One of the things I detested about being a little kid was that every time I thought of something really cool to do, I was invariably thwarted by my little brother shouting, “Mom! Kelvin's mixing rocket fuel in the bathtub again!” or “Mom! Kelvin's making a submarine out of the old refrigerator!”. Well, middle age has its drawbacks, but at least you can undertake a project like this without fear of getting nipped in the bud at the cry, “Mom! Kelvin's down in the basement bending spacetime!”. It's important to recall the distinction between “grownup” and “grown up”. Let's us grownups head for the basement to bend some serious spacetime.
Apart from rare and generally regrettable moments of free-fall, we spend our entire lives under the influence of the Earth's gravity, yet rarely, if ever, do we experience the universal nature of gravitation. It's a tremendous philosophical leap from “stuff falls” to “everything in the universe attracts everything else”. That leap, made by Isaac Newton in the 17th century, not only allowed understanding the motion of the Moon and the planets, but inoculated in Western culture the idea that the universe as a whole was governed by laws humans could discover. This realisation fueled the Enlightenment and the subsequent development of science and technology.
This page presents a “basement science” experiment which reveals the universality of gravitation by demonstrating the gravitational attraction between palpable objects on the human scale. The experiment deliberately uses only the crudest and most commonplace materials, permitting anybody who's so inclined to perform it. Einstein's 1915 theory of General Relativity explains gravitation as spacetime curvature created by matter and energy. So, by demonstrating how every object in the universe attracts everything else, we're bending spacetime in the basement.
But, if gravitation is ubiquitous, why was it not discovered millennia before Newton's 1687 Philosophiæ naturalis principia mathematica? The reason lies in the extraordinary weakness of the gravitational force.
Now you might say, “What do you mean, weak? I fell down a flight of stairs a couple of years ago, and gravity sure didn't feel weak to me!“. And yet, of the four forces of nature known to physics, gravitation is the weakest, by the mindboggling factor of 4.17×1042 (4 followed by 42 zeroes) times weaker than the electromagnetic force.
The stark difference in the strength of the electromagnetic and gravitational forces is evident in the picture to the left. The bright square in the jaws of the pliers is a 4 mm cubical magnet. It is lifting a spherical steel pétanque (a lawn bowling game popular in southern France and Switzerland) ball which weighs 550 grams. Consider this picture in the following way: we're pitting our valiant little magnet, with a volume of 0.064 cubic centimetres, weighing less than one gram, pulling up with the electromagnetic force, against the entire Earth, pulling down with gravity. And the winner is…the magnet. A one gram magnet (I'm being generous: I don't have a scale on which its weight reads other than zero) out-pulls the Earth, which weighs 5.9736×1027 grams and has a volume of more than 1027 cubic centimetres.
(The apparent discrepancy between the ratio of masses of the Earth and the magnet and the 4.17×1042 strength ratio of electromagnetism and gravitation is due to the fact that only an infinitesimal fraction of the mass of the magnet contributes to the [electro]magnetic attraction on the ball, while every gram of the Earth's mass exerts gravitational force. To obtain the correct ratio of force strengths, one must compare the gravitational attraction between two electrons at a given distance with the electromagnetic repulsion resulting from their charge. This calculation arrives at the correct strength ratio for the two forces.)
If gravity were not so weak compared to the electromagnetic force, you
wouldn't be reading this page; it's only because the electromagnetic
force that bonds the atoms in your body together so easily defeats the
Earth's gravity that you, along with all other solid objects,
don't slump into a puddle and eventually merge into a perfectly
spherical (actually, slightly ellipsoidal thanks to rotation) planet.
where G, the gravitational constant, is:
Even though the force of gravity between objects of modest mass is palpable compared to the weight of objects one can see, detecting such a tiny force seems a daunting, if not hopeless, endeavour for the basement tinkerer. Certainly, painstakingly designed and constructed laboratory apparatus has allowed measuring the gravitational constant to great precision and verifying the equivalence of gravitational and inertial mass, and precision gravitometers are routinely used in oil and gas exploration and mineral prospecting, but we're trying to see if we can experience the universal attraction of gravity without any high tech, high budget gear.
Measuring tiny gravitational forces would be easy if we were in deep space, far from any massive bodies. The only forces on objects in our space laboratory, then, would be those entirely under our control. As long as we made sure none of the objects we were experimenting with were magnetic or electrically charged (easily arranged, assuming they are conductive, simply by bringing them into contact so all excess charges equilibrate), the only force remaining between objects would be gravity, so however weak it be, we need only be sufficiently patient to observe its effects. (The other two forces, the strong and weak nuclear interactions, are limited in range to distances on the order of the size of an atomic nucleus and can be neglected on the human scale.)
