Pearls on a String: A Shape the Textbooks Drew Wrong, and the Membrane That Was Doing the Work
Unmyelinated axons are not smooth tubes. They rest in a beaded form set by membrane mechanics, and the beading changes how fast the signal travels — a measured case in which the form is a standing negotiation, not a container, and in which the old drawing was an artefact of the instrument.
The drawing that stood for a century
The unmyelinated axon has been drawn as a smooth cylinder for about as long as there have been textbooks to draw it in, with a standing corollary: when an axon shows a row of swellings along its length, something has gone wrong — injury, disease, degeneration. The beads were pathology. The tube was the thing itself.
The measurement reported in Nature Neuroscience in December 2024 says the resting shape is already beaded. Using high-pressure freezing and electron microscopy, which fixes the tissue by thermal speed rather than by chemistry, the groups found that unmyelinated axons at rest are non-uniform: a repeating alternation of wider bulges and narrower necks along the axon, a string of pearls. Chemical fixation — the century's workhorse — perturbs membrane tension and osmotic balance while it works, and it had been smoothing out the very feature under examination. The smooth cylinder was not an observation. It was what the preparation produced.
This is the volume's lens argument in its most literal form. A load-bearing feature of the object had been averaged away by the instrument, and the averaged result was read as a property of nature. It belongs beside the altermagnet whose bulk signal cancels to zero (Chapter 81): in both cases the missing quantity was never missing, and the first question when something load-bearing appears absent is what the instrument fails to resolve.
What sets the shape
The paper's title states the mechanism in the register this part of the volume has been working in: membrane mechanics dictate the morphology. The pearling is not built by a scaffold and then maintained; it falls out of the balance between membrane tension, bending resistance, and the adhesion and spontaneous curvature imposed on the bilayer by its surroundings. Change the balance and the shape moves. The investigators changed it — osmotic manipulation, tension manipulation — and the pearl geometry changed accordingly, in the direction the theory predicted.
The shape belongs to a family of classical instabilities in which a cylinder of fluid held by surface tension breaks into a periodic row of bulges because the beaded configuration costs less energy than the tube — the same physics that turns a thin stream of water from a tap into a chain of drops. A living membrane is not a simple fluid cylinder, and the paper's model carries bending stiffness, spontaneous curvature and confinement that the textbook instability does not. The kinship is a thing in common, not an identity, and it is borrowed for the shape of the argument rather than for its local mechanism.
Read through this part's governing principle, the pearl is not an object on a string. It is what a standing negotiation between tension, curvature and surroundings looks like while the negotiation holds — which is the drum head of Chapter 80 read at a different scale and with a different measured outcome. There the field concentrates where the charged sheet folds. Here the fold is the resting state, periodic along the axon, and it is set by the same class of terms.
And the function came with it
A morphological correction alone would be a good paper. What makes this one load-bearing for the volume is the second half of the title: morphology and function. Conduction along an unmyelinated axon depends on its geometry — axial resistance scales with cross-section, membrane capacitance with surface area — so a periodic alternation of bulges and necks is not cosmetic. The investigators report that manipulating the pearling changed conduction velocity. The bead is part of the signalling apparatus.
That closes the loop this part of the volume has been drawing. The shape is set by a relation among mechanical terms; the function is set by the shape; therefore the function is carried by the relation and not by any part of the axon considered alone. Remove the negotiation and you do not get a cleaner version of the same signal — you get a different conduction regime. It is the same sentence Chapter 81 was allowed to carry, arriving this time from an electrode rather than from a spectrometer: the information was never inside the thing alone, but in the relation that let the thing hold its form.
The reach of the finding should be stated narrowly. It is reported in unmyelinated axons; myelinated fibres are a different geometry with a different conduction mechanism, and nothing here transfers to them without its own measurement. How much of neural computation depends on pearling, whether pearl geometry is modulated dynamically as a signalling variable rather than merely set by resting conditions, and what other preparations have been quietly smoothing — all open. The result licenses the correction and the mechanism, not a theory of the brain.
The seam this leaves open
Two seams, and they run in opposite directions. The first is an audit rather than a leap: if a century of chemical fixation flattened a periodic resting shape in axons, the same preparation was used on nearly everything else with a membrane. Which other resting geometries are artefacts of the fixative? That is answerable with existing instruments and existing tissue, and it does not require a single speculative sentence.
