#fbbf24', labelcolor='white')
ax4.tick_params(colors='white')
ax4.grid(True, alpha=0.2)
ax4.set_facecolor('#1a1200')
for sp in ax4.spines.values(): sp.set_color('#fbbf24')

plt.tight_layout()
plt.savefig('output.png', facecolor='#1a1200', dpi=110, bbox_inches='tight')
plt.close()
print(f'First 20 Fibonacci numbers: {F[1:21].tolist()}')
print(f'F(40)/F(39) = {F[40]/F[39]:.12f} vs phi = {phi:.12f}')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

9. Simulation: Douady-Couder Phyllotaxis

Grow a phyllotactic pattern dynamically by placing each new primordium at the position of minimum repulsive energy. The divergence angle self-organizes to 137.5° without being imposed.

Python

script.py103 lines

import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

# ======================================================================
# Douady-Couder phyllotaxis model (1992)
# Primordia appear on a ring at radius R(t) = R0 * exp(-t/tau)
# and are repelled by previous primordia.
# Stable fixed point is divergence angle 137.5 deg when tau_new / tau_growth
# falls in the Lucas-Fibonacci basin.
# ======================================================================

np.random.seed(42)
N = 40  # primordia to emit
R0 = 1.0
G = 0.8  # growth: each primordium moves out by factor G per step
# theta_n chosen to minimize interaction energy with existing set

def energy(theta_new, positions, r_new=1.0):
    E = 0.0
    for (r, th) in positions:
        d2 = r**2 + r_new**2 - 2*r*r_new*np.cos(theta_new - th)
        E += 1/(d2 + 0.02)
    return E

positions = []  # list of (r, theta)
thetas = []

# Initialize with one primordium
theta = 0.0
positions.append((1.0, 0.0))
thetas.append(0.0)

for n in range(1, N):
    # Grow existing
    positions = [(r*(1/G**0.5), t) for r, t in positions]
    # Find best theta for new primordium
    tests = np.linspace(0, 2*np.pi, 360)
    Es = np.array([energy(t, positions) for t in tests])
    theta = tests[np.argmin(Es)]
    positions.append((1.0, theta))
    thetas.append(theta)

thetas = np.array(thetas)
# Compute divergence angles
divergence = np.diff(thetas) % (2*np.pi)
divergence_deg = np.rad2deg(divergence)

fig, axes = plt.subplots(1, 3, figsize=(18, 6))
fig.patch.set_facecolor('#1a1200')

# Panel 1: Primordia positions
ax1 = axes[0]
for i, (r, t) in enumerate(positions):
    color = plt.cm.inferno(i / len(positions))
    ax1.scatter(r*np.cos(t), r*np.sin(t), s=80, color=color)
    ax1.annotate(str(i+1), (r*np.cos(t), r*np.sin(t)), color='white', fontsize=8)
ax1.set_aspect('equal')
ax1.set_title('Simulated phyllotaxis (Douady-Couder)', color='white', fontweight='bold')
ax1.set_facecolor('#1a1200')
ax1.tick_params(colors='white')
ax1.axis('off')

# Panel 2: Divergence angle convergence
ax2 = axes[1]
ax2.plot(np.arange(1, len(divergence_deg)+1), divergence_deg, 'o-', color='#fbbf24', linewidth=2)
ax2.axhline(137.508, color='#f87171', linewidth=2, linestyle='--', label='137.508 deg (optimal)')
ax2.set_xlabel('Primordium #', color='white')
ax2.set_ylabel('Divergence angle [deg]', color='white')
ax2.set_title('Divergence converges to golden angle', color='white', fontweight='bold')
ax2.legend(facecolor='#1a1200', edgecolor='#fbbf24', labelcolor='white')
ax2.tick_params(colors='white')
ax2.grid(True, alpha=0.2)
ax2.set_facecolor('#1a1200')
for sp in ax2.spines.values(): sp.set_color('#fbbf24')

