Moments and the centre of mass for point masses and for a rod with varying density, the centroid of a plane region and the factor of one half in the vertical moment, the symmetry principle, and the Theorem of Pappus.
Subject: Calculus I · 74 slides · symbolic lesson
Open the interactive version of this deck
Title
Calculus I · Section 6.6
Where an object balances, and the volume that fact hands you for free
Objectives
Every application so far has integrated a quantity: area, volume, work. This one integrates position weighted by mass, and then divides. The quotient is a point, and a surprising amount of geometry follows from it.
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, pp. 610-625 — learning objectives 6.6.1 to 6.6.4
Everything in Chapter 6 so far has answered a how much question: how much area, how much volume, how much work. This section answers a where question instead, and the answer is a point rather than a number.
The idea behind it is one you already trust physically. A plate balances on a stick at one particular spot, and a seesaw balances when the heavier child sits closer in. Turning that instinct into arithmetic takes one new quantity, the moment, and one division.
Watch for two things as you go. First, the same three-step shape recurs for every object here: total mass, moment, divide. Only the way you compute the two totals changes, from a sum for point masses to an integral for a rod and to a pair of integrals for a plane region. Second, two shortcuts appear near the end. Symmetry hands you a coordinate with no work at all, and the Theorem of Pappus converts a centroid into the volume of a solid of revolution, which is the chapter's last and prettiest trick.
Warm-up
Discussion prompt
A child of 40 kilograms sits 2 metres from the pivot of a seesaw. Where must a 20 kilogram child sit to balance it? Write down the rule you used, in words, before reading on.
Write your answer before reading the reveal, even if you are sure. The rule you reach for here is the rule the whole section generalises, and it is worth knowing whether you had it in words already.
Most people say the lighter child must sit further out, which is right, but the useful version is sharper than that: what has to match on the two sides is mass multiplied by distance. Forty times two and twenty times four are both eighty, and that shared value is what the pivot feels.
Notice that you never needed to know anything about the plank, or about gravity, or about the units. Only the products matter. That is why the same arithmetic will work for kilograms on a number line, for a rod that gets denser along its length, and for a flat metal plate.
Section
Part 1
Concept
Figure (svg): A beam resting on a triangular fulcrum. A 40 kilogram block sits 2 metres to the left of the fulcrum and a 20 kilogram block sits 4 metres to the right; double-headed arrows mark the two distances.
Two masses on a beam balance exactly when the two products agree. Doubling a mass and halving its distance changes nothing.
\[ m_1 d_1 = m_2 d_2 \]
Distance alone does not balance a seesaw, and mass alone does not either. The quantity that does is the product, and it is called the moment about the pivot.
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 611 — Figure 6.62 and the balance condition
Look at the picture and check the arithmetic on it. The heavy block is half the distance out and twice the mass, and those two factors exactly cancel. If you slid the heavy block one metre further from the pivot, its product would rise to one hundred and twenty and the beam would tip.
The product of mass and distance is the quantity that matters, and it is called the moment about the pivot. This is the same word you may have met in a physics course as torque, and the reason is identical: it measures how hard something twists about a point.
One detail to store away now. The moment depends on which point you measure from. Move the pivot and every distance changes, so every moment changes too. The rest of this section is really the question of which point makes all the moments cancel.
Concept
Fix two masses at positions on a number line and ask where the pivot belongs. Call that unknown position the centre of mass. Each distance is now measured from it.
\[ m_1|\bar{x} - x_1| = m_2|x_2 - \bar{x}| \]
Drop the bars
Why: The pivot lies between the two masses, so both differences are positive.
\[ m_1(\bar{x} - x_1) = m_2(x_2 - \bar{x}) \]
Expand both sides
Why: Multiply through.
\[ m_1\bar{x} - m_1x_1 = m_2x_2 - m_2\bar{x} \]
Collect the unknown
Why: Move both terms carrying the unknown to the left.
\[ \bar{x}(m_1 + m_2) = m_1x_1 + m_2x_2 \]
Divide
Why: The total mass is positive, so this is legal.
\[ \bar{x} = \frac{m_1x_1 + m_2x_2}{m_1 + m_2} \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 611 — Figure 6.63 and the derivation
This is the derivation to be able to reproduce, because it is four lines long and it explains where the formula comes from. Start from the balance condition of the last slide, with the distances now measured from the unknown pivot position rather than given to you.
The absolute-value bars come off because the pivot must lie between the two masses, which makes both differences positive. After that it is ordinary algebra: expand, gather the terms containing the unknown on one side, and divide by the total mass.
The step that people rush is the third one, where both terms carrying the unknown are collected. Notice that they come from opposite sides of the equation and that one of them changes sign on the way, which is exactly why the total mass appears as a sum rather than a difference. Get that sign wrong and you will produce a formula with a difference of masses underneath, which blows up whenever the two masses happen to be equal.
Intuition
Read the last line again as an average. Each position appears once for every unit of mass sitting there, and the total is divided by how many units there were altogether.
\[ \bar{x} = \frac{m_1}{m_1+m_2}\,x_1 + \frac{m_2}{m_1+m_2}\,x_2 \]
The two fractions are positive and add to one, so the answer is always somewhere between the two positions, nearer the heavier one. That single fact will catch most arithmetic slips in this lesson.
Reading the result as a weighted average is the single most useful reframing in this lesson. Each position gets a share of the vote, and the size of its share is its fraction of the total mass.
Two consequences follow immediately and both are worth carrying. The shares are positive and add to one, so the answer must lie between the smallest and the largest position. And because only the shares matter, multiplying every mass by the same factor changes nothing at all.
You can use the first consequence as a free check on every problem in this lesson. If your balance point comes out beyond the ends of the object, you have either forgotten to divide or you have made a sign error in the moment. It takes two seconds to look and it catches the majority of slips.
Concept
moment about the origin — The sum of each mass multiplied by its position. It measures how strongly the system tends to turn about the origin, and it can be negative.
\[ M = \sum_{i=1}^{n} m_i x_i, \qquad m = \sum_{i=1}^{n} m_i, \qquad \bar{x} = \frac{M}{m} \]
| quantity | units | can it be negative? |
|---|---|---|
| total mass m | kilograms | no |
| moment M | kilogram-metres | yes, if mass sits left of the origin |
| centre of mass x̄ | metres | yes, and it always lies between the extreme positions |
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 612 — Theorem 6.9, equations 6.14 and 6.15
Three symbols, and they are easy to confuse because they all measure something about the same system. The table is there to keep them apart by their units, which is the quickest way to tell which one you are holding.
The total mass is a sum of positive numbers, so it is positive. The moment is a sum of signed terms, because a position left of the origin is negative, so the moment can come out negative or even zero. The centre of mass is a position, measured in the same units as the positions you started with.
Pay attention to the phrase about the origin. The moment is always taken about some point, and unless you are told otherwise that point is the origin. Later in the section you will meet moments taken about the two axes of the plane, and the same warning applies: name the thing you are balancing about before you write anything down.
Notation
Annotate
On: \( \bar{x} = \frac{\sum_{i=1}^{n} m_i x_i}{\sum_{i=1}^{n} m_i} \)
Work through the four annotations before moving on. The first two tell you what each half of the fraction is for, and the third one is the one that costs marks.