What we'd like to do, then, is cancel the Earth's gravity so that the much smaller gravitational forces between objects that fit in the basement become evident. Fortunately, we don't need a 25th century WarpMan to accomplish this, only a modest helping of 18th century technology.
One of the great all-purpose sledgehammers in the toolbox of physicists and engineers is differential measurement; in other words, don't worry about the absolute value of something, but only the difference between things you can measure. For example, it is common practice for linemen repairing high-voltage power transmission lines to work on them, without cutting power, from insulated baskets raised by a crane. As long as the lineman is insulated from the ground, only the voltage difference between his hands and the line he's working on matters; after attaching the basket to the line, this is zero, so he might as well be repairing a grounded conductor. Now if, while working on a conductor at, say, 200,000 volts above Earth potential, he should happen to touch the tower, grounded to Earth, that would make for a really bad day. The trick is keeping the difference small; you can live your entire life at 1 million volts, and as long as everything around you is near that value, there is no way, even in principle, you could discover the absolute potential. This is the consequence of all the forces of physics being gauge invariant: absolute values don't exist—only differences matter.
What we're looking for, then, is a device which responds only to differences in gravitational attraction, canceling out the much stronger constant gravitational attraction of the Earth. We need look no further than a slightly modified version of the same device Henry Cavendish used in 1798 to first measure the gravitational constant, G in the equations above. Ever since, the torsion balance has been the primary tool used both for measuring the gravitational constant and testing the equivalence principle, which states that all bodies experience the same gravitational force regardless of composition; Einstein's General Relativity showed this to be a fundamental consequence of the structure of space and time.
State of the art torsion balances have measured the gravitational constant to better than one part per million and confirmed the equivalence principle to more than 11 decimal places. This requires extraordinarily refined and delicate laboratory apparatus and experimental design, in which a multitude of subtle effects must be compensated for or canceled out. We, however, aren't going to measure anything—we're only interested in observing universal gravitation. This allows simplifying the torsion balance to something we can set up in the basement.
The principle of the torsion balance is extremely simple. Suspend a horizontal balance arm from a vertical elastic fibre. At each end of the balance arm are masses, much denser than material of the arm, which respond to the gravitational force. Once the suspending fibre, balance arm, and weights are set up and brought into balance, the downward force of gravitation acts equally on every component. The balance arm is then free to rotate without any hindrance from the Earth's gravity. It is constrained only by air friction and the torsional strength of the support fibre—its resistance to being twisted. We can then place test masses near the ends of the balance arm and observe whether the gravitational attraction between them and the masses on the arm causes the balance arm to move. When measuring the gravitational constant one must precisely calibrate the torsional strength of the fibre, but to simply observe gravitation we need only make sure the fibre is sufficiently limp to allow the gravitational force to overcome its resistance to twisting.
In practice, the balance arm is so free to move that once any force sets it into motion, it oscillates for a long period, spinning round and round if free or bouncing back and forth off the stops if constrained. To avoid this we need to damp the system so kinetic energy acquired by the bar is more rapidly dissipated. Well, nothing's more damp than water, so we add a water brake to the arm which turns in a fixed reservoir. The resulting drag as the balance arm moves is much greater than air resistance and frictional losses in the fibre, and reduces the oscillations to a tolerable degree.
“The time has come,” the Hacker said,
“To talk of many things:
Of plastic foam—and tuna cans—
Of chunks of lead—and string—
And how the force of gravity—
Will make the balance swing.”
So here's the sophisticated, high-tech, big science apparatus we'll use to observe the subtle curvature of spacetime. An aluminium ladder serves as the support from which the balance arm is suspended. Nylon monofilament fishing line, as shown above, is knotted to the middle of the third cross-beam at the back of the ladder, one above the brace bearing the little white box, about which more later. Using a ladder or similar movable support frame allows setting up the balance in the middle of the room. This is important because we are bending spacetime in the basement, in this case an underground storage room at Fourmilab. Ground level is about even with the ceiling of this room, about 45 cm above the top of the ventilation window at the upper right of the picture. An underground room is ideal because it minimises temperature variations and vibration which might perturb the balance arm. Both walls shown in this picture are sunk into solid limestone rock—if you set up the balance near one of these walls, the gravitational field from all that rock will mask that of the test masses, and the balance will assume a “gravity gradient” position with one of the ends of the bar pointing toward the wall, and will budge only slightly under the influence of the test masses. With the bar in the middle of the room, the tidal influence of the mass of the wall and the rock behind it is reduced to a negligible value. The pipe on the wall at the right is part of the serpentine pressurised hot water heating system; it was disabled to prevent air currents from disrupting the balance arm. In fact, since the room is underground, the heating system is rarely engaged, and only in the depths of winter, never in June.
