Module 8

Evolution & Diversity

Spider phylogenetics, web evolution, ballooning dispersal, venom neofunctionalization, and conservation

With over 48,000 described species in 130 families, spiders represent one of the most successful predatory radiations in animal history. Originating ~380 million years ago in the Devonian, they have colonized every terrestrial habitat from deserts to mountaintops to cave systems. Their key innovation — silk — evolved before webs, and web architecture has been gained, modified, and lost multiple times across the phylogeny. Understanding spider evolution requires integrating molecular phylogenetics, biomechanical innovation, and biogeographic dispersal — including the remarkable phenomenon of aerial ballooning.

8.1 Spider Diversity & Taxonomy

As of 2024, the World Spider Catalog lists >48,000 species in >4,200 genera and 130 families. New species are described at a rate of ~400 per year, and rarefaction analysis suggests the true diversity may exceed 120,000 species — meaning we have described less than half. All spiders are predatory, with the sole exception of Bagheera kiplingi, a jumping spider that feeds primarily on the protein-rich Beltian bodies of Acacia trees in Central America.

Species Accumulation & Rarefaction

The rate of species discovery follows a characteristic curve. If we model the discovery process as sampling from a large species pool, the expected number of species discovered after \(n\)sampling efforts follows a rarefaction curve:

Species Accumulation Model

\[ S(n) = S_{\max}\left(1 - e^{-n/\tau}\right) \]

where \(S_{\max}\) is the asymptotic total species richness and \(\tau\)is the characteristic discovery time. With current data (\(S \approx 48{,}000\),\(dS/dt \approx 400/\text{yr}\)), extrapolation gives\(S_{\max} \approx 120{,}000\text{--}170{,}000\).

Major Clades

Mesothelae — Liphistiidae (trapdoor spiders with segmented abdomens, ~100 species). Basal lineage, restricted to East/Southeast Asia.
Mygalomorphae — tarantulas, trapdoor spiders, funnel-web spiders (~3,000 species). Paraxial chelicerae, mostly ground-dwelling.
Araneomorphae — "true spiders" (~45,000 species, >90% of diversity). Diaxial chelicerae, most web-builders.

8.2 Orb Web Evolution

The orb web is often considered the pinnacle of web architecture — a geometrically regular, functionally optimized prey capture device. But did it evolve once or multiple times? This question has driven one of the most contentious debates in spider systematics.

The Monophyly vs. Convergence Debate

Two superfamilies build orb webs: Araneoidea (ecribellate webs with viscid glue spirals) and Deinopoidea (cribellate webs with dry hackled-band spirals). The debate centers on whether their orb webs are homologous (single origin) or convergent (independent origins):

Single origin hypothesis: Molecular phylogenetics (Bond et al. 2014; Fernandez et al. 2018) places Araneoidea and Deinopoidea within Orbiculariae, suggesting a single origin with subsequent diversification of silk types.
Multiple loss hypothesis: Many lineages within Araneoidea have secondarily lost the orb web (cobweb spiders, sheet-web spiders), suggesting web architecture is evolutionarily labile.

Ancestral State Reconstruction

Maximum likelihood ancestral state reconstruction uses Markov models of character evolution mapped onto a phylogeny. For a binary character (orb web present/absent) on a tree with \(N\) tips, the likelihood of the data given transition rates \(q_{01}\) (gain) and\(q_{10}\) (loss) is:

Continuous-Time Markov Model

\[ \mathbf{Q} = \begin{pmatrix} -q_{01} & q_{01} \\ q_{10} & -q_{10} \end{pmatrix}, \qquad P(t) = e^{\mathbf{Q}t} \]

The transition probability matrix over branch length \(t\) is\(P(t) = e^{\mathbf{Q}t}\). The likelihood is computed by the pruning algorithm (Felsenstein 1981), propagating conditional likelihoods from tips to root. If \(q_{10} \gg q_{01}\), the model favors a scenario where the orb web was gained once and lost many times — consistent with current molecular evidence.

8.3 Ballooning Dispersal

Ballooning is the aerial dispersal of spiders using silk threads to catch wind and/or atmospheric electric fields. Darwin observed ballooning spiders landing on HMS Beagle over 100 km from the nearest land. Mass ballooning events ("gossamer") can involve millions of spiderlings coating fields with silk on calm autumn mornings.