That third point deserves a slow read. The positions in the sum are signed coordinates, not distances. A mass sitting at minus three contributes a negative term, and that negativity is precisely how the arithmetic knows the mass is pulling the balance point to the left. If you quietly replace coordinates by distances, everything becomes positive and the answer will be nonsense.
The last annotation is the check you should run every single time. The result is a weighted average of the positions, so it has to land between the smallest and the largest of them. If it does not, stop and re-add.
Worked example
Find the moment about the origin and the centre of mass.
\[ m_1 = 30 \text{ at } x_1 = -2, \qquad m_2 = 5 \text{ at } x_2 = 3 \]
\[ m_3 = 10 \text{ at } x_3 = 6, \qquad m_4 = 15 \text{ at } x_4 = -3 \]
Write the moment term by term
Why: Each mass times its signed position.
\[ M = 30(-2) + 5(3) + 10(6) + 15(-3) \]
Add the four terms
Why: Two are negative, two positive.
\[ M = -60 + 15 + 60 - 45 = -30 \]
Add the masses
Why: Nothing signed here.
\[ m = 30 + 5 + 10 + 15 = 60 \text{ kg} \]
Divide
Why: Moment over total mass.
\[ \bar{x} = \frac{M}{m} = \frac{-30}{60} = -\frac12 \]
Figure (svg): A number line from minus 4 to 7 carrying four blocks whose heights show their masses: 15 kilograms at minus 3, 30 kilograms at minus 2, 5 kilograms at 3 and 10 kilograms at 6. A fulcrum sits just left of the origin at minus one half.
Check the answer lies between the extremes
Why: The positions run from minus 3 to 6, and minus one half is inside that range; it sits left of the origin because 45 of the 60 kilograms do.
\[ -3 \le -\tfrac12 \le 6 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 612 — Example 6.29
Nothing here is hard, so use it to fix the habits. Write the moment out term by term rather than in your head, because that is where the sign slips live: two of these four terms are negative.
The arithmetic gives a moment of minus thirty and a total mass of sixty. The division is the step people skip when they are hurrying, and the moment on its own, minus thirty, looks nothing like a position on a line that runs from minus three to six.
Look at the picture and see whether the answer feels right. Forty-five of the sixty kilograms sit to the left of the origin, so the balance point should be left of it, and it is. But it is only just left, at minus one half, because the ten kilograms at six are far enough out to nearly compensate. Being far out is worth a great deal in a moment, which is the whole content of the seesaw rule.
Fill the middle
The moment of the previous example was minus 30 and the answer was minus one half. Fill in the step between them.
Fill in the blanks
\bar60 = \frac______} = -\frac12
Why: The divisor is the TOTAL mass, 30 plus 5 plus 10 plus 15, which is 60 kilograms. Dividing by anything else, such as the number of masses, gives a number with the wrong units and no meaning as a position.
Fill in the blank before revealing it. The temptation, especially under time pressure, is to divide by the number of masses, which here would be four, because that is what an ordinary average does.
But this is a weighted average, so the divisor is the total weight, which is the total mass of sixty kilograms. Dividing by four would give minus seven and a half, a position nowhere near the objects.
A units check settles it instantly. The moment is in kilogram-metres. To get metres out, you must divide by something in kilograms, and the only kilogram quantity in sight is the total mass.
Worked example
\[ m_1 = 12 \text{ at } {-4}, \quad m_2 = 12 \text{ at } 4, \quad m_3 = 30 \text{ at } 2, \quad m_4 = 6 \text{ at } {-6} \]
Write the moment
Why: Signed positions again.
\[ M = 12(-4) + 12(4) + 30(2) + 6(-6) \]
Add
Why: The first two cancel exactly.
\[ M = -48 + 48 + 60 - 36 = 24 \]
Add the masses
Why: Same total as the previous example, by coincidence.
\[ m = 12 + 12 + 30 + 6 = 60 \text{ kg} \]
Divide
Why: Moment over total mass.
\[ \bar{x} = \frac{24}{60} = \frac{2}{5} = 0.4 \]
Check against the heavy mass
Why: Half the mass, 30 kilograms, sits at 2, and the rest is spread either side of the origin; a balance point just right of the origin is exactly what that picture predicts.
\[ -6 \le 0.4 \le 4 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 612 — Checkpoint 6.29
This one is set up to look like the previous example and behave differently. The total mass is again sixty kilograms, but the answer comes out on the other side of the origin.
Watch the first two terms cancel exactly: twelve kilograms at minus four and twelve at four contribute nothing at all to the moment, because they are mirror images of each other. That is symmetry doing work for you before you have even met the symmetry principle, and it is worth spotting because it removes two of the four products.
What is left is thirty kilograms at two, pulling right, against six kilograms at minus six, pulling left. The first wins comfortably, so the balance point lands at zero point four, just right of the origin, and safely inside the range of the positions.
Trap
A write-up that stops one line early:
\[ M = -30 \]
\[ \bar{x} = -30 \;? \]
Wrong. The division was never done.
The moment is measured in kilogram-metres, not metres, and it is far outside the range of the positions. A balance point of minus 30 for masses that live between minus 3 and 6 is impossible.
\[ \bar{x} = \frac{-30}{60} = -\frac12 \]
Two habits stop this: check the units, and check that the answer lies between the extreme positions.
This is the most common single-line error in the whole section, and it happens because the moment is the hard part and the division feels like a formality. You do the interesting work, write down minus thirty, and stop.
Of the two defences named on the slide, the units one is the more reliable, because it works even when you have no feel for the numbers: kilogram-metres are not metres, so a division by a mass is still owed. The range check is the faster one once you do have a feel for the problem.
The same trap reappears in every later part of this lesson, with moments of rods and moments of plane regions. The cure never changes. A moment is not a position until you have divided by the total mass.
Concept
Prove that the balance point earns its name: measured from it, the moments cancel.
Write the moment about a general point c
Why: Every position is now measured from c.
\[ M_c = \sum_{i=1}^{n} m_i (x_i - c) \]
Split the sum
Why: Sums distribute over the subtraction.
\[ M_c = \sum_{i=1}^{n} m_i x_i - c\sum_{i=1}^{n} m_i = M - cm \]
Set the total moment to zero
Why: That is what balancing at c means.
\[ M - cm = 0 \]
Solve for c
Why: Divide by the total mass.
\[ c = \frac{M}{m} = \bar{x} \]
Check on Example 6.29
Why: Measured from minus one half the four terms are 30 times minus 1.5, 5 times 3.5, 10 times 6.5 and 15 times minus 2.5, that is minus 45, 17.5, 65 and minus 37.5.
\[ -45 + 17.5 + 65 - 37.5 = 0 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 611 — the balance point of the system
This short proof is what justifies the name. Balancing at a point means that the turning effects about that point cancel, and turning effect is exactly what the moment measures, so balancing means the moment about that point is zero.
The algebra is three lines. Write the moment about a general point, split the sum into the original moment minus the point times the total mass, set the whole thing to zero, and solve. The point that comes out is the quotient you already had.
The check at the end is worth doing yourself on paper. Measure each of the four positions from minus one half, multiply by its mass, and add. The positive and negative contributions cancel exactly, which is the physical statement that the beam would balance if you put a pivot there. Once you have seen this, the formula stops being something to memorise and becomes something you could rederive in a minute.
