![]()
I. The Argument From Nature, and Its Unexamined Premise
There is a defense of traditional lutherie that is rarely stated outright because it is rarely felt to need stating. It runs roughly as follows: the materials of the violin family — spruce, maple, gut, pernambuco, horsehair — are organic, and organic materials carry within them a depth of structure that no analytical process designed and no synthetic process can reproduce. They are the products of evolutionary time. They possess what we might call a temporal thickness. The craftsman who selects them is not merely choosing a substance but inheriting a history.
I have made a version of this argument myself, and I still believe it. But it contains a premise that almost nobody examines, and the premise is false.
The premise is that the argument applies uniformly. That being organic is itself the credential. That a material's participation in biological time confers upon it a general fitness for whatever use a craftsman might discover for it.
This is not how evolution works, and it is not how these materials behave.
What follows is an attempt to state the discriminating criterion that the argument from nature actually requires — and then to apply it honestly, including where it tells against a material I would rather it did not.
II. The Criterion
Craft tradition is a form of memory. It stores solutions without storing reasons. A tradition that has selected spruce for soundboards for four hundred years does not need to know why spruce works; the knowing is distributed across generations of trial, and it is encoded in practice rather than in propositions. This is what makes sedimented intelligence formidable. It has run more experiments than any laboratory, over longer time, with the ear as its instrument.
But sedimented intelligence has a specific condition of validity, and identifying it is the whole business of this essay.
Craft tradition inherits a genuine optimization only when the material's evolved biological function is mechanically adjacent to the technical function the craftsman demands of it.
Where that adjacency holds, the tradition arrives at a material that natural selection has already refined for a structurally analogous problem. The craftsman's intuition then operates within an excellent solution, tuning at the margins of something already close to optimal. The tradition looks wise because it inherited wisdom it did not have to generate.
Where adjacency fails, the tradition inherits nothing but availability. And then centuries of accumulated practice amount not to accumulated wisdom but to accumulated accommodation — the slow, ingenious, and entirely undirected work of building a craft around a material's defects rather than around its virtues.
The criterion is not a philosophical nicety. It sorts the violin's materials, and it sorts them unevenly.
![]()
![]()
![]()
![]()
III. Ranking Adjacency
Spruce, in the soundboard — strong adjacency.
A conifer trunk is under lifelong selection for maximum stiffness at minimum mass, for efficient axial load transfer, and for resistance to buckling under wind loading. It must hold a heavy crown aloft for a century in a medium that pushes it sideways.
A soundboard requires high specific stiffness along the grain and efficient transport of vibrational energy across the plate.
These are not the same function. But the underlying material demand — pronounced anisotropic stiffness at the lowest achievable density — is very nearly isomorphic. Spruce arrives at the luthier's bench pre-optimized for a structurally analogous problem. The remarkable ratio of longitudinal to transverse stiffness, the low internal damping, the regularity of the growth-ring architecture: none of this was selected for music, but all of it was selected for something mechanically continuous with music's demands.
Maple, in the back, ribs, and neck — strong adjacency.
Selected for compressive strength, dimensional stability across seasons, and resistance to splitting. Asked, in the instrument, for very nearly the same, with the addition of reflective stiffness. The adjacency is close enough that maple's failure modes in an instrument are the same as its failure modes in a tree.
Gut, in the strings — moderate adjacency.
Intestinal tissue is under selection for tensile strength, hierarchical collagen alignment, and integrity under repeated deformation. A string demands tensile strength and stable elasticity under sustained load.
Adjacent, though less perfectly. The frequency regime — hundreds of cycles per second, sustained for hours — is entirely foreign to the tissue's evolutionary history, and gut's hygroscopic instability is the price. But the fundamental demand, a fiber that holds tension and returns energy, is one the tissue was built to meet.
Pernambuco, in the stick — moderate to strong adjacency.
Paubrasilia echinata is selected for the dense, stiff, low-damping heartwood characteristic of slow-growth tropical hardwoods, where structural economy and resistance to decay are both under pressure. The bow demands high specific stiffness and low loss tangent.