Physics of Ballooning

A ballooning spider experiences three main forces: gravity, aerodynamic drag on the silk thread, and (recently discovered) electrostatic force from the atmospheric potential gradient. The force balance:

Force Balance on Ballooning Spider

\[ m g = F_{\text{drag}} + F_{\text{electric}} = \frac{1}{2} C_D \rho_{\text{air}} v^2 A_{\text{silk}} + q E \]

where \(m \approx 1\,\text{mg}\) is the spider mass, \(C_D\) is the drag coefficient of the silk thread, \(A_{\text{silk}} = d \times L\) is the projected area of the silk thread (diameter \(d \approx 1\,\mu\text{m}\), length\(L \approx 1\text{--}3\,\text{m}\)), \(q\) is the charge on the silk, and \(E \approx 100\,\text{V/m}\) is the fair-weather atmospheric electric field.

Electrostatic Ballooning

Morley & Robert (2018) demonstrated that spiders can sense and exploit atmospheric electric fields for ballooning. Silk threads acquire negative charge through triboelectric effects and contact with the negatively charged ground surface. In the Earth's atmospheric potential gradient (\(\sim 100\,\text{V/m}\) fair weather, up to \(\sim 10\,\text{kV/m}\)near thunderstorms), this charge experiences an upward electric force:

\[ F_E = qE = \sigma A_{\text{silk}} E \]

where \(\sigma \approx 10^{-5}\,\text{C/m}^2\) is the surface charge density on silk. For a 2 m thread of diameter 1 \(\mu\)m at \(E = 1\,\text{kV/m}\):\(F_E \approx 10^{-5} \times \pi \times 10^{-6} \times 2 \times 10^3 \approx 6 \times 10^{-8}\,\text{N}\). This is comparable to \(mg \approx 10^{-6} \times 9.81 \approx 10^{-5}\,\text{N}\)for a 1 mg spider only in enhanced fields, but combined with wind drag, electric forces significantly reduce the minimum wind speed needed.

Maximum Dispersal Range

The horizontal range of a ballooning spider depends on the wind speed profile with altitude, the terminal settling velocity, and flight duration:

\[ R = \int_0^{T} v_{\text{wind}}(z(t))\,dt, \qquad z(t) = \frac{F_{\text{up}} - mg}{6\pi\mu r_{\text{eff}}} t \]

With updrafts of ~1 m/s carrying spiders to altitudes of 3–5 km (recorded by aircraft surveys), and wind speeds of 10–30 m/s at altitude, dispersal ranges of 100–1,600 km are theoretically achievable — explaining the colonization of remote oceanic islands.

8.4 Ancient Lineage & Molecular Clock

Spiders (order Araneae) originated approximately 380 MYA in the Devonian period. The earliest fossil spiders include Attercopus fimbriunguis from the late Devonian (~380 MYA) andPalaeothele from the Carboniferous (~300 MYA). Silk glands evolved before webs — originally used for egg sacs, draglines, and burrow lining. The first true aerial webs appeared ~200 MYA in the Triassic.

Molecular Clock Estimation

Divergence times are estimated using relaxed molecular clock models. The substitution rate is modeled as varying across branches according to an autocorrelated or uncorrelated log-normal distribution. For a branch of length \(t_i\) with rate \(r_i\), the expected number of substitutions is:

Relaxed Molecular Clock

\[ d_i = r_i \cdot t_i, \qquad \ln(r_i) \sim \mathcal{N}(\mu_r, \sigma_r^2) \]

Calibration points from the fossil record (e.g., Attercopus at ~380 MYA, amber-preserved Araneoidea at ~130 MYA) constrain the model. Bayesian inference with MCMC sampling yields posterior distributions for divergence times. Key estimates: Mygalomorphae split from Araneomorphae ~300 MYA; major araneomorph radiation ~180–120 MYA; orb web origin ~160–200 MYA.

8.5 Venom Evolution & Molecular Adaptation

Spider venom is a complex cocktail containing hundreds of bioactive peptides that have evolved through a process of gene duplication followed by neofunctionalization. Ancestral toxin genes duplicate via unequal crossing-over, and the copies diverge under positive selection to target different molecular targets (ion channels, receptors, enzymes).