Prediction
Predict first
Every mass in a system is doubled, and none of them moves. What happens to the centre of mass?
Correct: It does not move
Why: Doubling every mass doubles the moment and doubles the total mass, and the quotient of the two is unchanged. The centre of mass depends only on the ratios between the masses, never on their absolute sizes, which is why the units of mass never matter in these problems.
Commit to an option before reading the explanation. The reason to ask is not that anyone finds this hard, but that the answer tells you something structural about the formula.
Doubling every mass doubles the moment and doubles the total mass, and the quotient is unchanged. So the centre of mass depends only on the ratios between the masses, never on their absolute sizes.
That is why you can drop the units of mass entirely in a centroid problem, and it is the same reason that the constant density of a lamina will cancel later in this lesson. Whenever a quantity appears once on top and once on the bottom, it is not really part of the answer.
Concept
A mass in the plane is a certain distance from each axis, so there are two moments. The distance to the x-axis is the second coordinate, and the distance to the y-axis is the first, so the labels look crossed over.
\[ M_x = \sum_{i=1}^{n} m_i y_i, \qquad M_y = \sum_{i=1}^{n} m_i x_i \]
\[ \bar{x} = \frac{M_y}{m}, \qquad \bar{y} = \frac{M_x}{m} \]
Read the subscript as the axis you are balancing about, not as the coordinate you are about to get. The moment about the x-axis delivers the second coordinate of the answer.
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 613 — Theorem 6.10, equations 6.16 and 6.17
In the plane a mass has two distances that matter, one to each axis, so there are two moments and the recipe runs twice. The mechanics are identical: a moment, a total mass, a division.
The part to slow down over is the labelling, which looks crossed. The moment about the horizontal axis uses the vertical coordinates, because the distance from a point to the horizontal axis is measured vertically. The moment about the vertical axis uses the horizontal ones, for the mirror-image reason.
Say the sentence out loud a couple of times: moment about an axis, arm measured to that axis. If you can reconstruct that sentence in an exam you will never have to remember which subscript goes where, because the geometry will tell you.
Notation
Annotate
On: \( M_x = \sum m_i y_i \qquad M_y = \sum m_i x_i \)
Step through the annotations and, as you read each one, point at the axis in question and at the arm that reaches it. This is a piece of notation that only makes sense geometrically.
The third annotation names the error to expect from yourself. Writing the moment about the horizontal axis with horizontal coordinates in it is natural, because the labels seem to match, and it is wrong. The label names what you are balancing about, not what you are multiplying by.
One more thing to notice. Both moments divide by the same total mass, so you compute that once and use it twice. If the two divisions give you a point far away from every mass, one of the two moments has a sign error in it.
Worked example
\[ m_1 = 2 \text{ at } (-1, 3), \quad m_2 = 6 \text{ at } (1,1), \quad m_3 = 4 \text{ at } (2,-2) \]
Add the masses
Why: Nothing signed.
\[ m = 2 + 6 + 4 = 12 \text{ kg} \]
Take the moment about the y-axis
Why: Use the first coordinates.
\[ M_y = 2(-1) + 6(1) + 4(2) = -2 + 6 + 8 = 12 \]
Take the moment about the x-axis
Why: Use the second coordinates.
\[ M_x = 2(3) + 6(1) + 4(-2) = 6 + 6 - 8 = 4 \]
Divide each by the total mass
Why: Two divisions, one point.
\[ \bar{x} = \frac{12}{12} = 1, \qquad \bar{y} = \frac{4}{12} = \frac13 \]
Figure (svg): Three dots in the coordinate plane with areas proportional to their masses: 2 kilograms at minus 1 and 3, 6 kilograms at 1 and 1, and 4 kilograms at 2 and minus 2. A yellow cross marks the centre of mass at 1 and one third.
Check the point against the picture
Why: Half the mass sits at the point one comma one, so the answer should be close to it, and the remaining 2 and 4 kilograms pull in nearly opposite vertical directions; a point just below it is right.
\[ (\bar{x}, \bar{y}) = \left(1, \tfrac13\right) \text{ m} \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, pp. 613-614 — Example 6.30
Run the same three-step recipe twice over. Add the masses once, then take the two moments, then do two divisions. Writing all three sums out term by term makes the signs visible.
In the second moment, the one about the horizontal axis, the contributions are plus six, plus six and minus eight. They nearly cancel, which is why the second coordinate comes out as only one third even though two of the masses sit three units and two units from that axis.
Compare your answer with the picture. The six-kilogram mass is half the total, so the balance point should be close to it, and it is: one unit across and a third of a unit up, just below that dot. The other two masses are on opposite sides vertically and pull against each other. If your answer had come out somewhere far from all three dots, a moment would have been mis-signed.
Worked example
\[ m_1 = 5 \text{ at } (-2,-3), \quad m_2 = 3 \text{ at } (2,3), \quad m_3 = 2 \text{ at } (-3,-2) \]
Add the masses
Why: Ten kilograms altogether.
\[ m = 5 + 3 + 2 = 10 \text{ kg} \]
Moment about the y-axis
Why: First coordinates.
\[ M_y = 5(-2) + 3(2) + 2(-3) = -10 + 6 - 6 = -10 \]
Moment about the x-axis
Why: Second coordinates.
\[ M_x = 5(-3) + 3(3) + 2(-2) = -15 + 9 - 4 = -10 \]
Divide
Why: Both moments happen to be the same here.
\[ \bar{x} = \frac{-10}{10} = -1, \qquad \bar{y} = \frac{-10}{10} = -1 \]
Check the sign
Why: Seven of the ten kilograms sit in the third quadrant, down and to the left, and the answer is down and to the left of the origin, inside the triangle formed by the three masses.
\[ (\bar{x}, \bar{y}) = (-1, -1) \text{ m} \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 614 — Checkpoint 6.30
Work this one before revealing the steps. The arithmetic is small and the point is to get the two moments into the right columns without looking back at the theorem.
The two moments happen to come out equal here, both minus ten, which is a coincidence of the numbers and not a general fact. Do not let it tempt you into computing only one of them.
For the check, count where the mass sits. Five kilograms are at a point down and to the left, two more are further down and left, and only three are up and to the right. Seven of the ten kilograms pull toward the third quadrant, so a balance point one unit down and one unit left of the origin is exactly what you should expect.
Section
Part 2
Concept
Now spread the mass out along a rod lying on an interval, with a density that changes from place to place. Cut the rod into thin pieces; a piece near a point has almost constant density, so its mass is the density there times its width.
\[ \Delta m_i \approx \rho(x_i^*)\,\Delta x, \qquad \Delta M_i \approx x_i^*\,\rho(x_i^*)\,\Delta x \]
Add the pieces and take the limit
Why: Both sums are Riemann sums.
\[ m = \int_a^b \rho(x)\,dx, \qquad M = \int_a^b x\,\rho(x)\,dx \]
Divide, exactly as before
Why: The definition of the balance point does not change.
\[ \bar{x} = \frac{\displaystyle\int_a^b x\,\rho(x)\,dx}{\displaystyle\int_a^b \rho(x)\,dx} \]
The denominator is the mass integral from Section 6.5. The numerator is the same integral with one extra factor of the position. Nothing new is being measured, only weighted.
