Gui Ruiyan

Christiaan Huygens · 克里斯蒂安·惠更斯
On the Motion of Colliding Bodies · 论碰撞物体的运动

On Huygens’ Second Hypothesis and the Experiment that Proves It

Paper / 论文

Of the three hypotheses Huygens employs at the start of his treatise, On the Motion of Colliding Bodies, Hypothesis II seems peculiar. The first hypothesis has been familiar to us since the time of Galileo, which states that a body keeps moving uniformly if nothing disturbs it. This is a universal hypothesis that is employed in all investigations of motion in classic physics. The third hypothesis, which posits the relativity of motion, is well recognized in our daily life, as Huygens himself shows when he gives the example of a passenger on a boat observing two balls moving. This is also a universal hypothesis. The second hypothesis, however, which states that ‘when two hard bodies, equal to each other and having equal speeds, directly collide with one another, each rebounds with the same speed with which it arrived’, is the only one of the starting hypotheses that is particular to the problem of collision. Therefore, the experiment that verifies this hypothesis can only be carried out through a setting of colliding bodies.

The ideal setting of this experiment should therefore entail: 1. two perfectly hard and equal bodies 2. that move with equal speeds through friction-less space and 3. collide ‘in the same straight line which passes through the centers of gravity of the two bodies’.

The two colliding bodies need to be hard because only such bodies can be perfectly elastic without varying in shape. Elasticity guarantees that the total quantity of motion, or what Descartes would call the ‘total force’, of the system is preserved and is not transmitted to the infinite inside of matter. Constancy of shape, on the other hand, makes sure that in any instant of collision or contact, the center of gravity does not change place, so that the collision is always direct.

The space needs to be friction-less because otherwise the resistance might cause the body to decelerate. Though it is possible that one can apply a force to the body so that resistance is overcome and the body keeps moving equably, it is extremely difficult to do so in this experiment, without affecting the final results.

The collision needs to be direct, i.e. kept at the straight line that passes through the two centers of gravity, because first, if the motion or contact takes place at an angle, the absolute rebound ‘speed’ of each body will not conform to the hypothesis, for the body rebounds in a direction different from the one whence it comes, and in this case, the concept of speed becomes inadequate since it does not take account of the change in direction; second, the impact during contact will set the body to revolve around itself, rendering the analysis more complicated.

The ideal setting of any experiment, however, can never be reached in practical conditions. The compromises one makes to reconcile theory with practice for this particular experiment are roughly the following.

To start with, we take two iron balls to be our colliding bodies. We choose iron as the material since it simulates the perfect hardness better. The body takes a spherical shape because, first, the curvature of it reduces the effect of air resistance; second, the center of gravity of each body is easily determinable, since sphere is a regular geometrical shape and every part of the outside of a sphere is equidistant to its center; finally, the contact surface between spheres is minimal, hence no air pressure difference might greatly affect the results. Next, we weigh the two bodies on a balance and make sure their masses are equal. Since they are both spheres made out of iron, their surface areas will be equal if their masses are equal, and the air resistance each with equal speed suffers will also roughly be equal. As this is done, we have determined all the qualities belonging to the two bodies themselves that are conducive to the success of this experiment, and the first condition for this experiment, i.e. to have two perfectly hard and equal bodies, is fulfilled as nearly as possible.

The second condition of this experiment is that these two bodies should move with equal speeds through friction-less space. The way we approximate this ideal condition is through the following construction. First, we substitute the two bodies moving on the same straight line, which is our original conception of the experiment in its ideal setting, with two pendulums. We take the two balls that are already selected to be our colliding bodies and suspend them in midair, their centers of gravity lined up and their surfaces barely touching when at rest, then we displace them, release them and let them collide. Next, we set a light source directly above the two bodies in order to cast their shadows. A table is placed below the two bodies, and a large white paper is placed on the table to reflect the cast shadows, making our measurement taking easier. Finally, we need to determine our measure. The place we measure from is the initial position at which each body is at rest and touching the other. That tangent which touches the shadows of both bodies cast on the white paper is marked down as this initial position. The speed at which the two bodies move toward each other and the speed at which they rebound are measured by the corresponding horizontal displacement from this initial position. The former is measured by the displacement from the initial position to the position at which each body is to be released. The latter is measured by the displacement from the initial position to the position at which each body reaches its maximum displacement after impact.