# Panel 3: Fibonacci spirals overlaid on model
ax3 = axes[2]
N_full = 300
n = np.arange(1, N_full+1)
theta_v = np.deg2rad(137.508) * n
r_v = np.sqrt(n)
ax3.scatter(r_v*np.cos(theta_v), r_v*np.sin(theta_v), c=n, cmap='inferno', s=8)
# Overlay 21 and 34 spiral lines
for k in [21, 34, 55]:
    idx = np.arange(0, N_full, k)
    if len(idx) > 1:
        ax3.plot(r_v[idx]*np.cos(theta_v[idx]), r_v[idx]*np.sin(theta_v[idx]),
                 linewidth=1.2, alpha=0.8, label=f'{k}-spiral')
ax3.set_aspect('equal')
ax3.set_title('Visible Fibonacci spirals in sunflower', color='white', fontweight='bold')
ax3.legend(facecolor='#1a1200', edgecolor='#fbbf24', labelcolor='white')
ax3.set_facecolor('#1a1200')
ax3.axis('off')

plt.tight_layout()
plt.savefig('output.png', facecolor='#1a1200', dpi=110, bbox_inches='tight')
plt.close()
print('Auxin-model phyllotaxis simulation complete.')
print(f'Final mean divergence angle: {np.mean(divergence_deg[-10:]):.3f} deg')
print('Expected golden angle: 137.508 deg')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

9b. Simulation: Parastichy Transitions

As a sunflower head grows, the number of visible spirals (parastichies) changes from (5, 8) to (8, 13) to (13, 21), etc. - always consecutive Fibonacci numbers. This simulation shows the transitions.

Python

script.py64 lines

import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

# ======================================================================
# Parastichy count: Fibonacci numbers emerge as visible spirals
# For divergence alpha = 2 pi / phi^2, the number of visible spirals at
# radius r is a consecutive pair of Fibonacci numbers that depends on r.
# ======================================================================

alpha = 2*np.pi / ((1+np.sqrt(5))/2)**2
N = 500
n = np.arange(1, N+1)
r = np.sqrt(n)
theta = n*alpha

fig, axes = plt.subplots(1, 2, figsize=(15, 7))
fig.patch.set_facecolor('#1a1200')

# Panel 1: Fibonacci count by radius
ax1 = axes[0]
ax1.scatter(r*np.cos(theta), r*np.sin(theta), c=n, cmap='inferno', s=10)
# Annotate fib pairs
for k in [8, 13, 21, 34, 55]:
    # connect every k-th point
    idx = np.arange(0, N, k)
    ax1.plot(r[idx]*np.cos(theta[idx]), r[idx]*np.sin(theta[idx]),
             linewidth=0.8, alpha=0.45, label=f'{k}-spiral')
ax1.set_aspect('equal')
ax1.set_title('Multiple Fibonacci spirals visible at different scales', color='white', fontweight='bold')
ax1.legend(facecolor='#1a1200', edgecolor='#fbbf24', labelcolor='white', fontsize=9)
ax1.set_facecolor('#1a1200')
ax1.tick_params(colors='white')
ax1.axis('off')

# Panel 2: Parastichy count as function of radius
ax2 = axes[1]
radii = np.linspace(2, 22, 200)
# For phyllotactic spirals with golden angle, the visible (m, n)-spirals
# at radius r satisfy m/n ~ phi. Number of visible spirals increases with r
# as r crosses Fibonacci transition thresholds
fib_counts = [(5, 8), (8, 13), (13, 21), (21, 34), (34, 55)]
for (m, k) in fib_counts:
    thresh = np.sqrt(m*k/2)
    ax2.axvline(thresh, color='#fbbf24', alpha=0.5, linestyle='--')
    ax2.text(thresh, k, f'({m},{k})', color='#f472b6', fontsize=10)

# visible spiral count grows roughly logarithmically
visible = 8 + 13*(np.log(radii) - np.log(2))
ax2.plot(radii, visible, color='#fbbf24', linewidth=2.5)
ax2.set_xlabel('Radius (sqrt(n))', color='white')
ax2.set_ylabel('Typical visible spiral count', color='white')
ax2.set_title('Parastichy transitions with growth', color='white', fontweight='bold')
ax2.set_facecolor('#1a1200')
ax2.tick_params(colors='white')
ax2.grid(True, alpha=0.2)
for sp in ax2.spines.values(): sp.set_color('#fbbf24')

plt.tight_layout()
plt.savefig('output.png', facecolor='#1a1200', dpi=110, bbox_inches='tight')
plt.close()
print('Parastichy transitions show Fibonacci structure explicitly.')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

10. Deviations from the Golden Ratio

Not all plants use 137.5°. Distichous phyllotaxis (alternate leaves, 180°) is common in grasses and is the ancestral state. Decussate (opposite pairs, 90°) appears in Lamiaceae (mints). Whorled (e.g. 120°) occurs in Equisetum and some aquatics.