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 624 — Exercises 260 to 269, the centre of mass of a rod
This is the same move you have made in every application in Chapter 6. Cut the object into pieces small enough that one number describes each piece, add up the pieces, and let the pieces shrink.
Here each piece is a short segment of the rod. Over a short enough stretch the density hardly changes, so the mass of the piece is the density at that spot times the width, and its moment is that mass times its position. Both sums are Riemann sums, so both become integrals.
Look carefully at the two integrands. The lower one is just the density, which is Section 6.5's mass integral. The upper one is the same thing with an extra factor of the position, the arm. If you can hold on to the idea that a moment is always mass multiplied by arm, you will never have to memorise either formula, here or in the harder plane-region case that follows.
Picture it
Figure (svg): The density of a rod plotted against position: a straight line rising from 1 at the left end to 4.2 at the right end, with eight shaded strips under it growing taller to the right. A dashed line marks the midpoint at 2 and a dot marks the balance point at 2.41.
The strips on the right carry more mass than the strips on the left, so the balance point is pulled to the right of the midpoint. Read the dashed line and the dot together: the whole effect of the varying density is that gap of about four tenths of a metre.
The curve is the density, not the rod. Read the height at a point as how much mass is packed in near that point, and read the area of each shaded strip as the mass that strip actually contains.
The strips on the right are taller, so they hold more mass, so the balance point is dragged rightward, past the geometric midpoint at two. Notice how small the effect is: the balance point only moves to about two point four one.
That is worth a moment's thought, because the density more than quadruples across the rod. The reason the balance point barely moves is that the light left-hand end still contributes real mass at a real distance, and a weighted average is stubborn. Extreme answers need extreme densities.
Worked example
A rod occupies the interval from 0 to 4 metres and its density in kilograms per metre is one plus eight tenths of the position. Find the balance point.
\[ \rho(x) = 1 + 0.8x, \qquad 0 \le x \le 4 \]
Integrate the density for the mass
Why: Term by term.
\[ m = \int_0^4 (1 + 0.8x)\,dx = \left[x + 0.4x^2\right]_0^4 = 4 + 6.4 = 10.4 \]
Write the moment integrand
Why: Position times density; expand the product.
\[ x\,\rho(x) = x(1 + 0.8x) = x + 0.8x^2 \]
Integrate it
Why: Power rule twice.
\[ M = \left[\frac{x^2}{2} + \frac{0.8x^3}{3}\right]_0^4 = 8 + \frac{51.2}{3} = \frac{376}{15} \]
Divide
Why: Moment over mass.
\[ \bar{x} = \frac{376/15}{10.4} = \frac{94}{39} \approx 2.410 \]
Check the uniform case
Why: Setting the slope to zero gives density 1, mass 4, moment 8 and a balance point of exactly 2, the midpoint — so the extra four tenths is entirely the work of the varying density.
\[ \rho \equiv 1: \quad \bar{x} = \frac{8}{4} = 2 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 624 — the pattern of Exercises 260 to 269
Two integrals and a division, and both integrals are ordinary power-rule work once you expand the products. The step worth being deliberate about is writing the moment integrand, because that is where the arm gets forgotten.
Multiply the density by the position before you integrate, not after. The product of the position with one plus eight tenths of the position is the position plus eight tenths of its square, and integrating that termwise is routine.
The check at the end is the pattern to copy in every problem of this type. Set the interesting feature of the density to zero, so that it becomes constant, and see whether the formula gives the obvious answer. A uniform rod must balance at its midpoint, and it does. That single check would have caught a dropped factor of the position, a wrong limit of integration, or a division performed the wrong way round.
Worked example
A rod of length L has density equal to the square of the position. Where does it balance?
\[ \rho(x) = x^2, \qquad 0 \le x \le L \]
Integrate the density
Why: Power rule.
\[ m = \int_0^L x^2\,dx = \frac{L^3}{3} \]
Integrate position times density
Why: One more power.
\[ M = \int_0^L x^3\,dx = \frac{L^4}{4} \]
Divide, and cancel
Why: The cube of L cancels.
\[ \bar{x} = \frac{L^4/4}{L^3/3} = \frac{3L}{4} \]
Check the position and the scaling
Why: Three quarters of the way along, well past the midpoint, which is right for a rod whose far end is much denser; and doubling L doubles the answer, as any length must.
\[ \frac{3L}{4} > \frac{L}{2} \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 624 — Exercise 261
Working with a general length instead of a number costs nothing here and tells you more. Both integrals are single powers, and the length cancels down to a clean fraction.
Three quarters of the way along is a long way past the midpoint, and it should be. With density equal to the square of the position, the far end of the rod is enormously heavier than the near end, so the balance point is pulled well out.
The second half of the check is about scaling, and it is a habit worth building. The answer is a length, so doubling the rod must double it, and it does. If an answer to a length question came out as a pure number, or as something with the length squared in it, you would know at once that a cancellation had gone wrong.
Worked example
A rod runs from 0 to the value of pi and its density is the sine of the position. Find the balance point.
\[ \rho(x) = \sin x, \qquad 0 \le x \le \pi \]
Look for a symmetry first
Why: Sine is symmetric about the quarter turn: the density at that point plus a shift equals the density at the same shift on the other side.
\[ \sin\left(\tfrac{\pi}{2} + t\right) = \sin\left(\tfrac{\pi}{2} - t\right) \]
Read off the balance point
Why: A rod whose density is a mirror image about its midpoint balances at that midpoint.
\[ \bar{x} = \frac{\pi}{2} \]
Confirm with the mass integral
Why: The sine integrates to minus cosine.
\[ m = \int_0^{\pi}\sin x\,dx = \left[-\cos x\right]_0^{\pi} = 2 \]
Confirm with the moment integral
Why: Integrate by parts, with u equal to x.
\[ M = \int_0^{\pi} x\sin x\,dx = \left[-x\cos x + \sin x\right]_0^{\pi} = \pi \]
Check the two routes agree
Why: The quotient of the two integrals reproduces the symmetry answer exactly, and the symmetry took one line against five.
\[ \bar{x} = \frac{\pi}{2} \approx 1.571 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 624 — Exercise 263
This slide is really about looking before you compute. The density is a sine curve over half a period, which is a perfect arch: the value a little way past the middle is exactly the value the same distance before the middle.
A density that is a mirror image about the midpoint means equal mass on either side at equal distances, so the rod must balance at the midpoint. That is one line of reasoning and no calculus at all.
The two integrals afterwards are there so you can see the shortcut confirmed rather than taken on trust. The moment integral needs integration by parts and the mass integral needs a sign kept straight, and together they take five lines to reproduce what symmetry gave in one. Remember the feeling, because symmetry is about to become a named theorem in Part 4 and it will save you far more than four lines there.
Error analysis
Annotate
On: \( M = \int_0^4 (1 + 0.8x)\,dx = 10.4 \;\Longrightarrow\; \bar{x} = \frac{10.4}{10.4} = 1 \)
Read the line before you open the annotations and try to name the mistake yourself. Everything about the calculation is correct except for what is being calculated.
The integrand is the density on its own, so the integral is the mass of the rod. Calling it the moment and then dividing by the mass is guaranteed to give one, whatever the rod, which is a strong signal that something structural is wrong rather than merely arithmetical.
