Module 2: Trunk Biomechanics

The adult trunk is approximately 2 m long, 140 kg in mass, and contains no skeletal elements whatsoever. The structure is organised as a densely interwoven matrix of three fibre orientations:

• Longitudinal fascicles arranged dorsally and ventrally, running the full length of the organ.
• Radial (transverse) fascicles squeezing the trunk to narrow its cross-section.
• Oblique / helical fascicles mediating torsion and fine control.

Longren et al. (2023) used synchrotron microCT to count ~40 000 distinct muscle fascicles in adult Loxodonta africana trunks — an order of magnitude more than any other known muscular organ. Comparative estimates from octopus arms reach ~20 000 fascicles, and mammalian tongues a few thousand. The trunk is therefore the most kinematically redundant actuator in nature.

Critically, there are no bones or cartilages to act as levers. Force is transmitted through continuous muscle-to-muscle contact mediated by the extracellular matrix. The biomechanical implications of this “zero-lever” arrangement were first systematised by Kier & Smith (1985) and underlie all modern muscular-hydrostat theory.

The defining feature of a muscular hydrostat is incompressibility of the muscle-water matrix. For any slender segment of cross-sectional area A and length L, volume conservation demands:

For a circular cross-section with radius r, A = πr², so a 10% radial contraction (\(r \to 0.9 r\)) elongates the segment by (\(1 - 0.9^2)^{-1} - 1 \approx 23\%\). This is the primary mechanism of trunk extension during reaching and snorkeling.

Bending via differential longitudinal activation

Pure radial contraction extends the trunk without bending. Bending arises from asymmetric activation of dorsal vs. ventral longitudinal fibres. The local radius of curvature κ follows from Euler-Bernoulli kinematics applied to the hydrostat:

Because radial contraction reduces r, the same differential longitudinal strain produces sharper curvature when the trunk is simultaneously extending. This coupling permits exquisite fine control at the trunk tip, where r is smallest.

Telescoping extension

Experimental kinematic studies (Kier & Smith 1985; Dagnino-Subiabre 2019) show that trunk extension propagates distally as a telescoping wave of radial contraction, rather than a uniform lengthening. Segments near the base activate first, propelling the distal tissue forward; the wave reaches the tip in ~ 0.3 s. The simulation below reproduces this telescoping behaviour.

At the distal end the trunk terminates in fleshy lobes often called pseudo-digits or “fingers”. The geometry differs between African and Asian elephants:

• African (Loxodonta): two opposable pseudo-digits — one dorsal, one ventral — enabling precision pinching.
• Asian (Elephas): single dorsal pseudo-digit; ventral grasping relies on wrapping the trunk around the object.

Dagg & Foster (1976) and more recent tactile-sensitivity studies show that the trunk tip can detect surface textures at roughly 1 μm resolution — comparable to human fingertip Pacinian corpuscle density. This extraordinary sensory resolution permits selective foraging: an elephant can pluck a single blade of grass from amongst rooted vegetation without uprooting the clump.

The contrast in dexterity — picking one grass blade versus lifting a 150 kg log — illustrates the enormous dynamic range of the trunk, spanning ~7 orders of magnitude in force and ~4 orders in positional precision.

Wilson et al. (2015, Journal of the Royal Society Interface) used pneumotachography on resting and water-drinking elephants to measure airflow velocities in the trunk. Their key result: during sniff and inhalation manoeuvres, airflow through the trunk can reach 150 m/s — roughly 30× faster than human inhalation. In terms of Mach number this is\(\mathrm{Ma} \approx 0.44\), approaching the transonic regime.

Such high velocities are made possible by the muscular hydrostat acting as a controlled-geometry nozzle. Radial contraction narrows the cross-section from ~80 cm² at rest to ~8 cm² during powerful sniffs, producing 10× acceleration of the airflow (continuity\(A_1 v_1 = A_2 v_2\)). Drinking water is uptaken by similar suction: volumetric flux

The trunk therefore operates in two very distinct flow regimes: a high-velocity air regime for olfaction and snorkeling, and a low-velocity, high-volume liquid regime for drinking. The ability to alternate between them on demand relies on the same muscular-hydrostat mechanics that controls reaching and bending.