Here the adjacency is real but partly accidental, and the accident is instructive. Research on pernambuco's exceptionally low internal damping has attributed a substantial portion of it not to the cellulose-lignin architecture but to the wood's extractives — brazilin among them — such that removing the extractives measurably raises damping. The property that makes pernambuco the finest bow wood is thus a byproduct of a chemistry evolved for entirely other purposes, defensive or metabolic. Selection produced it; nothing selected for it in the acoustic sense. This is adjacency at its most contingent, and it should temper any account that treats the fit between organic material and musical function as anything other than fortunate.
Horsehair, in the ribbon — adjacency failure.
A tail hair is under selection for insect deflection, thermal regulation, social and sexual display, and tolerance of abrasion against the animal's own body and its environment.
It is asked, in the bow, to sustain constant axial tension for months. To maintain dimensional stability under that sustained load. To undergo cyclic mechanical loading at hundreds of hertz. To execute a controlled frictional release against a hard surface, hundreds of times per second, mediated by an applied resin.
Not one of these demands appears anywhere in its evolutionary brief. The functional distance between deflect flies and excite a string via stick-slip friction at four hundred hertz under sustained ribbon tension is not a matter of degree. It is categorical.
IV. The Failure Modes Are the Diagnosis
If the criterion is correct, it should predict not merely that horsehair is suboptimal but how it fails. And it does, with a precision that ought to be uncomfortable.
Horsehair breaks under tension — routinely, unpredictably, mid-performance.
It stretches under sustained load, and stretches further with ambient humidity. Keratin's moisture regain runs on the order of 15 to 30 percent by mass depending on conditions, which means the working tension of the ribbon is a function of the weather. The player compensates by adjusting the screw, continuously, for the life of the hair.
It varies substantially strand to strand in diameter and surface condition, such that a significant fraction of any hank must be discarded before use.
It requires chemical bleaching to reach acceptable uniformity of appearance, a process that degrades the fiber further.
It oxidizes, loses its capacity to hold rosin, and must be wholly replaced twice a year at meaningful cost.
These are not the signatures of a mature technology approaching its optimum. They are precisely what the adjacency criterion predicts when a material is conscripted into a mechanical role for which nothing ever selected it. Spruce does not fail this way in a soundboard, because spruce in a soundboard is doing something structurally continuous with what it was built to do. When spruce fails, it fails from abuse or from the luthier's error. When horsehair fails, it fails from use.
And the tradition's response to five centuries of this was not to revisit the choice. It was to optimize around it.
Rosin chemistry developed to compensate for the fiber's variable grip. Camber and stick geometry evolved partly to manage the ribbon's compliance. Conventions of hair tension, of hair count, of hank width — all of these are accommodations. And eventually, most tellingly, bowing technique itself absorbed the material's limitations: the player learns to accommodate a ribbon that thins over months, that changes tension with the room, that will occasionally shed a strand mid-phrase. Technique became the final layer of compensation.
This is what accumulated accommodation looks like from the inside. It is indistinguishable, to its practitioners, from accumulated wisdom. It produces genuine expertise, genuine subtlety, genuine mastery. What it does not produce is any occasion to ask whether the material at the center of it was ever the right one.
V. Why No One Noticed
It would be too easy, and false, to conclude that the great makers were simply 'inattentive'. They were not. The correct explanation is structural, and it has three parts.
The bow was not a variable until Tourte made it one. Before the standardization achieved by François Xavier Tourte in the closing decades of the eighteenth century, stick geometry, length, camber, head form, and hair width varied so widely between makers and traditions that the hair's contribution would have been swamped by everything else. One cannot perceive a second-order variable while the first-order variables remain in flux. It is at least arguable that Tourte's standardization is what created the conditions under which the ribbon could become perceptible as a parameter at all.
The playing context did not expose it. Baroque performance was largely ensemble, in resonant rooms, on gut strings whose own damping and slower speech dominate the articulation profile. A compliant, slow-speaking string in a reverberant space within a continuo texture masks differences in excitation. This is not a claim about the acuity of seventeenth-century ears. It is a claim about signal.