The general protection is a units check on the integrand. Density times a length is a mass; a moment needs a mass times a length, so the integrand must carry one more factor with the units of length, and that factor is the position. Run that check before integrating, not after, because the integration is the expensive part.
Ranking
Put in order
Put the steps of a centre-of-mass calculation for a rod into the order you would actually do them.
Why: Symmetry is free and can finish the problem in one line, so it goes first. The two integrals can be done in either order, but the mass is usually easier and gives you a feel for the size of the answer. The division is last, and the range check costs nothing.
Drag the steps into order before checking. Only two of the orderings are genuinely debatable, and thinking about which is which is the point.
Symmetry goes first because it is free and can end the problem immediately, as the sine example just showed. The range check goes last because it is also free and catches the errors that the earlier steps let through.
The two integrals in the middle can honestly be done in either order. Most people do the mass first, because it is usually the simpler of the two and because knowing the total mass gives you a feel for how big the final quotient should be.
Sorting
Sort into buckets
Each rod occupies the interval from 0 to 1. Does its balance point sit left of the midpoint, at the midpoint, or right of it?
Sort these by reasoning about the shape of the density, not by integrating. You are practising the judgement that makes the range check useful.
Two of the six densities increase along the rod, which piles mass at the right-hand end and drags the balance point that way. Two decrease, which does the opposite. A constant density obviously balances at the midpoint.
The sixth one is the interesting case. One plus the square of the distance from the middle is not constant at all, but it is a mirror image about the midpoint, so the extra mass it adds near the left end is matched exactly by the extra mass near the right. That balances at the midpoint too. Symmetry, not constancy, is what puts the balance point in the middle.
Section
Part 3
Concept
lamina — A sheet of material thin enough to treat as a plane region. When its density is constant, its centre of mass depends only on the shape.
centroid — The geometric centre of a plane region: the centre of mass of a lamina of that shape with constant density.
The constant density multiplies the mass and both moments alike, so it cancels in every quotient. That is why a centroid is a fact about a shape, and why every calculation below may simply set the density to one.
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 614 — laminas, centroids and Theorem 6.11
Two words are being defined here and it is worth keeping them distinct. A lamina is a physical object, a thin sheet with a mass. A centroid is a property of a shape, with no physics in it at all.
They coincide exactly when the density is constant, which is the only case this section treats. Then the density multiplies the mass and both moments by the same factor, and that factor cancels in every quotient.
This is why you will see the density set to one in every worked example that follows, including in the book. It is not an approximation and it is not a special case; it is a genuine simplification, and the answer is the same for a sheet of lead as for a sheet of paper. If a problem ever asks you for the mass or the moment itself, rather than the centroid, put the density back in.
Concept
Figure (svg): The region under the curve y equals the square root of x from 0 to 4, with one vertical strip drawn between x equals 2 and 2.5. A dot marks the top of the strip and a second, brighter dot marks its midpoint at half the height.
Partition the interval and stand a thin rectangle on each piece, reaching from the axis up to the curve. Choose the sample point at the middle of each piece, so the rectangle's own centre sits directly above it.
\[ \text{strip: width } \Delta x, \; \text{height } f(x_i^*), \; \text{centre } \left(x_i^*, \tfrac{1}{2}f(x_i^*)\right) \]
A rectangle has its centre of mass where its diagonals cross, half way up and half way across. Treating the strip as a point mass placed there changes neither its mass nor its moments.
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 615 — Figures 6.65 and 6.66
The strip is the whole idea, so spend time on the picture. The region is cut into thin vertical rectangles, each standing from the horizontal axis up to the curve.
The choice of sample point matters here in a way it did not in Chapter 5. Take it at the middle of each little interval, so that the strip's own centre of mass sits directly above it. A rectangle balances where its diagonals cross, which is half way across and half way up.
Now use the fact you proved back in Part 1: replacing an object by a point mass at its own centre of mass changes none of its moments. So each strip can be treated as a single dot of mass, sitting at the position you chose and at half the curve's height. That reduction from a strip to a dot is what lets you reuse the point-mass formulas, and the half in the height is about to become the factor that everyone forgets.
Picture it
Figure (svg): Two copies of the same region and the same strip. On the left a horizontal arrow runs from the y-axis to the strip, labelled arm equals x. On the right a vertical arrow runs from the x-axis to the strip's midpoint, labelled arm equals f of x over 2.
Every part of the strip stands at the same horizontal distance from the vertical axis, so that arm is just the position. Vertically the strip is spread from the axis up to the curve, and it acts at its middle, so that arm is half the height. One picture, two different answers.
The two panels are the same region and the same strip. Only the arrow changes, and the arrow is the arm.
On the left the strip is being balanced about the vertical axis. Every scrap of the strip stands at the same horizontal distance from that axis, so the arm is simply the position, and the whole strip acts at once.
On the right it is being balanced about the horizontal axis, and now the strip is not all at one distance: it stretches from the axis up to the curve. It acts at its own middle, half way up, so the arm is half the height. If you can see why the two pictures give different arms, you will never again have to remember which formula gets the factor of one half.
Concept
Add the strip masses
Why: Density times area, summed and taken to the limit.
\[ m = \lim_{n\to\infty}\sum_{i=1}^{n} \rho f(x_i^*)\,\Delta x = \rho\int_a^b f(x)\,dx \]
Moment about the x-axis: mass times the vertical arm
Why: The arm is half the height, so the height appears twice.
\[ M_x = \lim_{n\to\infty}\sum_{i=1}^{n} \rho\,\frac{[f(x_i^*)]^2}{2}\,\Delta x = \rho\int_a^b \frac{[f(x)]^2}{2}\,dx \]
Moment about the y-axis: mass times the horizontal arm
Why: The arm is the position itself.
\[ M_y = \lim_{n\to\infty}\sum_{i=1}^{n} \rho\,x_i^*f(x_i^*)\,\Delta x = \rho\int_a^b x\,f(x)\,dx \]
Divide, and watch the density cancel
Why: Each quotient has one factor of the density above and one below.
\[ \bar{x} = \frac{M_y}{m}, \qquad \bar{y} = \frac{M_x}{m} \]
Check the formulas on a rectangle
Why: For a rectangle of height h on the interval from 0 to w the mass is hw, the y-moment is h w squared over 2 and the x-moment is h squared w over 2, giving the centre of the rectangle, as symmetry demands.
\[ \bar{x} = \frac{hw^2/2}{hw} = \frac{w}{2}, \qquad \bar{y} = \frac{h^2w/2}{hw} = \frac{h}{2} \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, pp. 615-616 — Theorem 6.12, equations 6.18 to 6.20
Follow the derivation with the strip picture in mind, because every line is mass times arm and nothing else. The mass of a strip is density times height times width. Multiply that by the arm to the vertical axis, which is the position, and you get the first moment; multiply it instead by the arm to the horizontal axis, which is half the height, and you get the second.
That second multiplication is where the square appears. One factor of the height came from the strip's area, the other from its arm, and the arm also carries the divisor of two. If you ever find yourself asking why the height is squared, the answer is that it appears for two completely different reasons.
The check at the end is small and worth doing. Run the three formulas on a plain rectangle, where you already know the answer must be the middle by symmetry, and watch them deliver it. A formula that gets a rectangle wrong will get everything wrong.