Detecting Positive Selection: Ka/Ks Ratio

The ratio of nonsynonymous (\(K_a\), amino acid-changing) to synonymous (\(K_s\), silent) substitution rates reveals the mode of molecular evolution:

Ka/Ks (dN/dS) Selection Test

\[ \omega = \frac{K_a}{K_s} = \frac{d_N}{d_S} \begin{cases} < 1 & \text{purifying selection (conserved)} \\ = 1 & \text{neutral evolution} \\ > 1 & \text{positive selection (adaptive divergence)} \end{cases} \]

Spider venom toxin genes consistently show \(\omega > 1\) at functionally important residues (the "pharmacophore" — the surface that contacts the target ion channel), while the structural scaffold (cystine knot framework) remains under purifying selection (\(\omega < 0.3\)). This is a hallmark of modular evolution.

Convergent Toxin Evolution

Remarkably, unrelated spider families have independently evolved similar toxin architectures. The inhibitor cystine knot (ICK) motif appears in toxins from Hexathelidae (funnel-web spiders), Theridiidae (black widows), and Theraphosidae (tarantulas) — families that diverged >200 MYA. This convergence reflects the limited structural solutions for stable, disulfide-rich peptides that can target highly conserved ion channel architectures.

8.6 Conservation & Biodiversity Threats

The vast majority of spider species remain undescribed, making conservation assessment extremely challenging. The IUCN Red List includes only ~200 spider species, yet habitat destruction threatens thousands of unknown taxa — particularly in tropical forests, caves, and island ecosystems.

Cave Spiders as Ecosystem Indicators

Troglobiontic (cave-obligate) spiders are particularly vulnerable because they have extremely restricted ranges, often a single cave system. They serve as indicators of karst ecosystem health. Species like Trogloraptor marchingtoni (discovered in Oregon caves in 2012 — a newfamily, Trogloraptoridae) highlight how much diversity remains undiscovered even in well-studied regions.

~60% of spider species are known from fewer than 5 specimens
Many species are described from single localities (micro-endemics)
Island spiders face extreme extinction risk from invasive species
Hawaiian Tetragnatha radiation: ~60 endemic species, several already extinct
Cave spiders cannot disperse if their habitat is destroyed

Computational Models

Species Rarefaction, Ballooning Dispersal, Molecular Clock & Venom Evolution

Python

script.py245 lines

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

fig, axes = plt.subplots(2, 2, figsize=(14, 10))

# ========== Panel 1: Species Accumulation / Rarefaction Curve ==========
ax1 = axes[0, 0]

# Historical species description data (approximate)
years = np.arange(1757, 2025)
# Logistic-like accumulation
S_max_est = 120000
tau = 150  # characteristic time in years
t_start = 1757
S_described = S_max_est * (1 - np.exp(-(years - t_start) / tau))

# Actual data approximation (slower early, faster mid, still climbing)
S_actual = np.zeros_like(years, dtype=float)
for i, yr in enumerate(years):
    if yr < 1850:
        rate = 20
    elif yr < 1920:
        rate = 80
    elif yr < 1970:
        rate = 200
    elif yr < 2000:
        rate = 350
    else:
        rate = 400
    S_actual[i] = S_actual[i-1] + rate + np.random.normal(0, rate * 0.1) if i > 0 else 100

# Clip to realistic range
S_actual = np.clip(S_actual, 0, 48500)

ax1.plot(years, S_actual, color='#a78bfa', linewidth=2.5, label='Described species')
ax1.plot(years, S_described, color='#fbbf24', linewidth=1.5, linestyle='--',
         label=f'Projected (S_max = {S_max_est:,})')

# Extrapolation
future_years = np.arange(2025, 2100)
S_future = S_max_est * (1 - np.exp(-(future_years - t_start) / tau))
ax1.plot(future_years, S_future, color='#fbbf24', linewidth=1.5, linestyle=':', alpha=0.5)

ax1.axhline(y=48000, color='#22c55e', linewidth=1, linestyle='--', alpha=0.5)
ax1.text(1780, 50000, 'Current: ~48,000', color='#22c55e', fontsize=10)

ax1.fill_between(future_years, S_max_est * 0.8 * (1 - np.exp(-(future_years - t_start) / tau)),
                  S_max_est * 1.4 * (1 - np.exp(-(future_years - t_start) / tau)),
                  alpha=0.1, color='#fbbf24')