The balance arm is a 5 × 5 × 30 cm bar of plastic foam, hacked from a 5 cm thick slab of packing material with a Swiss Army knife. The bar is suspended in a cradle made of insulated telephone wire. The bar is held in its cradle by friction and the indentation made in the soft plastic foam due to the weights at either end of the bar; it's easier to adjust the bar for proper alignment this way than if it were glued to the cradle.
The nylon monofilament that suspends the cradle is barely visible at
the top of the picture—it is fastened by a knot to a loop formed
into the cradle wires by twisting them. The monofilament is a very
fine “six pound test” (about 3 kg capacity) fishing line
manufactured in Japan; a 300 metre spool of it costs about US$9. The
masses which cause the bar to turn when a gravitational force acts
upon them are lead “sinkers” used by fishermen, each
weighing 169 grams. Two are placed on each end of the balance beam,
giving it a total weight of 676 grams. Be sure to place the weights
on both ends of the beam simultaneously so it doesn't topple, then
adjust the placement so the beam is horizontal. Nylon monofilament is
very elastic: when you put the weights on the beam the support line
will stretch and the beam will end up closer to the ground. You may
have to adjust the attachment of the line to the ladder (or other
support) or, as I did, twist the cradle wires to restore the beam to
the desired height. Finally, when you first hang the beam, it may take
some time to release stresses in the fibre remaining from the
manufacturing process and from its having been rolled onto a spool.
It's best to let the arm hang for a couple of days, free to turn, to
allow these initial stresses to equalise before attempting any
experiments with gravitation.
The height of the beam is important because of the need for it to fit
properly with the water brake. If the beam is allowed to swing
freely, it will be terribly underdamped—once it starts to swing,
only air friction and the minuscule losses in the fibre will act to
stop it. This causes the beam to bounce around incessantly, masking
the steady influence of gravitation. The water brake dissipates the
energy of these unwanted oscillations precisely as an automotive shock
absorber does; the flap's motion does work on a viscous fluid, water
in this case, and deposits its energy in heating it.
The water brake consists of a flap which projects downward from the balance arm (in this case, a piece of aluminium cut with scissors from the tray of a “heat and eat” meal, fixed with white glue into a slot cut into the bottom of the balance beam). The flap projects into a reservoir (a tuna fish can) filled with water. A more viscous fluid such as salad oil would provide greater damping and less bouncing than water, but I opted for water since it's less icky to clean up when the inevitable spill occurs and can be disposed of when the experiment's done without a visit to the village recycling barrel.
If I were rebuilding the balance beam, I would use a longer and
narrower flap and/or a larger and deeper water reservoir. If
the flap is only slightly smaller than the inside diameter of the
reservoir, you have to be very careful that the flap and reservoir are
centred on the beam. Otherwise, the flap will touch the edge of
the reservoir and freeze the beam in place, as that frictional force is many
orders of magnitude greater than the gravitational force we wish
the beam to respond to. The water reservoir can be as large as
you like, as long as it doesn't interfere with placing the
test masses; the larger it is, the less you have to worry
about its being precisely centred.
Blocks of plastic foam support the test masses so their centre of gravity is at the same height as the masses at the ends of the balance beam, maximising the attraction. The foam also keeps the balls from tending to roll away. The black rectangle, actually an inverted mouse pad, serves as a background for the time display superimposed by the video camera, rendering it more readable when images are reduced in scale so movies download more rapidly.