West (2002) noted that elephants can remain submerged with only the trunk tip breaking the surface — a posture possible only because their respiratory system lacks a pleural cavity (M1, Section 11). Under immersion the water column exerts transmural pressure\(\Delta P = \rho g h \approx 10\text{ kPa}\) per metre of depth on the lungs. A pleural cavity would collapse under this load; the elephant “pleural adhesion” resists it directly via tensioned connective tissue.

The trunk-as-snorkel constrains aquatic diving depth: at depths\(> 2\) m the transmural pressure on the lungs exceeds alveolar elastic recoil and further inhalation becomes impossible. Observational data place elephant swimming depth typically at < 1.5 m trunk-tip height.

Dagg & Foster (1976) documented that foraging elephants perform a stereotyped trunk-to-mouth cycle with period ~3–4 s. Each cycle comprises: (i) reach, (ii) grasp or pluck, (iii) flex proximally, (iv) deposit into the mouth, (v) extend back to forage position. The cycle is rhythmic with a well-defined motor pattern generator likely residing in the cerebellum (Module 6).

At ~ 4 s per cycle and ~ 10 hours of active foraging per day, an elephant performs ~ 9000 trunk-to-mouth cycles daily. Over a 65-year lifespan that is > 200 million cycles — placing enormous cumulative load on the muscle fibres and highlighting why muscle turnover and regeneration are so robust in this organ (histology shows exceptionally high satellite-cell density).

The trunk is exquisitely innervated by the infraorbital and ethmoidal branches of the trigeminal nerve. The infraorbital nerve of Loxodonta is the largest peripheral nerve of any mammal, containing ~400 000 myelinated fibres — 5× the human optic nerve. This vast afferent bandwidth is matched by an equally remarkable olfactory receptor repertoire:

Niimura et al. (2014) sequenced and annotated the olfactory receptor repertoire across 13 mammalian species. Loxodonta africana has the largest functional OR gene set yet documented, consistent with the elephant’s documented ability to discriminate between very close-related hydrocarbons (important for identifying individual conspecifics by urinary bouquet) and for detecting water sources at distances > 20 km downwind.

The trunk tip also houses a dense population of Meissner-like corpuscles and a network of TRPV1 receptors (five paralogs in Loxodonta, see M0). This combination of mechanosensation, thermosensation, and nociception in a single terminal structure makes the trunk the richest sensory organ per unit surface area in any mammal.

Two limits bracket the trunk’s mechanical envelope. At the precision end, an elephant can pluck a single blade of grass (< 1 g) or a peanut without crushing it. At the power end, wild elephants routinely lift logs of 100–150 kg and captive working elephants have been documented lifting up to 250 kg over short distances (Longren et al. 2023).

This 7-order-of-magnitude force range is made possible by the recruitment logic of the 40 000 fascicles. For delicate grasping only the distalmost few hundred fascicles are activated, producing ~ 1–10 N of tip pinch force. For heavy lifting nearly all fascicles are synergistically recruited.

At 250 kg the trunk operates near its isometric maximum (\(F_0 \approx 2500\) N) — an observation that anchors the Hill model simulation below. Because muscle is intrinsically rate-limited by the shortening velocity, a fully loaded trunk lift is slow: the empirical steady-state velocity at 200 kg payload is only ~0.1 m/s.

Kinematic studies suggest that an extending trunk does not contract uniformly. Instead, activation waves propagate distally from the base along the longitudinal fascicles. Let\(\alpha(s, t)\) denote the activation field parametrised by arclength s and time t. The observed wave has the form\(\alpha(s, t) \approx \alpha_0 \Theta(s - c t)\) with wavespeed\(c \approx 4\,\text{m/s}\), reaching the tip in\(t \approx L/c \approx 0.5\,\text{s}\).

This propagation mode resembles a peristaltic wave in an earthworm, but instead of moving chyme it moves mass outward and lengthens the organ. The telescope-like extension is faster and more energy-efficient than uniform contraction because only a small subset of fascicles is active at any instant, keeping metabolic heat production localised.

A serial kinematic chain with 40 000 independently addressable muscle fascicles has effectively infinite degrees of freedom: the problem of mapping a desired tip pose to a unique activation pattern is wildly underdetermined. The nervous system must therefore solve an inverse kinematics problem under redundancy constraints (cf. Bernstein’s degrees-of-freedom problem in motor control).

