Whip Crack: The Tip That Outruns Sound

Anatomy of a Crack: Tens of Milliseconds from Wrist to Bang

A bullwhip is not a rope. It is a graded stack of parts, each one lighter than the last. A typical 6-foot (1.8 m) whip weighs 0.4-1 kg and consists of a rigid handle; a plaited leather thong around 20 mm across at the heel that tapers continuously along its length; a short unplaited fall; and finally the cracker or popper, a 10-20 cm strand of nylon or horsehair around a millimetre thick or less, with a mass of only a few tenths of a gram. Nothing in that assembly stores elastic energy the way a bow limb or a mantis shrimp latch does. The whip is a pure kinematic amplifier: energy enters slowly, spread over a large mass, and leaves quickly, concentrated in a minute one.

The sequence is short. The hand accelerates the butt of the whip to roughly 10 m/s and then decisively stops, injecting on the order of 10-30 J. That abrupt reversal folds the thong into a transverse loop - a hairpin, not a smooth sine wave - which rolls outward toward the tip, crossing the whole whip in a few tens of milliseconds. For most of that journey the loop travels at an unremarkable 20-60 m/s.

Everything interesting is crammed into the last millisecond, in the final few centimetres. There the loop collapses onto the cracker, the tip is flung forward at several hundred metres per second, and the local flow around it goes supersonic. The pressure disturbances the tip has been radiating pile up into a single steep-fronted pulse. That pulse - an N-wave, a sharp compression followed by a rarefaction and a second sharp recovery, lasting well under a millisecond - is the crack.

The Rolling Loop: Why the Crest Moves at Twice the Loop Speed

The most useful mental picture is a wheel. A travelling loop on a flexible line is a fold that translates along the line while string feeds continuously into one leg and out of the other. In the frame moving with the fold, the material streams through it at the local wave speed. In the laboratory frame the segment on one side of the loop is instantaneously at rest, exactly like the contact patch of a rolling wheel, and the crest of the loop therefore moves at twice the loop's translation speed - just as the top of a rolling wheel moves at twice the speed of its hub.

That factor of two is free, and it is not the only gain. Alain Goriely and Tyler McMillen showed in 2002 (Physical Review Letters 88, 244301, Shape of a Cracking Whip) that in an idealised inextensible, perfectly flexible string the loop does not simply translate at constant speed: it shrinks as it travels, and the tip speed grows as a power law that formally diverges in finite time. Their model treats the whip as a tapered, inextensible line, and two things feed the divergence together: the loop tightens as it advances, and it runs into ever lighter material.

Three effects therefore stack in a real whip:

• Loop kinematics. The crest runs at twice the loop translation speed.
• Loop collapse. The fold tightens as it advances, concentrating the moving material into an ever-shorter arc.
• Taper. The loop runs into progressively lighter material, so the same energy and momentum drive an ever-smaller mass.

Taper as an Impedance Transformer: The Governing Numbers

The relation everyone starts from is the transverse wave speed on a string, c = √(T/μ), where T is tension and μ is mass per unit length. Toward the free end of a whip the tension is modest in absolute terms - of order tens of newtons, generated dynamically by the loop's own curvature rather than imposed from outside - but μ collapses faster than T does, so c climbs steeply along the whip.

Put numbers on the taper. Plaited leather at a density of about 900 kg/m^3 and 20 mm diameter gives μ of roughly 0.3 kg/m at the heel. The last few centimetres of the cracker - a filament a few tenths of a millimetre across at about 1140 kg/m^3 - give μ near 10^-4 kg/m. That is a fall of some three orders of magnitude, around a factor of 3000.

Two complementary arguments turn that into a speed. The first is a simple energy budget: if the kinetic energy of the moving material is roughly conserved as the moving mass shrinks, then v scales as m^-1/2, and a 3000-fold drop in μ buys a factor of about 55. The second is the adiabatic-taper result familiar from horns and transmission lines. The mechanical impedance of a string is Z = √(Tμ) = μc, the transmitted power goes as Z times the square of the transverse velocity, and if the taper is gradual on the scale of the disturbance, conserving that power gives v ∝ Z^-1/2 = (Tμ)^-1/4. Both routes say the same thing: shedding impedance buys speed.

Neither idealisation is more than an order-of-magnitude guide, and a real whip loses heavily along the way: drag rises as v^2 and worsens sharply transonically, the plait dissipates energy internally, and much of the injected energy stays behind in the trailing thong. The observed end-to-end amplification is roughly 30-70x, turning a 10 m/s hand into a 300-700 m/s tip. At 20 °C the speed of sound in air is 343 m/s, so even the low end of that range crosses Mach 1 and the high end reaches Mach 2. Over the final millisecond the tip changes speed by several hundred metres per second, an acceleration of order 10^5 m/s^2, roughly 10,000 g.

Seeing the Shock: A Century of Measurement

The sonic-boom explanation is old. Otto Lummer - better known for the blackbody radiation measurements with Ernst Pringsheim that helped pin down Planck's law - argued around 1905 that the crack must be an acoustic shock from a supersonic tip. In 1927 Z. Carrière published shadowgraphs that actually showed the shock front in the air beside a cracking whip, using the same Toepler-family optics that had made ballistic shock waves visible decades earlier.

Direct kinematic proof came in 1958, when B. Bernstein, D. A. Hall and H. M. Trent published On the dynamics of a bull whip in the Journal of the Acoustical Society of America (vol. 30, p. 1112). Their high-speed films showed the tip exceeding the sound speed and tied the bang to that moment.

