I came up with an idea of motion of center of mass the day before yesterday but couldn't find any solid answer for the cause of its motion.

Take the example of a man (say of mass $50$$kg$) and a tank (of negligible mass) filled with water kept at some distance (whose total weight is also about $50$ $kg$). Now suppose there is a pipe of uniform cross section (whose mass can be ignored) connecting the mouth of the man to an opening at the bottom of the tank. Also the surface on which they are standing is sufficiently rough (so that they don't move).

enter image description here

At $t=0$ the system is released from rest. So the center of mass of the system is exactly in the midway at $t=0$ (represented by the dashed line). But as water comes out, the center of mass of the system shifts towards the man . This means that the center of mass accelerated from its initial position. And for this to happen we need an external force.

There are only two external forces in play :

1: Static Friction force on the tank ($f_{tank}$), in the left direction

2: Static Friction force on the man ($f_{man}$), in the right direction

The direction of the acceleration of the center of mass is always towards the man but this may not be true with the net frictional force. Either of the two static frictional forces can be greater than the other or they may even be equal (in which case the center of mass shouldn't accelerate).


1: How can the frictional force alone can account for the direction of the acceleration of the center of mass ?

2: Is such type of motion of center of mass possible in which the components of the system don't move but the center of mass does ?

Can someone clarify whether it's friction or gravity causing the acceleration of the center of mass ?

Note : I have taken the tank , the man and the pipe as my system. And I am looking for the horizontal motion of the center of mass only so didn't mention gravity or atmospheric pressure on the upper surface of water in the tank.

  • 3
    $\begingroup$ You have forgotten gravity. If there is no gravity then the water does not flow and the COM does not move. $\endgroup$
    – gandalf61
    Commented Jun 17, 2021 at 5:58
  • 1
    $\begingroup$ even if the man wasn't there the COM would shift right? $\endgroup$
    – lineage
    Commented Jun 17, 2021 at 5:58
  • $\begingroup$ @gandalf61 seconds, just seconds $\endgroup$
    – lineage
    Commented Jun 17, 2021 at 5:59
  • $\begingroup$ @gandalf61 why would gravity make the center of mass move horizontally ? $\endgroup$
    – Ankit
    Commented Jun 17, 2021 at 6:00
  • $\begingroup$ @Ankit Pressure $\endgroup$
    – lineage
    Commented Jun 17, 2021 at 6:05

3 Answers 3



  1. its friction that is responsible for the COM's acceleration.
  2. No.

Earlier on $\ldots$

This is an interesting question. As formulated right now - a vertical water tank that has sprouted a leak in its wall- it is not evident that even though in the absence of gravity the COM wouldn't move, gravity itself isn't important to the spirit or solution of the question.

Indeed, in an earlier answer, I had wrongly argued that the star of the show - friction on the tank - plays no role. And I had good reason to. (see the detailed rebuttal at the end). I had ignored the knot in my stomach that was forming the moment I had accepted that a vertical force could lead to horizontal acceleration; its a fluid so maybe $\ldots$

New perspective

Let us simplify your system without loosing any of the physics. Let us replace the leaky water tank with a clamped or constrained rocket, at rest wrt. lab . The water is now replaced with the exhaust gasses.

We see that while switching off friction (or clamping, in the case of rocket) leads to the same behavior of COM in both cases - no movement - switching of gravity doesn't : while the water doesn't move, the rocket's exhaust is unaffected. This shows that gravity only plays the role of providing the kinetic energy to the water plume. Even thought it was a vertical external force that provided the energy, it need not necessarily be so, as in the case of the rocket, where exhaust energy could have come from say, burning fuel.

So the rocket accelerates some gas and the gas accelerates the rocket back equally and oppositely, right? Only in the frictionless case.

In the 'friction-full' case , the rocket (tank) is clamped. So the gas pushing of the rocket is the same as pushing off the ground - the constraint force immediately pushes back keeping the rocket stationary and the gas accelerating.$^1$

Since the COM moves with the accelerating gas, its friction which is responsible for the movement. Strangely enough, in this sense the acceleration of the gas or water is no different than ordinary walking, where one similarly pushes of the ground, on earth or on moon.