Use the densest objects you can obtain for the ends of the balance beam and as test masses: lead sinkers, steel balls, plutonium hemispheres, etc. Density is important because the gravitational force varies as the inverse square of the distance between the centres of mass of two objects. With a dense substance, the centre of mass is closer to the surface, so you can get the centres of mass closer together and enhance the gravitational force. For example, consider two pairs of one-kilogram spheres, the first made of lead (density 11.3 g/cm³), the second of pine wood (density about 0.43 g/cm³), placed so the surfaces of each pair of spheres are 1 cm apart. A one kilogram lead sphere has a radius of 2.76 cm, so the centres of mass are separated by 1+2.76×2, or 6.52 cm. A one kilogram sphere of pine has a radius of 8.22 cm, by comparison, so the centres of mass of the two pine spheres are 1+8.22×2 = 17.44 cm apart. Taking the square of the ratio of these distances shows that the gravitational force between the lead spheres is more than 7 times that of the pine spheres. Since attraction is linear by mass but inverse square in distance, you're better off with a modest mass of high-density material than a large mass of a substance with lesser density.
It's best to use a nonmagnetic material like lead for the weights on the ends of the balance arm. The forces we're working with are so small that if you use, for example, steel ball bearings on the arms, you may end up accidentally reinventing the compass instead of detecting the force of gravity.
“So what's the little white box on the back of the
ladder?”, you ask. Okay…. It's a BSR Model 500
surveillance camera which lets me observe the state of the experiment
as it runs. The Sony camcorder I use to make movies doesn't generate
video output while recording, so I can't use its video feed to monitor
what's happening. Popping into the room where the experiment's
running is a no-no—air currents from opening and closing the
door, not to mention walking around in the room could seriously
disrupt things. The BSR camera and accompanying 13 cm (diagonal)
monitor allows keeping tabs on what's happening in a non-intrusive
manner. I made a custom interface of the BSR camera/monitor cable to
Fourmilab's ubiquitous RJ-45 cabling, so I can place two BSR cameras
anywhere on the site and monitor either from anywhere else. At the
right is an image from the surveillance camera taken at the end of an
experiment, confirming that the balance beam has come to a stop
against the foam block supporting the mass at the top. The camera is
sensitive to infrared and includes infrared LEDs to illuminate nearby
objects, and has a microphone linked to a speaker in the monitor.
This makes it ideal for anxious parents who wish to monitor their
sleeping baby; spacetime hackers can use the infrared illumination to
view the balance beam without the thermal disruption of incandescent
lamps or direct sunlight. The storage room where I ran this
experiment has fluorescent strip lighting on the ceiling, and I
observed no detrimental effects from its being illuminated. Of
course, if the room you're using is equipped with that low-tech
miracle called a window, you can dispense with all this complexity.
The following time-lapse movies (about 30 seconds per frame) show the
torsion balance responding to the gravitational field generated by two
740 gram competition pétanque balls. The picture at left shows the
camera angle employed in both movies. In each, the movie begins with
the bar stationary, in contact with one of the balls or the foam
supporting it. The balls are then shifted to the opposite corners,
where they attract the lead weights on the ends of the bar. The bar
then turns, slowly at first and then with increasing speed as it is
accelerated by the gravitational force growing as the inverse square
of the decreasing distance between the masses. The bar bounces when
it hits the stop on the other end, and finally, after a series of
smaller and smaller bounces as the water brake dissipates its kinetic
energy, comes to rest in contact with the closer ball or support.
This is the lowest energy state, at which the bar will always arrive
at the end of the experiment.
Pay no attention to the plastic robot ant—she's just curious. It's a long story.
What we've demonstrated by these experiments is the universality of gravitation; there is nothing special about the Earth that makes objects fall toward it. Everything attracts everything else; the Earth's attraction is greater simply because the Earth is so much more massive than the objects we encounter in everyday life. Only by canceling out the Earth's gravitation by means of a torsion balance were we able to observe the gravitational attraction between masses of less than a kilogram.
The universality of gravitation means that every object in the universe is interlinked in a web of mutual attraction; the universe is transparent to gravitation. The most distant galaxies exert a pull on you, as you do upon them—immeasurably tiny to be sure, but present just the same. From a practical standpoint, universality means there's no way to shield your torsion balance from the gravitational attraction of masses in its vicinity; you can only set it up sufficiently far from other massive objects so the attraction of the test masses predominates. One interesting massive object to consider is yourself (I use “massive” only in the sense of "possessing mass", not pejoratively; if you took it that way, perhaps you should check out my on-line diet book).