Notation
Annotate
On: \( M_x = \rho\int_a^b \frac{[f(x)]^2}{2}\,dx, \qquad M_y = \rho\int_a^b x\,f(x)\,dx \)
These are the two formulas you will use most in this lesson, so read all four annotations rather than skimming to the examples.
The second annotation is the one to take to heart. The factor of one half is not cosmetic and it is not a constant that came out of an integration. It is the arm itself, and leaving it out makes the second coordinate of your centroid exactly twice too big, which is a mistake that produces plausible-looking wrong answers rather than obvious ones.
The fourth annotation is your permission slip. Because the density sits in all three integrals, it cancels in both quotients, so you may set it to one whenever the question asks only for a centroid. Just be sure it really is constant; a density that varies across the region is a multivariable problem and these formulas do not cover it.
Worked example
Find the centroid of the region under the square-root curve on the interval from 0 to 4. Take the density to be one.
\[ f(x) = \sqrt{x}, \qquad 0 \le x \le 4 \]
Integrate for the mass
Why: Raise the exponent to three halves and divide by it.
\[ m = \int_0^4 \sqrt{x}\,dx = \left[\tfrac23 x^{3/2}\right]_0^4 = \tfrac23(8) = \frac{16}{3} \]
Build the x-moment integrand
Why: The square of the square root is the position itself.
\[ \frac{[f(x)]^2}{2} = \frac{x}{2} \]
Integrate it
Why: A single power.
\[ M_x = \int_0^4 \frac{x}{2}\,dx = \left[\tfrac14x^2\right]_0^4 = 4 \]
Integrate for the y-moment
Why: Position times the square root is the position to the three halves.
\[ M_y = \int_0^4 x^{3/2}\,dx = \left[\tfrac25 x^{5/2}\right]_0^4 = \tfrac25(32) = \frac{64}{5} \]
Divide twice
Why: Each moment over the same mass.
\[ \bar{x} = \frac{64/5}{16/3} = \frac{12}{5}, \qquad \bar{y} = \frac{4}{16/3} = \frac34 \]
Figure (svg): The region under the square root curve from 0 to 4, shaded, with a dot at 2.4 and 0.75 marking its centroid and dashed lines dropping to each axis.
Check the point sits inside the region
Why: At the position 12 over 5, or 2.4, the curve has height about 1.55, and three quarters is below that, so the centroid is inside the shaded region and right of the midpoint 2, as the widening shape demands.
\[ \sqrt{2.4} \approx 1.549 > 0.75 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, pp. 616-617 — Example 6.31
This is the model solution for a region above the axis, and every line in it earns its place. Three integrals, then two divisions, and nothing else.
Notice the small simplification in the second step: squaring the square root gives the position back, so the integrand of the harder moment turns out to be the easiest one of the three. That will not always happen, but it is always worth simplifying the squared height before integrating rather than after.
The check at the end does two jobs at once. It confirms the point is inside the region, by comparing the centroid's height with the curve's height directly above it, and it confirms that the point is right of the midpoint, which is what a region that widens to the right demands. Both are free, and between them they would catch almost any arithmetic slip you could make here.
Fill the middle
A student wrote the moment of this region about the x-axis without its arm. Restore the missing divisor.
Fill in the blanks
M_x = \int_0^4 \frac2___}\,dx
Why: The strip's arm to the x-axis is half its height, so one of the two factors of the height carries a divisor of 2. Leave it out and the second coordinate of the centroid comes out exactly twice too large.
Fill in the blank before revealing. The answer is short, but say to yourself why it is there rather than just typing it.
The strip's arm to the horizontal axis is half its height, and that half is the only source of the divisor. The two in the denominator is not an integration constant and it does not come from the power rule; it is a distance.
If you forget it, nothing in the algebra afterwards will complain. Your integral will evaluate cleanly, your division will work, and you will get an answer that is silently twice too large. That is why this one is worth a slide of its own.
Worked example
\[ f(x) = x^2, \qquad 0 \le x \le 2 \]
Integrate for the mass
Why: Power rule.
\[ m = \int_0^2 x^2\,dx = \left[\tfrac13x^3\right]_0^2 = \frac83 \]
Integrate for the y-moment
Why: Position times the height is the cube.
\[ M_y = \int_0^2 x^3\,dx = \left[\tfrac14x^4\right]_0^2 = 4 \]
Integrate for the x-moment
Why: Square the height, then halve it.
\[ M_x = \int_0^2 \frac{x^4}{2}\,dx = \left[\tfrac1{10}x^5\right]_0^2 = \frac{16}{5} \]
Divide twice
Why: Same mass both times.
\[ \bar{x} = \frac{4}{8/3} = \frac32, \qquad \bar{y} = \frac{16/5}{8/3} = \frac65 \]
Check both coordinates for plausibility
Why: At the position 1.5 the curve is at 2.25 and the centroid height 1.2 is below it, so the point is inside; and the centroid is right of the midpoint 1, which is right for a region that is far wider near the right-hand end.
\[ (\bar{x}, \bar{y}) = (1.5,\; 1.2) \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 617 — Checkpoint 6.31
Do this one on paper before revealing. It is the same shape of calculation as the example, with an easier function, so it is a fair test of whether the three integrands have stuck.
The one place to be careful is the moment about the horizontal axis. Square the height first, giving the fourth power, and only then halve it. Squaring after halving would introduce an extra factor of four.
For the check, look at both coordinates against the picture. The centroid is at one and a half horizontally, right of the midpoint at one, because the region is much wider near its right-hand end. And its height of one point two is well below the curve's height of two and a quarter directly above it, so the point lies inside the region. Both facts should feel obvious from the shape, which is what makes them useful checks.
Trap
For the parabola of the last slide:
\[ M_x = \int_0^2 x^4\,dx = \frac{32}{5} \]
\[ \bar{y} = \frac{32/5}{8/3} = \frac{12}{5} \]
Wrong. The arm was taken as the whole height.
The curve only reaches a height of 4 at the very right-hand end, and almost all of the region is far below that, so a balance height of 2.4 is impossible. The correct value is 1.2, exactly half as large.
\[ \bar{y} = \frac{16/5}{8/3} = \frac65 \]
A quick guard: the second coordinate of a centroid can never exceed the greatest height of the region, and for a region that sags toward the axis it will be well under half of it.
This is the same omission the cloze slide warned about, now shown doing its damage. The answer produced is two point four, exactly double the correct one point two.
The reason it is dangerous is that nothing looks wrong. The integral is easy, the division works, and the number is not absurd. Nothing downstream of the mistake will complain, so you have to catch it by thinking about the shape instead.
Try the guard on this region and see how fast it works. The parabola reaches four only at the single point at the far right edge, and everywhere else it is lower, most of it much lower, so a balance height of two point four is out of the question before you check a single line of algebra. Ten seconds of looking at the picture beats five minutes of re-checking the integration.
Matching
Match the pairs
Why: The height alone gives the area, and with a density it gives the mass. Multiplying by the horizontal arm x gives the moment about the y-axis; squaring the height and halving it builds the vertical arm into the moment about the x-axis. Only a quotient is a coordinate.
Match these before revealing. The exercise is not about arithmetic at all; it is about recognising which integrand carries an arm and which does not.
The height on its own carries no arm, so its integral is an area, or a mass once a density is present. Multiplying by the position builds in the horizontal arm. Squaring the height and halving it builds in the vertical arm, for the two separate reasons you saw in the derivation.