ax1.set_xlabel('Year', fontsize=12, color='white')
ax1.set_ylabel('Cumulative Species Described', fontsize=12, color='white')
ax1.set_title('Spider Species Accumulation Curve', fontsize=14, fontweight='bold', color='#c4b5fd')
ax1.legend(fontsize=10, facecolor='#1a1a2e', edgecolor='#7c3aed', labelcolor='white')
ax1.set_facecolor('#0a0a1a')
ax1.tick_params(colors='white')
ax1.set_xlim(1757, 2100)
for spine in ax1.spines.values():
    spine.set_color('#7c3aed')

# ========== Panel 2: Ballooning Dispersal Model ==========
ax2 = axes[0, 1]

# Spider mass range
masses = np.logspace(-7, -4, 100)  # kg (0.1 mg to 100 mg)

# Silk parameters
L_silk = 2.0  # m
d_silk = 1e-6  # m
A_silk = d_silk * L_silk  # projected area

# Air properties
rho_air = 1.2  # kg/m^3
mu_air = 1.8e-5  # Pa.s
C_D = 1.5  # for thin cylinder perpendicular to flow

# Terminal velocity (drag only)
v_term_drag = np.sqrt(2 * masses * 9.81 / (C_D * rho_air * A_silk))

# With electric field assist
E_fields = [0, 100, 500, 1000, 5000]  # V/m
sigma_charge = 1e-5  # C/m^2
q = sigma_charge * np.pi * d_silk * L_silk

for E in E_fields:
    F_electric = q * E
    F_net = masses * 9.81 - F_electric
    F_net = np.maximum(F_net, 0.001 * masses * 9.81)  # prevent negative
    v_term = np.sqrt(2 * F_net / (C_D * rho_air * A_silk))
    # Dispersal range (assume 2 hours flight at 10 m/s wind)
    flight_time = 2 * 3600  # seconds
    wind_speed = 10.0  # m/s
    R_km = wind_speed * flight_time / 1000 * np.minimum(1.0, 0.01 / v_term)  # simplified model
    # Better: range = wind * flight_time, flight_time limited by settling
    altitude = 3000  # m (max)
    t_settle = altitude / v_term
    t_flight = np.minimum(t_settle, 8 * 3600)  # max 8 hours
    R_km_better = wind_speed * t_flight / 1000

label = f'E = {E} V/m' if E > 0 else 'Drag only'
    ls = '-' if E > 0 else '--'
    ax2.plot(masses * 1e6, R_km_better, linewidth=2, linestyle=ls, label=label)

ax2.axhline(y=100, color='#f472b6', linewidth=1, linestyle=':', alpha=0.5)
ax2.text(0.15, 110, 'Darwin observation (~100 km)', color='#f472b6', fontsize=9)

ax2.set_xlabel('Spider Mass (mg)', fontsize=12, color='white')
ax2.set_ylabel('Max Dispersal Range (km)', fontsize=12, color='white')
ax2.set_title('Ballooning Range: Wind + Electric Field', fontsize=14, fontweight='bold', color='#c4b5fd')
ax2.legend(fontsize=9, facecolor='#1a1a2e', edgecolor='#7c3aed', labelcolor='white')
ax2.set_facecolor('#0a0a1a')
ax2.tick_params(colors='white')
ax2.set_xscale('log')
ax2.set_yscale('log')
for spine in ax2.spines.values():
    spine.set_color('#7c3aed')

# ========== Panel 3: Molecular Clock Divergence ==========
ax3 = axes[1, 0]

# Simplified divergence time estimation
# Using substitution rates and branch lengths
np.random.seed(42)

# Key divergence events (MYA)
events = {
    'Araneae origin': 380,
    'Meso/Opisthothelae': 350,
    'Mygalo/Araneomorphae': 300,
    'RTA clade': 180,
    'Orbiculariae': 200,
    'Araneoidea': 160,
    'Araneidae': 130,
    'Theridiidae': 120,
    'Salticidae': 100,
    'Linyphiidae': 110,
}