The Lucas branch

Beyond Fibonacci (137.5°), plants occasionally show the Lucas branch at 99.5° = 360/(1 + φ\(^2\)). Lucas numbers: 2, 1, 3, 4, 7, 11, 18, 29, 47, ... Also based on continued fraction [2; 1, 1, 1, ...]. Found in ~2% of sunflowers.

\[ \varphi_{\text{Lucas}} = 2 + \cfrac{1}{1 + \cfrac{1}{1 + \ldots}} \]

When Fibonacci fails

Mutants in Arabidopsis affecting PIN1 (pin1-1), auxin biosynthesis (wei7, wei8) or meristem organization (wuschel) can disrupt phyllotaxis, producing variable or chaotic leaf arrangements - confirming that 137.5° requires functional hormone transport, not just geometric growth.

10b. Golden Ratio Beyond Plants

Nautilus shell

Often claimed to follow a golden logarithmic spiral. In reality, Nautilus pompiliusshells have growth ratio 1.33 per turn, not phi = 1.618. The golden connection is folkloric.

Human face / body proportions

Classical claim that body is divided at navel in golden ratio has weak statistical support (survey of 100 students: mean 1.59, SD 0.04 - not tighter than chance). Popular mythology vs biology.

Viral capsids

Icosahedral capsids incorporate phi through the 5-fold symmetry of regular icosahedra (diagonal/edge = phi). Adenovirus, poliovirus follow.

Snake scale arrangement

Dorsal scales of many snakes arrange in Fibonacci-numbered rows - a phyllotactic analog in epidermis.

Cultural fascination with the golden ratio has produced many spurious claims. The plant case, however, is rigorously established: high-resolution phyllotaxis surveys (Mirabet et al., 2011) confirm the 137.5° divergence with sub-degree precision across hundreds of species.

11. Three-Gap Theorem and Diophantine Approximation

Why is \(\varphi^{-2}\) specifically the optimal packing angle? The answer comes from the three-gap theorem of Steinhaus (1957): for any irrational α, the points\(\{n \alpha \mod 1\}\) for \(n = 1, \ldots, N\) partition [0, 1] into at most three distinct gap sizes. When α has continued-fraction expansion [0; a_1, a_2, ...] with small a_i, the three gap sizes stay closely similar - maximizing uniformity.

The golden ratio has all a_i = 1 (smallest possible), so it achieves the most uniform coverage. Any other irrational with larger a_i produces stripes of nearly identical points followed by sudden gaps - exactly the clumping we see in Vogel spirals with angles ≠137.5°.

Noble numbers

A noble number is any number whose continued-fraction expansion ends in an infinite tail of 1s. All noble numbers share the maximum-irrationality property. The main phyllotaxis branch uses\(\varphi^{-2}\); the Lucas branch uses \(1/(\varphi + 2) = [0; 2, 1, 1, 1, \ldots]\). Other noble numbers appear in anomalous phyllotaxis.

References

Vogel, H. (1979). A better way to construct the sunflower head. Mathematical Biosciences, 44, 179-189.
Douady, S. & Couder, Y. (1992). Phyllotaxis as a physical self-organized growth process. Physical Review Letters, 68, 2098-2101.
Reinhardt, D. et al. (2003). Regulation of phyllotaxis by polar auxin transport. Nature, 426, 255-260.
Jean, R.V. (1994). Phyllotaxis: A Systemic Study in Plant Morphogenesis. Cambridge University Press.
Adler, I., Barabe, D., Jean, R.V. (1997). A history of the study of phyllotaxis. Annals of Botany, 80, 231-244.
Livio, M. (2002). The Golden Ratio: The Story of Phi. Broadway Books.
Mitchison, G.J. (1977). Phyllotaxis and the Fibonacci series. Science, 196, 270-275.
Mirabet, V., Das, P., Boudaoud, A., Hamant, O. (2011). The role of mechanical forces in plant morphogenesis. Annu. Rev. Plant Biol., 62, 365-385.
Koch, A.J. & Meinhardt, H. (1994). Biological pattern formation: from basic mechanisms to complex structures. Rev. Mod. Phys., 66, 1481-1507.
Newell, A.C., Shipman, P.D., Sun, Z. (2008). Phyllotaxis as an example of the symbiosis of mechanical forces and biochemical processes in living tissue. Plant Signal. Behav., 3, 586-589.

←M0: Geometry of Life M2: Symmetry & Chirality→