Another line of thought which leads to the same result and would have avoided the pitfalls of reasoning in terms of the internals of the system is the blackbox argument - when in doubt use a blackbox.

Lets put the entire rocket/tank inside a blackbox. Since the COM is moving, we see the box moving horizontally. The only horizontal force acting on the box is friction so it must be the cause for the COM's horizontal acceleration. (Why its acting at all - only those inside the box know).

The second question

2: Is such type of motion of center of mass possible in which the components of the system don't move but the center of mass does ?

No. The COM is used to characterize the state of motion of the system as a whole. Its the average of their motions. If every component of a system is at rest wrt. a lab frame, so would be their average and hence the COM.

Rebuttal of the earlier answer

In my previous answer I had given a similar case involving force redirect and COM movement resulting from gravity:

Consider the case of a mass hanging from a rope on a pulley. The other end of the rope, going over the pulley, attaches to another mass but this one initially rests on a horizontal table. Upon progressing time, though gravity acts vertically on the first mass and accelerates it downwards, the second mass accelerates horizontally. Why is that?

I has used the above analogous scenario along with the following observation

Its a similar reasoning for the acceleration of the COM of a sprouting water body. The weight of the water column in the tank is responsible for generating a pressure in the bulk of the liquid (const. at a given depth). This pressure by Pascal's law acts uniformly from all directions at any given point. This nature of pressure makes it a "3D pulley" - it can redirect the force of gravity in all directions.

to argue that it was the redirection of gravity by gauge pressure that was responsible for the COM's movement. After all, if it could do it once, it could do it again.

What about the friction on the tank? Here's what I had said earlier:

$\ldots$ the exhaust of the water plume as it gushes out on the left under the gauge pressure from the right, exerts a force on the tank. Static friction counters this force. It therefore isn’t responsible for the shifting COM.

These three deceptively simple arguments had together convinced me that gravity was the protagonist of this problem. Alas, upon further (and tiresome) scrutiny, they fall apart:

  1. In the case of the pulley, the net horizontal acceleration of the COM isn't provided by gravity. It is provided by the normal reaction at the pulley. This is most clearly seen in the blackbox argument as there is no other horizontal force in this system to begin with.

  2. Yes, Pascal's law allows for a redirection of weight but it is still an internal force albeit sourced because of an external force. The fact that the same pressurization of fluid can be obtained using internal mechanisms bears out this point. Moreover, the pressure is hidden in the blackbox argument and so its gravitational origin is of no concern.

  3. The fact that friction counters the exhaust's force on the tank is correct. But by the same reasoning, it also supplies the reaction on the exhaust. Since there is no other force on the exhaust, it must accelerate and take the COM with it, thus establishing friction as the driving force.


$^1$ In fact both friction and the force on exhaust can be calculated to be $\rho v_{exhaust}^2 A$ where $\rho$ is fluid density, $H$ is column height, $A$ is orifice cross-section.

A.1 The man seems irrelevant to the discussion about the accelerating COM. "The COM moves from some initial value" - that alone is enough to inquire about the external force. Using an equal mass oppositely placed, besides bringing the initial COM to center, plays no differentiating role.

A.2 Thanks to Dale for motivating further scrutiny.

  • 1
    $\begingroup$ Nice example of rope and pulley. The water tank is like the rope and pulley, but turned upside down and run in reverse. $\endgroup$
    – gandalf61
    Commented Jun 17, 2021 at 6:32
  • 1
    $\begingroup$ the man is definitely seems ready to drink 50L water though :-) $\endgroup$ Commented Jun 17, 2021 at 6:48
  • $\begingroup$ @lineage thanks 😊.. $\endgroup$
    – Ankit
    Commented Jun 17, 2021 at 7:32
  • $\begingroup$ @lineage take a look at Dale's comment .. in the Pranshu's answer.. $\endgroup$
    – Ankit
    Commented Jun 17, 2021 at 12:52
  • 1
    $\begingroup$ @lineage wow! I just noticed the revisions. Very nice work! +1 This answer is now quite a bit better than mine $\endgroup$
    – Dale
    Commented Jun 20, 2021 at 2:58

To simplify this problem why not replace the person with just another tank, having a person there adds no value to this problem as he is not actively sucking on the liquid.