# Generate posterior distributions (normal approximation)
event_names = list(events.keys())
event_times = list(events.values())

y_positions = np.arange(len(event_names))

for i, (name, time) in enumerate(zip(event_names, event_times)):
    # Posterior uncertainty (wider for older divergences)
    std = time * 0.08 + 5
    samples = np.random.normal(time, std, 1000)

# Violin-like horizontal plot
    bins = np.linspace(time - 3*std, time + 3*std, 50)
    hist, edges = np.histogram(samples, bins=bins, density=True)
    centers = 0.5 * (edges[:-1] + edges[1:])
    hist_scaled = hist / hist.max() * 0.35

ax3.fill_betweenx([i - h for h in hist_scaled], centers,
                       [i + h for h in hist_scaled],
                       alpha=0.4, color='#a78bfa')
    ax3.plot(time, i, 'o', color='#fbbf24', markersize=8, zorder=5)
    ax3.plot([time - 2*std, time + 2*std], [i, i], color='#7c3aed',
             linewidth=1.5, alpha=0.5)

ax3.set_yticks(y_positions)
ax3.set_yticklabels(event_names, fontsize=10, color='white')
ax3.set_xlabel('Divergence Time (MYA)', fontsize=12, color='white')
ax3.set_title('Molecular Clock Estimates', fontsize=14, fontweight='bold', color='#c4b5fd')
ax3.invert_xaxis()
ax3.set_facecolor('#0a0a1a')
ax3.tick_params(colors='white')
for spine in ax3.spines.values():
    spine.set_color('#7c3aed')

# Geological period shading
ax3.axvspan(542, 485, alpha=0.05, color='#22c55e')
ax3.axvspan(419, 359, alpha=0.05, color='#3b82f6')  # Devonian
ax3.axvspan(252, 201, alpha=0.05, color='#ef4444')  # Triassic
ax3.text(390, -0.7, 'Devonian', fontsize=9, color='#3b82f6', alpha=0.7)
ax3.text(226, -0.7, 'Triassic', fontsize=9, color='#ef4444', alpha=0.7)

# ========== Panel 4: Venom Ka/Ks and Diversity ==========
ax4 = axes[1, 1]

# Simulate Ka/Ks data for toxin gene families
np.random.seed(99)
n_families = 8
family_names = ['ICK-1', 'ICK-2', 'LSTX', 'DkTx', 'Magi', 'PhTx', 'HwTx', 'CsTx']
family_colors = ['#ef4444', '#f97316', '#fbbf24', '#22c55e', '#3b82f6', '#8b5cf6', '#f472b6', '#a78bfa']

# Ka/Ks for pharmacophore vs scaffold residues
for i, (name, color) in enumerate(zip(family_names, family_colors)):
    # Scaffold residues: purifying selection
    n_scaffold = np.random.randint(8, 15)
    ka_ks_scaffold = np.random.exponential(0.15, n_scaffold)
    ka_ks_scaffold = np.clip(ka_ks_scaffold, 0.01, 0.5)

# Pharmacophore residues: positive selection
    n_pharma = np.random.randint(5, 10)
    ka_ks_pharma = np.random.lognormal(0.5, 0.6, n_pharma)
    ka_ks_pharma = np.clip(ka_ks_pharma, 0.5, 8)

# Plot
    x_scaffold = np.full(n_scaffold, i - 0.15) + np.random.normal(0, 0.03, n_scaffold)
    x_pharma = np.full(n_pharma, i + 0.15) + np.random.normal(0, 0.03, n_pharma)

ax4.scatter(x_scaffold, ka_ks_scaffold, s=25, color=color, alpha=0.5, marker='o')
    ax4.scatter(x_pharma, ka_ks_pharma, s=35, color=color, alpha=0.8, marker='^')

# Mean lines
    ax4.plot([i - 0.25, i - 0.05], [np.mean(ka_ks_scaffold)] * 2, color=color, linewidth=2)
    ax4.plot([i + 0.05, i + 0.25], [np.mean(ka_ks_pharma)] * 2, color=color, linewidth=2)

ax4.axhline(y=1, color='#fbbf24', linewidth=1.5, linestyle='--', label='Neutral (Ka/Ks = 1)')
ax4.fill_between([-0.5, 7.5], 0, 1, alpha=0.05, color='#3b82f6')
ax4.fill_between([-0.5, 7.5], 1, 10, alpha=0.05, color='#ef4444')
ax4.text(6.5, 0.3, 'Purifying selection', color='#3b82f6', fontsize=10, fontstyle='italic')
ax4.text(6.5, 3.0, 'Positive selection', color='#ef4444', fontsize=10, fontstyle='italic')