Further explanations are required in order to fully justify our current construction. To start with, we have to elucidate why it is necessary to substitute the two bodies moving on a straight line in our ideal setting with two pendulums. There are a few difficulties involved in this original and ideal conception of the experiment. First, in reality, the plane on which the two bodies move always has significant friction. But pendulums move through only air, which exerts much less friction. Second, the measure of speed in the original conception would require the observer to mark down either the time it takes the body to traverse a given distance or the distance traversed by the body in a given time. To measure with accuracy a quantity that always changes linearly, i.e. the time or distance traversed by a moving body, is much more strenuous than to measure something recurring and cyclic, such as simple displacement in the case of a pendulum, which always returns to the same rebound position unless air resistance hampers it. Due to these problems, it is certainly desirable if one could replace the original conception with a construction much more expedient in real practice.

Next, we need to prove that the pendulums indeed are legitimate substitutes. A few assumptions are employed in this process. First, when displacement is small, we learn, through geometrical demonstration, that the final speed of the pendulum can be considered as proportional to the displacement. We also know that the motion of a pendulum can be analyzed into an infinity of instantaneous linear motions that are tangential to the curve which the pendulum traces along its way. Finally, we know through Galileo and Huygens that height, whose subduplicate is proportional to displacement, and speed are mutually determined, that is, as a body falling from a height determines a certain ultimate speed, so a speed ultimately moves up to a determinate height: therefore, the displacement at which a body is released determines its ultimate speed, and vice versa, the rebound speed determines its final displacement. With these assumptions, we set the initial position where the two bodies are at rest and barely touching each other, so that the experiment can best simulate the direct collision of uniformly moving bodies: only so can we make sure that as the two bodies collide, they will reach the bottom where the tangents on which the two bodies move at the instant of impact converge into one straight line, and since acceleration at that point is 0, they can be considered to move uniformly. By these tokens, we see that as long as displacement is kept small, the simple pendulums thus set up are good substitutes for the original conception of this experiment: at the instant of impact, the two pendulums can be considered to be two bodies directly colliding, and their respective speeds are represented through their displacements. One other assumption, despite being not useful for our present end of substantiating Hypothesis II, will nevertheless be most crucial if we later move on to substantiate other propositions with this experiment setting. The assumption is that the impact between two pendulums always takes place when the two bodies reach their respective lowest positions, i.e. when they reach the initial position, if the displacement is small and if the two bodies are released simultaneously. This assumption is reasonable, since we know that pendulums move isochronously within certain limits. This means that when displacement is still small, whether the body is dropped from a larger or a smaller displacement does not affect the time it takes for the body to reach the bottom. Therefore, two bodies always collide at the same place when displacement is small and when they are released simultaneously, for they reach the initial position at the same time and make contact there. It is not helpful for our present end because two equal bodies with equal speeds always directly collide at their common center of gravity. This assumption is thus not needed to guarantee that the two bodies should collide always at the same position.

The data we gather from this experiment are the following. When the two bodies are released both from displacement 20 cm, Body A rebounds to displacement 18.5 cm, whereas Body B rebounds to displacement 18.3 cm. When the two bodies are released from displacement 40 cm, Body A rebounds to 33 cm, whereas Body B rebounds to 35 cm. A third trial shows that as the two bodies are released from 10 cm, Body A rebounds to 10 cm, while Body B rebounds to 8 cm.

I judge that we see Hypothesis II substantiated by this experiment. For our final readings for rebounds follow closely our readings for releases. Sources of error are multiple, here I only include the most manifest. First, all bodies suffer air resistance in their motion. That is why nearly all readings for rebounds are less than readings for releases. Moreover, if the observer misses the first reading for rebound, they will have to wait for the next rebound, and during that wait time the motion of the body might be retarded even more by air resistance. Second, the two bodies are to be released simultaneously, otherwise they will not reach the bottom at the same time, and impact will not occur where we expect it to occur. To reduce this error, we agree to count to three and release. Yet human reaction time might still affect this execution. Third, since the smaller the displacement, the more accurate are the results, we see in our three trials that as the displacement is the greatest, the difference between rebound readings and release readings is also the greatest in value. Nonetheless, we do have an anomaly in our third trial, where one rebound reading is twenty percent less than the other. This might be due to a malfunction of the pair of strings that suspend Body B. Finally, since we are measuring from the cast shadows, the positions of which, however, do not entirely overlap with the positions of bodies, there is also room for latent error.