The decisive modern study is Peter Krehl, Stephan Engemann and Dieter Schwenkel at the Ernst-Mach-Institut in Freiburg, published in Shock Waves (vol. 8, pp. 1-9, 1998) as The puzzle of whip cracking - uncovered by a correlation of whip-tip kinematics with shock wave emission. By combining high-speed shadowgraphy with frame-by-frame tip tracking they recorded peak tip speeds near Mach 2, and delivered the key correction to a century of textbook accounts: the bang is emitted close to the tip's maximum speed, not at the instant it first crosses Mach 1. The tip goes supersonic some way before the audible crack, and the shock only coalesces once the tip is deep into the supersonic regime.

Reproducing this means respecting the timescales: tip kinematics need 10^4-10^5 frames per second, and the pressure signature needs a quarter- or eighth-inch pressure-field condenser microphone able to resolve a shock front whose rise time is microseconds. An A-weighted, slow-response meter under-reports such a peak by tens of decibels, which is why casual measurements of whip loudness scatter so badly.

Look-alikes, and the Explanations That Are Wrong

Several tidy-sounding accounts of the crack are simply false, and one true-sounding one is subtly off.

• The tip slaps the air. Slapping air is not a mechanism for a 130 dB report. A 0.2 g cracker moving subsonically radiates a soft swish; the crack appears only once the tip is supersonic, and it disappears if you shorten or remove the cracker so the tip never gets there.
• The whip strikes itself. High-speed film shows whips cracking cleanly in free air with no self-contact, and cracks happen on strokes where the geometry makes contact impossible.
• The cracker breaks. Crackers do wear out and eventually shred, but that is fatigue from transonic drag and 10,000 g accelerations, not the sound source. A fresh cracker cracks just as loudly.
• The bang happens at Mach 1. This is the Krehl correction. The crossing of Mach 1 is not the acoustically special instant; peak tip speed is.

The whip is also routinely conflated with two other bangs that share the same waveform. A supersonic rifle bullet produces an N-wave too, but its source is a body in near-steady flight dragging an attached conical Mach cone; an observer beside the trajectory hears the ballistic crack first and the muzzle blast afterward. An aircraft sonic boom - the Bell X-1 at Mach 1.06 on 14 October 1947, or Concorde at Mach 2 - is that same cone sustained for minutes and swept across the ground as a boom carpet, with bow and tail shocks separated enough to be heard as a double boom over tens to hundreds of milliseconds. The whip's source, by contrast, exists for about a millisecond: it accelerates hard, radiates one coalesced pulse, and is gone. Same waveform, radically different source history.

Where the Model Breaks, and What Is Still Open

The power-law divergence in the idealised inextensible-string model is a signal that the model has run out of physics, not that the tip speed is unbounded. Several effects cut it off in the real world:

• Bending stiffness. A plait cannot fold to zero radius, which puts a floor under the loop size and therefore a ceiling on the crest speed.
• Aerodynamic drag. Drag grows as v^2 and then jumps again as wave drag once the tip goes transonic, so the last increments of speed are the most expensive.
• Material strength. The tension needed to keep accelerating a tip at 10^4 g is small in newtons but is applied to a sub-millimetre filament; the cracker is genuinely near its working limit.
• Three-dimensionality. Real cracks are not planar. Circus and Australian competition strokes deliberately use out-of-plane loops, and the standard planar models do not capture them.

The energy audit is instructive. Of the 10-30 J the arm supplies, the radiated pulse carries only a small fraction, plausibly a percent or less; the rest goes into the whip's residual motion and drag. The whip is a superb speed amplifier and a mediocre loudspeaker, which is exactly why the bang is so brief and so sharp.

Open questions remain. It is unsettled whether the shock is emitted from the very tip or from a short terminal segment of the fall, and how much acoustic energy comes from the tip's violent deceleration rather than its supersonic advance. Optimal taper profiles have never been systematically solved. And one spectacular application is unresolved: Nathan Myhrvold and Philip Currie argued in Paleobiology (vol. 23, pp. 393-409, 1997) that the long, strongly tapered tails of diplodocid sauropods such as Apatosaurus had the right mass distribution to crack supersonically, perhaps as signalling. The biomechanics of vertebral joints, soft tissue and tail-tip strength are still argued over, but the underlying amplifier is exactly the one in a bullwhip.

The whip crack next to other supersonic and near-supersonic bangs

Source	Peak speed of the emitter	Acoustic signature	Duration and level
Bullwhip cracker	~350-700 m/s (Mach 1-2)	Single N-wave from a point source that accelerates and decelerates within a millisecond	~0.1-0.5 ms; ~120-150 dB peak at 1-2 m
Rifle bullet (5.56 mm)	~900 m/s (Mach ~2.6)	Attached conical Mach cone dragged along by a body in near-steady flight; crack then muzzle thump	~0.2 ms N-wave; ~140-150 dB a few metres off the line of fire
Bell X-1 (1947) / Concorde	Mach 1.06 / Mach 2.0	Sustained bow and tail shocks sweeping a carpet across the ground, heard as a double boom	~50-300 ms; ~50-100 Pa overpressure at the surface
Lightning channel	initially far above Mach 1	Cylindrical blast from an explosively heated channel, refracted and scattered into thunder	0.1-1 s of rumble; ~120 dB at 100 m
Wet-towel snap	typically ~10^2 m/s, subsonic	Aerodynamic and impact noise, no shock front	few ms; ~90-110 dB