The second is the leap, and it is bounded. This part of the volume has a charged drum head (Chapter 80), a tuned skin whose admissible notes move with its tension (Chapter 76), and now a measured case in which membrane mechanics set a periodic geometry that changes what a cell can do. The jump asks whether periodicity in membrane form is more widely functional than the beading of one axon type — and it leaps from those measurements, not from mystery. The carrying test is the volume's standing one: a system whose function tracks a mechanically set periodic geometry, changes when the mechanical terms are changed, and survives thermal and sham controls. The microtubule and mitochondrial-membrane test beds named earlier in this part are where such a test would run.
The checkpoint question stays where it was. Each newly resolved shape is being handed to us by a better instrument, so: what does this resolved period let the system do that the smoothed description could not account for — and what is the smallest perturbation that still lets you read the answer without becoming the artefact you went looking for? The century of smooth cylinders is the honest warning that the second half of that question is not rhetorical.
Equations borrowed
- Griswold, J. M., Bonilla-Quintana, M., Pepper, R., Lee, C. T. et al. (2025). Membrane mechanics dictate axonal pearls-on-a-string morphology and function. Nature Neuroscience, 28, 49–61. DOI 10.1038/s41593-024-01813-1. Published online 2 December 2024; open access. The measured result and its accompanying model: the primary anchor of the chapter.
- Helfrich bending energy of a lipid bilayer (Helfrich, 1973), and the classical Rayleigh–Plateau instability of a surface-tension-held cylinder, borrowed for the shape of the beading argument and not for its local mechanism — things in common, not homology.
- Cable-theory scaling of axial resistance and membrane capacitance with axon geometry, textbook, as the reason a periodic change in cross-section is expected to change conduction velocity.
- The zero-and-infinity suspicion and the bounded-jump method, the volume's own (Chapter 81), borrowed back as the instruments of the chapter; the lens argument of Chapter 80 and the tuned skin of Chapter 76 as the frames this result is read alongside.
Validity band
The measured claims hold as reported: in unmyelinated axons, prepared by high-pressure freezing rather than chemical fixation, the resting morphology is periodically beaded; the geometry is set by membrane-mechanical terms and moves when those terms are manipulated; conduction velocity changes with the geometry. Nothing here is established for myelinated fibres, for whole-circuit computation, or for pearling as a dynamically modulated signalling variable — those are open. The Rayleigh–Plateau kinship is an analogy of form, not a claim that a living membrane is a simple fluid cylinder. The wider reading — that functional form in membranes is generally a standing negotiation rather than a container — is interpretation, labelled as such, propped on this result together with Chapters 76, 80 and 81, and it licenses no biological claim beyond them.
Falsifier
The chapter's primary anchor fails if independent groups, using preparations that do not perturb membrane tension or osmotic balance, find unmyelinated axons at rest to be uniform cylinders — i.e. if the beading proves to be an artefact of high-pressure freezing rather than chemical fixation being an artefact of the fixative. The mechanism fails if pearl geometry is unresponsive to controlled changes in membrane tension, osmotic pressure and spontaneous curvature, or if it is shown to be imposed by a cytoskeletal scaffold with the mechanical terms playing no causal role. The functional claim fails if conduction velocity is unchanged when pearl geometry is manipulated with all other variables held. The chapter's interpretive reading fails if a membrane-set functional geometry is exhibited whose function is fully recovered after the mechanical relation setting it is removed. The bounded jump fails if no further system is found in which function tracks a mechanically set periodic geometry and survives thermal and sham controls.
Where this chapter is weakest
The chapter rests on a single paper, and a single paper — however well controlled, however open its data — is not yet a replicated correction to a century of drawings. The honest position is that the smooth cylinder has been credibly overturned in one preparation by one collaboration, and that the correction is waiting on independent confirmation this chapter cannot supply. The Rayleigh–Plateau borrowing is doing more rhetorical work than mechanistic work, and the phrase 'string of pearls' is attractive enough to carry an argument it has not earned. The functional half of the claim is the part a reader is most likely to overextend: a change in conduction velocity in an unmyelinated axon is not a theory of computation, and this chapter's framing of the result as a fourth resonance risks making the four look like a programme when they are four separate measurements that happen to agree. The lens argument is the chapter's strongest material, and it is also the easiest to turn into a general suspicion of all instruments, which would be the nihilist mirror of credulity rather than the discipline this volume is trying to keep.
The volume-wide audit of these weak points is collected in Where This Volume Is Weak.