And the division of labor placed the question outside the violin maker's competence. Cremona built instruments; bows came from Mirecourt, Nuremberg, London. That Stradivari made no bows is a fact about guild structure, not about his evaluative judgment — he made no strings either, and no one infers from this that he thought strings unimportant.
The conditions for noticing arrived late: standardized bows, projecting strings, dry halls, exposed solo playing, and finally recorded scrutiny. The historical silence about bow hair is not evidence that bow hair does not matter. It is evidence that the apparatus for hearing it is recent.
VI. The Field Has Revisited a Material Before
There is a counterexample to the whole picture of traditional inertia, and it is worth confronting because it is the strongest evidence that the field can reopen a settled material question.
The strings did change. Gut gave way to steel and then, for most players and most instruments, to synthetic cores.
What makes this instructive is that the winning material would have failed every a priori materials screen. Consider what a naive analysis would have concluded about nylon 6 — Perlon — around 1970. Its Young's modulus is on the order of two to three gigapascals against steel's two hundred: two orders of magnitude inferior. Its internal damping is roughly a hundredfold higher. Its density is a fraction of steel's. On every axis a materials engineer would name, the polymer is not marginally worse but catastrophically worse.
It won anyway. And it won in the more demanding case, since the string is the oscillator — its internal losses attenuate its own partials directly.
Two lessons follow. The first is that datasheet screening mispredicts, because the properties that matter are the ones measured under operating conditions, in the relevant architecture, at the relevant frequency and strain. The second, and more useful: the synthetic core succeeded not through chemistry alone but through architecture. A solid nylon monofilament at the required diameter would have been far too stiff in bending, producing inharmonicity — partials running sharp of the harmonic series, heard as a metallic and unfocused tone. Stranding thousands of fine filaments decouples axial stiffness from bending stiffness: the bundle carries tension like a solid and bends like a rope.
The material deficit was engineered around at the level of structure. That is the transferable insight, and it is the one most likely to matter for whatever replaces horsehair.
It is also worth noting, as a small lesson in the sociology of materials, that the industry quietly abandoned the name. Perlon became "synthetic core" became a series of proprietary names disclosing nothing, not because the polymer changed but because the culture's regard for polymers did.
VII. The Reversal
There is one axis on which horsehair may be genuinely adjacent after all, and it happens to be the axis that matters quite a bit.
Keratin's cuticle architecture and its surface chemistry are under real selection — for handling contact, for resisting abrasion, for interacting with oils and particulates. Whatever else a tail hair is not built for, it is built to have a surface that does work.
And if rosin adhesion is governed principally by surface polarity rather than by bulk mechanical properties — as I believe the evidence indicates, and as my own bench work has repeatedly suggested — then horsehair may be accidentally excellent at the single property on which the entire function depends, while remaining poor at everything else.
This would explain the historical record with remarkable economy. A century of synthetic substitutes has failed. But examine how they failed. Not on stiffness. Not, primarily, on damping. Almost invariably on rosin retention — which is to say, on the surface. Each attempt replaced the bundle entire, discarding along with the fiber's manifest defects the one component that evolution had inadvertently optimized. The failures were not evidence that keratin is irreplaceable. They were evidence that no one had identified which part of keratin was doing the work.
VIII. Decomposition
The correct conclusion, then, is neither that horsehair is sacred nor that it is worthless. It is that horsehair is a bundle — a set of properties that happen to travel together in a tail hair because they travel together in a horse. Some are adjacent to the technical function. Most are not. The surface may be excellent. The bulk is demonstrably poor.
Tradition had no way to separate them. A material harvested whole must be accepted whole, and every subsequent refinement — of rosin, of camber, of technique — is a refinement of the accommodation rather than of the material.
Manufacture permits what neither nature nor craft ever could. It permits the bundle to be taken apart: to retain the surface chemistry where evolution did its accidental good work, and to redesign the bulk where it did none. To specify a fiber that holds rosin as keratin does and holds tension as keratin cannot.
This is not a rejection of the argument from nature. It is the argument from nature taken seriously enough to ask the question it has always declined to ask — not is this material organic, but which part of the organism was ever doing the work.
The tradition could not ask it. We can.
Nikolai Rogich is a luthier in Cleveland, Ohio, and the founder of Pegasus Bows.