The last pairing is the one to read twice. Only a quotient is a coordinate. An integral by itself, however carefully built, is a moment or a mass, and it is not a point until you have divided.
Tweak it
Parameter explorer
The curve is y = xᵖ on the interval from 0 to 1. Its centroid sits at ((p+1)/(p+2), (p+1)/(2(2p+1))). Slide p and predict both coordinates before you compute them: which way does each move as the region flattens against the axis?
\[ y = x^{{p}}, \quad 0 \le x \le 1 \]
Start with the exponent at two and move it slowly upward. Watch the region flatten against the horizontal axis and then swing up sharply near the right-hand end.
Predict each coordinate before you compute it. As the exponent grows the area concentrates near the right edge, so the first coordinate should climb toward one, and the formula agrees: the ratio of the exponent plus one to the exponent plus two heads to one. The second coordinate is less obvious. It falls, and it settles at a quarter rather than at zero, because the thin sliver that survives near the right edge is tall.
Now run the slider the other way, down toward small exponents. The region fattens upward, the first coordinate slides back toward a half and the second one rises. At an exponent of exactly one the region is a triangle and the centroid is at two thirds and one third, which you can check against the triangle result later in this part.
Concept
Now let the region be bounded above by one curve and below by another. The strip runs between them, so its height is the difference, and its own centre is at the average of the two heights.
\[ \text{height } f(x) - g(x), \qquad \text{centre at } \tfrac12\left[f(x) + g(x)\right] \]
Build the x-moment from mass times arm
Why: Multiply the strip's area by the height of its centre.
\[ \left[f(x)-g(x)\right]\cdot\tfrac12\left[f(x)+g(x)\right] = \tfrac12\left([f(x)]^2 - [g(x)]^2\right) \]
Collect all three integrals
Why: The y-moment only needs the height, so nothing changes there.
\[ m = \rho\int_a^b (f-g)\,dx, \quad M_x = \rho\int_a^b \tfrac12(f^2-g^2)\,dx, \quad M_y = \rho\int_a^b x(f-g)\,dx \]
Check that the earlier case is included
Why: Setting the lower curve to zero recovers the three integrals of Theorem 6.12 exactly.
\[ g \equiv 0: \quad \tfrac12(f^2 - 0) = \tfrac{f^2}{2} \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 618 — Theorem 6.13, equations 6.21 to 6.23
Only one thing changes when a second curve becomes the lower boundary, and it is the arm to the horizontal axis. The strip no longer starts on the axis, so its middle is no longer at half its height.
The strip now runs from the lower curve to the upper one, so its centre sits at the average of the two heights. Multiplying that arm by the strip's area, which is the difference of the heights, produces a difference of two squares, halved. The product of a sum and a difference is a difference of squares, which is why the formula looks the way it does.
The other two integrals need only the height, so they simply get the difference in place of the single function. And the check at the end matters: setting the lower curve to zero must give you back the earlier formulas, and it does. A new formula that does not reduce to the old one in the old case is a new formula with a mistake in it.
Worked example
Find the centroid of the region between the two curves below.
\[ f(x) = 1 - x^2, \qquad g(x) = x - 1 \]
Find where they meet
Why: Set them equal and factor.
\[ 1 - x^2 = x - 1 \;\Longrightarrow\; x^2 + x - 2 = 0 \;\Longrightarrow\; x = -2, \; 1 \]
Integrate the height for the mass
Why: The height is two minus the position squared minus the position.
\[ m = \int_{-2}^{1}(2 - x^2 - x)\,dx = \left[2x - \tfrac13x^3 - \tfrac12x^2\right]_{-2}^{1} = \frac92 \]
Build the x-moment integrand
Why: Difference of the two squares, halved.
\[ \tfrac12\left[(1-x^2)^2 - (x-1)^2\right] = \tfrac12\left(x^4 - 3x^2 + 2x\right) \]
Integrate it
Why: Power rule term by term.
\[ M_x = \tfrac12\left[\tfrac{x^5}{5} - x^3 + x^2\right]_{-2}^{1} = -\frac{27}{10} \]
Integrate for the y-moment
Why: Position times the height.
\[ M_y = \int_{-2}^{1}(2x - x^3 - x^2)\,dx = \left[x^2 - \tfrac{x^4}{4} - \tfrac{x^3}{3}\right]_{-2}^{1} = -\frac94 \]
Divide twice
Why: Both moments are negative, so both coordinates will be.
\[ \bar{x} = \frac{-9/4}{9/2} = -\frac12, \qquad \bar{y} = \frac{-27/10}{9/2} = -\frac35 \]
Figure (svg): The parabola y equals 1 minus x squared and the line y equals x minus 1 crossing at minus 2 and minus 3 and at 1 and 0, with the lens-shaped region between them shaded and a dot at minus one half, minus three fifths.
Check against the crossing points
Why: The region runs from minus 2 to 1, and minus one half is inside; vertically it runs from minus 3 to 1, and minus three fifths is inside and below the midpoint, which is right because the region is much deeper on the left.
\[ -2 < -\tfrac12 < 1, \qquad -3 < -\tfrac35 < 1 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, pp. 618-619 — Example 6.32
The first step is the one that is easy to skip. You cannot integrate until you know the limits, and here they come from setting the two functions equal and solving a quadratic.
Once the limits are in hand the rest is routine, with one place to be slow: expanding the difference of the two squares. Write out both squares in full before subtracting, because the cross terms are where sign errors breed.
Both moments come out negative here, so both coordinates of the centroid are negative, and the picture shows why. The region hangs mostly below the horizontal axis and mostly left of the vertical one. For the check, read the crossing points off the figure and confirm the centroid lies between them in both directions. It does, and it sits low, because the region is far deeper on its left-hand side than on its right.
Worked example
\[ f(x) = 6 - x^2, \qquad g(x) = 3 - 2x \]
Find the crossing points
Why: Set them equal and factor.
\[ 6 - x^2 = 3 - 2x \;\Longrightarrow\; x^2 - 2x - 3 = 0 \;\Longrightarrow\; x = -1, \; 3 \]
Integrate the height
Why: Three plus twice the position minus the position squared.
\[ m = \int_{-1}^{3}(3 + 2x - x^2)\,dx = \left[3x + x^2 - \tfrac13x^3\right]_{-1}^{3} = \frac{32}{3} \]
Integrate for the y-moment
Why: Position times that height.
\[ M_y = \int_{-1}^{3}(3x + 2x^2 - x^3)\,dx = \left[\tfrac32x^2 + \tfrac23x^3 - \tfrac14x^4\right]_{-1}^{3} = \frac{32}{3} \]
Integrate for the x-moment
Why: Difference of squares, halved.
\[ M_x = \tfrac12\int_{-1}^{3}\left[(6-x^2)^2 - (3-2x)^2\right]dx = \frac{416}{15} \]
Divide twice
Why: The first quotient is exactly one.
\[ \bar{x} = \frac{32/3}{32/3} = 1, \qquad \bar{y} = \frac{416/15}{32/3} = \frac{13}{5} \]
Check the first coordinate against symmetry
Why: The two crossing points are minus 1 and 3, whose midpoint is 1; the height three plus twice the position minus its square is a downward parabola symmetric about that same line, so a first coordinate of exactly 1 is forced.
\[ \frac{-1 + 3}{2} = 1 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 619 — Checkpoint 6.32
Try this one yourself first. It has the same structure as the example but the numbers are friendlier, and the final check is more interesting than the calculation.