Using the gravitational force calculator earlier in this document, we can compute the gravitational attraction between the 338 gram mass at the end of the balance beam and the 740 gram test mass at the 14 cm distance when the beam is at the midpoint between the masses to be 0.000085 dynes. Now suppose you're crouching down in order to move the test masses, with your centre of gravity one metre from the closer test mass, and that you weigh 65 kg. Plugging these numbers into the calculator shows that your own gravitational attraction on the nearer end of the beam is 0.000147 dynes, 1.7 times as great as that of the test mass. Your actual influence on the motion of the balance arm is less, however, since what matters is the difference in force exerted on the masses at the two ends of the balance arm. Since your centre of gravity is more distant than the test masses, the difference is less.
Let's work it out. Assume the centres of gravity of the two masses on the balance arm are 25 cm apart, and that you're crouching so the arm makes a 45° angle with your centre of gravity, one metre from the centre of the arm. The nearer mass is then 17.68 cm closer than the more distant one and the difference in gravitational attraction (or tidal force) on the two masses is the difference in attraction on a mass 91.16 cm distant and one 108.84 cm away. The calculator gives the attraction on the near end of the arm as 0.0001764 dynes and the far end as 0.0001238 dyne, with a difference of 0.0000527 dynes. Now recall that the force exerted by the test mass was 0.000085 dynes, only 1.6 times as large, so even taking into account the reduced tidal influence due to your greater distance, the force you exert on the balance cannot be neglected. This makes it essential to remotely monitor the experiment so your own mass doesn't disrupt it.
In practice, air currents due to your motion and resulting from convection driven by your body's temperature being above room temperature may exert greater forces on the balance arm than the gravitational field generated by your mass. In any case, it's best to let the experiment evolve on its own, observed from elsewhere.
Nineteen centuries elapsed between the death of Archimedes in 212 B.C. and the publication of Newton's Principia in 1687. Given the philosophical implications of Newton's theory, it's interesting to speculate what might have happened had Archimedes discovered the universal nature of gravitation.
To do this, he would have had to suspect that attraction was
universal, suggest an experiment to confirm this, and perform that
experiment, with results validating the hypothesis. Here is
information in Archimedes' possession which might have suggested the
universality of gravitation.
Suppose then that, given these facts, Archimedes embarked upon the following chain of reasoning.
It seems plausible, then, given the knowledge at hand and a chain of
inference which, in retrospect at least, appears straightforward, that
Archimedes could have suspected the universality of gravitation. But
could he have demonstrated it? Unlike many scholars in
ancient Greece who contented themselves with philosophical arguments,
Archimedes was an intensely practical man, renowned as a military
engineer as well as a mathematician and philosopher. His laws of the
lever and buoyancy were tested experimentally, and so we should expect
he would subject any inference about gravitation to experimental
confirmation. Now that we've succeeded in bending spacetime in the
basement with common household materials of the late 20th century,
let's see if the experiment can be done using only materials
Archimedes might have employed.
Let's try to redesign the torsion balance using only materials available in antiquity.
In choosing and using any natural fibre support, you have to approach the project with a willingness to learn by trial and error and a great deal of patience. Each kind of fibre has its own “personality”, and the quirks can take some time to understand. For example, many plant and animal fibres are sensitive to moisture—if a summer thunderstorm increases the relative humidity from 50% to 99% in the space of an hour, your balance arm may start to swing wildly as the fibre absorbs moisture from the air. Further, plant fibres tend to tear, both under tensile stress and when twisted. This can cause your balance beam to “spontaneously” shift to a different equilibrium point or, after having been displaced, return to a different location than the starting point.
Would Archimedes have appreciated the importance of choosing a supple and well-behaved support fibre? I think so. From the radius of the Earth, which he knew, and assuming its density to be the same as rocks such as limestone (about half the actual density of the Earth), the ratio of the Earth's mass to that of whatever test masses were employed could be estimated within a factor of two. Making the simplest assumption (which has the additional merit of being correct) that attraction is proportional to mass, it is clear that the force acting on the masses at the ends of the balance arm is minuscule, so a fibre which offers the least possible resistance to twisting should be employed.
The concrete floor would also seem strange to Archimedes, but it is irrelevant to the experiment. A smooth stone floor, as existed for millennia before, would produce identical results.
The following movies demonstrate universal gravitation with an apparatus which, as argued above, could have been conceived by Archimedes and built from materials he could readily obtain.
Suppose this had happened. Consider how easily it could have. Would such a discovery in Archimedes' time have had an impact comparable to Newton's or, occurring in a very different social and intellectual milieu, would it have been regarded as no more than a curiosity? How might human history have played out had the Enlightenment begun 1900 years before Newton?
Click on titles to order books on-line from