# Legend entries
ax4.scatter([], [], s=25, color='white', marker='o', label='Scaffold residues')
ax4.scatter([], [], s=35, color='white', marker='^', label='Pharmacophore residues')

ax4.set_xticks(range(n_families))
ax4.set_xticklabels(family_names, fontsize=10, color='white', rotation=45)
ax4.set_ylabel('Ka/Ks (dN/dS)', fontsize=12, color='white')
ax4.set_title('Venom Toxin Molecular Evolution', fontsize=14, fontweight='bold', color='#c4b5fd')
ax4.legend(fontsize=9, facecolor='#1a1a2e', edgecolor='#7c3aed', labelcolor='white')
ax4.set_facecolor('#0a0a1a')
ax4.tick_params(colors='white')
ax4.set_yscale('log')
ax4.set_ylim(0.01, 10)
for spine in ax4.spines.values():
    spine.set_color('#7c3aed')

fig.patch.set_facecolor('#0a0a1a')
plt.tight_layout()
plt.savefig('output.png', dpi=150, bbox_inches='tight', facecolor='#0a0a1a')
plt.close()
print("Evolution & Diversity Analysis Complete")
print(f"Current described species: ~48,000")
print(f"Estimated total: ~{S_max_est:,}")
print(f"Discovery rate: ~400 species/year")
print(f"Ballooning range for 1mg spider (drag only): ~{wind_speed * (3000 / np.sqrt(2 * 1e-6 * 9.81 / (C_D * rho_air * A_silk))) / 1000:.0f} km")
print(f"Spider origin: ~380 MYA (Devonian)")

Click Run to execute the Python code

Code will be executed with Python 3 on the server

References

World Spider Catalog (2024). World Spider Catalog. Version 25.0. Natural History Museum Bern. Available at: https://wsc.nmbe.ch.
Bond, J.E., Garrison, N.L., Hamilton, C.A., Godwin, R.L., Hedin, M. & Agnarsson, I. (2014). Phylogenomics resolves a spider backbone phylogeny and rejects a prevailing paradigm for orb web evolution. Current Biology, 24(15), 1765–1771.
Fernandez, R., Hormiga, G. & Giribet, G. (2018). Phylogenomics of spiders: reconstructing the deep evolutionary history of a megadiverse order. BMC Evolutionary Biology, 18, 146.
Morley, E.L. & Robert, D. (2018). Electric fields elicit ballooning in spiders. Current Biology, 28(14), 2324–2330.
Felsenstein, J. (1981). Evolutionary trees from DNA sequences: a maximum likelihood approach. Journal of Molecular Evolution, 17(6), 368–376.
King, G.F. & Hardy, M.C. (2013). Spider-venom peptides: structure, pharmacology, and potential for control of insect pests. Annual Review of Entomology, 58, 475–496.
Meehan, C.J., Olson, E.J., Reudink, M.W., Kyser, T.K. & Curry, R.L. (2009). Herbivory in a spider through exploitation of an ant-plant mutualism. Current Biology, 19(19), R892–R893.
Griswold, C.E., Audisio, T. & Ledford, J.M. (2012). An extraordinary new family of spiders from caves in the Pacific Northwest. ZooKeys, 215, 77–102.
Selden, P.A. & Penney, D. (2010). Fossil spiders. Biological Reviews, 85(1), 171–206.
Garb, J.E., González, A. & Gillespie, R.G. (2004). The black widow spider genus Latrodectus: phylogeny, biogeography, and invasion history. Molecular Phylogenetics and Evolution, 31(3), 1127–1142.
Darwin, C. (1839). Journal of Researches into the Natural History and Geology of the Countries Visited during the Voyage of H.M.S. Beagle. Henry Colburn, London.
Gillespie, R.G., Benjamin, S.P., Brewer, M.S., Rivera, M.A.J. & Roderick, G.K. (2018). Repeated diversification of ecomorphs in Hawaiian stick spiders. Current Biology, 28(6), 941–947.

M7. Reproduction & Development Course Overview