This leaves us with two identical tanks connected with a pipe. One tank is filled with water the other is empty. Now the only thing that can make this water move through the pipe is the force of gravity which moves the centre of gravity from the water-filled tank to the centre of the pipe.

You might wonder how a downward force can make the centre of mass move laterally, this is where the static frictions you talk about come into play. if the tanks had no friction relative to the floor, the system will move in the opposite direction on the flow of fluid through the pipe making sure the centre of mass has no lateral motion. Since we have static friction, the tanks have a lateral force acting on them enabling our center of mass to move laterally.

  • $\begingroup$ hello 🤗.. thanks for ur answer. BTW I introduced friction so that my components of the system don't move but the center of mass does .. to create a sort of illogical phenomenon.. ;) $\endgroup$
    – Ankit
    Commented Jun 17, 2021 at 7:13
  • 1
    $\begingroup$ Ok then I misinterpreted your problem, I thought you were talking about the friction on the fluid. let me fix my answer $\endgroup$ Commented Jun 17, 2021 at 7:17
  • $\begingroup$ This answer is correct. The other answers are wrong $\endgroup$
    – Dale
    Commented Jun 17, 2021 at 11:16
  • $\begingroup$ @Dale but if it's true , why is the direction of net static friction force always along the left direction ? It could be zero also (I guess) .. or it could be along the right depending on the conditions .. am I wrong ?? $\endgroup$
    – Ankit
    Commented Jun 17, 2021 at 12:14
  • $\begingroup$ @Ankit in this example the net static friction force at all times acts in the direction of the acceleration of the center of mass. Remember, the static friction force is given by an inequality, so it can take any arbitrary direction and any magnitude less than the maximum. $\endgroup$
    – Dale
    Commented Jun 17, 2021 at 12:31

Can someone clarify whether it's friction or gravity causing the acceleration of the center of mass ?

It is 100% unambiguously clear that it is the static friction which causes the horizontal acceleration of the center of mass (COM). Place the experiment on a smooth surface and the water will flow but the tank and everything will slip to the right with the COM remaining in place. When the friction force is absent the COM does not move horizontally, when it is present the COM does move horizontally. It is clearly and unambiguously the force responsible for the horizontal acceleration.

The root of the question seems to be an incorrect thought that static friction should not be able to produce acceleration. This is incorrect, in fact, it is static friction which produces acceleration in most land vehicles, including automobiles, trains, bicycles etc. Static friction even produces the acceleration for walking. As you are probably personally familiar, reducing the static friction makes walking or even standing quite difficult.

So the acceleration of the COM by static friction is not merely possible, it is a common part of everyday life.

Some of the answers have incorrectly stated that it is gravity which causes the horizontal acceleration of the COM. This is incorrect. Per Newton’s 2nd law the acceleration of the COM requires a net external force in the direction of the acceleration. Gravity does not contribute any horizontal component to the net external force, so it does not contribute to the horizontal acceleration. There is nothing in Newton’s laws corresponding to “redirecting” a force.

What gravity does contribute, in this case, is energy. The water flows down and gravity does work on it as it flows down. Work is different from force and energy is different from momentum, so it is incorrect to conflate them.

While the force to accelerate the COM must be external, the energy may be internal. For example, we could replace gravity by a piston on the top of the tank pushed down by an internal spring. Provided we somehow maintain the friction force, the COM would move exactly as before. The acceleration of the COM would still come from the external static friction force, but now the energy would come internally from the spring.

  • $\begingroup$ presuming you are talking about my answer, I did say static friction is accelerating the system, but I argue against the notion that it is the cause of motion (which in a way both these forces are so it is kind of inane to ask) as OP was asking whether friction force is alone responsible for motion of the centre of mass in their first qn $\endgroup$ Commented Jun 19, 2021 at 6:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.