As noted in Module 1, ~98&percnt; of elephant brain neurons reside in the cerebellum. Ullmann et al. (2011) argued that this cerebellar dominance reflects trunk motor control: it is the only structure in the elephant brain with sufficient capacity to learn and cache the motor programs associated with the roughly 10⁴ distinct trunk behaviours documented in the field (precision pluck, spray, wrap-lift, dust-bath, vocal resonator, etc.).

Mathematically the problem is one of minimising a cost functional over muscle activation vectors α(s):

The classical Hill (1938) force-velocity equation for concentric muscle contraction is:

Solving for v(F) and plotting against load mg gives the quintessential hyperbolic force-velocity curve. Maximum mechanical power (\(P = F v\)) peaks at approximately (\(F \approx 0.3 F_0\) and\(v \approx 0.3 v_{\max}\)) — a feature common to all striated muscle and directly responsible for the observed “optimal payload” of ~70–100 kg for sustained elephant work.

With F₀ ≈ 2500 N (from the 250 kg lift observation) and\(v_{\max}\) ≈ 0.8 m/s at the tip, peak power output of a single full-trunk contraction reaches ~500 W — a large but not extraordinary fraction of total sustained metabolic power (~1500 W). The simulation below traces this curve across the physiologically relevant load range.

Muscular hydrostats are found across the animal kingdom in very different guises: octopus arms, squid tentacles, mammalian and reptilian tongues, the mammalian penis, and the elephant trunk. Each exploits the same volume- conservation principle, but differs in fibre architecture, innervation density, and sensory specialisation:

• Octopus arm: ~ 20 000 fascicles; all eight arms simultaneously active; distributed nervous system (~2/3 of the ~500 million neurons reside in the arms themselves).
• Mammalian tongue: a few thousand fascicles; subject to palatal constraint (must fit within the oral cavity); shorter reach.
• Giraffe tongue: prehensile and up to 45 cm long; similar hydrostat principles apply at reduced scale.
• Elephant trunk: 40 000 fascicles; ~400 000 trigeminal afferents; by far the most complex terrestrial muscular hydrostat.

Remarkably, all examples show convergent solutions to the bending problem: the Kier-Smith curvature formula\(\kappa = (\varepsilon_d - \varepsilon_v) / (2 r)\) applies equally to octopus arms, giraffe tongues, and elephant trunks. The only substantial difference is scale and fibre count.

Because the trunk is so densely innervated, injuries are catastrophic. Trunk amputation — common in poaching-snare victims — reduces the elephant’s ability to drink, forage, and socialise. Survival is possible but entails dramatic behavioural compensation (kneeling to drink, rooting with the mouth). Welfare protocols for captive and rehabilitated elephants therefore prioritise trunk health above most other interventions.

Stereotypic trunk-swaying in captive settings is a recognised welfare indicator: the behaviour correlates with chronic cortisol elevation and diminished central dopaminergic tone. Longren et al. (2023) noted that microCT-derived fascicle-density estimates remain essentially invariant between wild and captive specimens, but the behavioural deployment of that fascicle repertoire is dramatically impoverished under stress.

The elephant trunk integrates three extreme biomechanical problems: (i) a boneless structure supporting its own weight under gravity, (ii) a dexterous actuator spanning 7 orders of magnitude in force, and (iii) a high-velocity aerodynamic nozzle tuned for chemoreception and snorkeling. Each is a singular evolutionary achievement, and the integration into a single organ is without parallel.

The Kier-Smith formalism underwrites the first two: volume conservation converts muscular contraction into positional control with essentially no mechanical loss. The third — 150 m/s airflow — is made possible by the same radial muscles acting as a variable-geometry nozzle. The Hill force-velocity curve then predicts the quantitative limits of lifting work, and the simulation below reproduces the full 2 500 N isometric ceiling corresponding to the reported 250 kg lift capacity.

Outstanding research questions include (a) whether the 40 000 fascicles are truly independently addressable or clustered into “motor synergies” at lower dimensionality; (b) how the cerebellar motor programs are learned and stored; (c) how the trunk’s dynamic response compares with engineered soft robotic arms (bio-inspired continuum manipulators are now a major field); and (d) how trunk morphology differs between African and Asian lineages at the genomic level (see Module 0). Subsequent modules build on this foundation: thermoregulation through ears and trunk (M3), the role of tusks in feeding (M4), and trunk vocal resonance in infrasound signalling (M5).