The two crossing points are at minus one and three, and their midpoint is one. Now look at the height of the region, which is three plus twice the position minus its square. That is a downward parabola, and it is symmetric about the very same line.
So the region really is symmetric about a vertical line, and the first coordinate of the centroid is forced to be one before any integral is done. The calculation confirms it, with the two integrals coming out numerically equal, which is a coincidence of the numbers. The lesson here is that symmetry can hide inside a region built from two unsymmetrical-looking curves, and the next part of the lesson makes that into a theorem.
Worked example
A triangular plate has vertices at the origin, at the point a on the horizontal axis, and at the point b on the vertical axis. Find its centroid.
\[ f(x) = b\left(1 - \frac{x}{a}\right), \qquad 0 \le x \le a \]
Integrate the height
Why: Half of base times height, as expected.
\[ m = \int_0^a b\left(1 - \frac{x}{a}\right)dx = b\left[x - \frac{x^2}{2a}\right]_0^a = \frac{ab}{2} \]
Integrate for the y-moment
Why: Position times the height.
\[ M_y = b\int_0^a\left(x - \frac{x^2}{a}\right)dx = b\left(\frac{a^2}{2} - \frac{a^2}{3}\right) = \frac{a^2b}{6} \]
Integrate for the x-moment
Why: Square the height and halve it.
\[ M_x = \frac{b^2}{2}\int_0^a\left(1 - \frac{x}{a}\right)^2dx = \frac{b^2}{2}\cdot\frac{a}{3} = \frac{ab^2}{6} \]
Divide twice
Why: Both quotients lose the factor of a b.
\[ \bar{x} = \frac{a^2b/6}{ab/2} = \frac{a}{3}, \qquad \bar{y} = \frac{ab^2/6}{ab/2} = \frac{b}{3} \]
Figure (svg): A right triangle with vertices at the origin, at 6 and 0, and at 0 and 4, shaded, with its three medians drawn as dashed lines meeting at a dot at 2 and one and a third.
Check against the medians
Why: A third of the way along each leg is exactly where the three medians of a triangle cross, the centroid every geometry course names, so the integrals reproduce a result known without calculus.
\[ \left(\tfrac{a}{3}, \tfrac{b}{3}\right) = \tfrac13\left[(0,0) + (a,0) + (0,b)\right] \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 624 — Exercise 271
Doing this one with letters rather than numbers is worth the small extra effort, because the answer is a fact you will use again and again.
The height of the region is a straight line falling from the vertical intercept to zero at the horizontal one, and the three integrals are all polynomial. Watch the factor of the base times the height cancel out of both quotients, leaving nothing but a third of each side.
The check is the reason to do this at all. A third of the way along each leg is exactly where the three medians of a triangle meet, and that is the centroid you may have met in geometry years before you met an integral. Calculus and classical geometry agreeing here is not a coincidence: the median construction is really a symmetry argument in disguise. Keep the result, because the very last worked example of this lesson uses it to produce the volume of a cone in three lines.
Comparison
Comparison matrix
| quantity | integrand | what it gives after dividing |
|---|---|---|
| mass m | f(x) | nothing; it is the divisor |
| moment M_y | x f(x) | x-bar |
| moment M_x | f(x)^2 / 2 | y-bar |
| mass, two curves | f(x) − g(x) | nothing; it is the divisor |
| moment M_x, two curves | (f(x)^2 − g(x)^2) / 2 | y-bar |
Fill in every blank before checking any of them. The table is the whole of Part 3 compressed into five rows, and being able to reconstruct it is a good test of whether the arms have stuck.
The rows to think hardest about are the ones with the squares. Both of them describe a moment about the horizontal axis, and both carry a divisor of two, because in both cases the strip acts at its own middle rather than at its top.
Notice also what the right-hand column says about the mass rows: they give you nothing directly. The mass is never a coordinate, it is only ever the divisor. If you find a table row where a single integral is being reported as an answer, something has gone wrong.
Prediction
Predict first
Suppose you sliced the same region into horizontal strips instead of vertical ones. What would happen to the centroid you compute?
Correct: It would come out the same, but the roles of the two formulas swap
Why: The centroid is a property of the region, not of how you chose to cut it, so the point is unchanged. What changes is which formula carries the factor of one half: with horizontal strips the arm to the y-axis is half the strip's length, and the arm to the x-axis is just the height. Slicing the other way is often much easier for a region described by x as a function of y.
Commit to an option before reading the explanation. The question is really asking whether the centroid is a property of the region or a property of your method.
It is a property of the region, so the point cannot move. What changes is the bookkeeping: with horizontal strips, the arm to the vertical axis is now half the strip's length, and the arm to the horizontal axis is just the height of the strip. The factor of one half changes sides.
This is genuinely useful, not just a curiosity. When a region is described naturally by giving the horizontal extent as a function of the vertical position, horizontal strips need no rearranging and vertical ones would force you to split the region into pieces. Choosing the slicing direction is a real decision, exactly as it was for volumes in Sections 6.2 and 6.3.
Section
Part 4
Concept
the symmetry principle — If a region is symmetric about a line, its centroid lies on that line.
This is stated in the book without proof, but the reason is short: reflecting the region in its axis of symmetry leaves the region, and therefore the balance point, unchanged, and the only points fixed by a reflection are the ones on the mirror line.
Used in the right place it removes half the work. Used in the wrong place it invents an answer, so the first job is always to check that the symmetry is real.
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 614 — Theorem 6.11
The principle is stated in the book without proof, which is fair enough, because the reason is almost too short to call a proof. Reflecting a region in its own axis of symmetry gives back the same region, so it must give back the same balance point, and the only points a reflection leaves alone are the ones on the mirror line.
Used well it is the biggest labour saver in this section. One of your two integrals simply disappears, and with it half the chances of an arithmetic error.
Used carelessly it invents answers, which is why the next few slides spend so much time on checking that the symmetry is real. The test is always physical rather than algebraic: fold the picture along the candidate line and see whether the two halves land on each other. A symmetric formula is not the same thing as a symmetric region.
Worked example
Make the argument concrete for a region symmetric about the vertical axis, so its height is an even function.
\[ f(-x) = f(x) \quad \text{on} \quad -a \le x \le a \]
Look at the y-moment integrand
Why: Position times height, with the position replaced by its negative.
\[ (-x)f(-x) = -x\,f(x) \]
Name what that means
Why: The integrand is an odd function on an interval symmetric about the origin.
\[ h(-x) = -h(x), \quad h(x) = x\,f(x) \]
Integrate an odd function over a symmetric interval
Why: The two halves cancel exactly.
\[ M_y = \int_{-a}^{a} x\,f(x)\,dx = 0 \]
Divide
Why: A zero moment gives a zero coordinate.
\[ \bar{x} = \frac{M_y}{m} = 0 \]
Figure (svg): The curve y equals x times the quantity 4 minus x squared between minus 2 and 2. The lobe left of the origin is below the axis and shaded red; the lobe right of the origin is above it and shaded green; the two have equal area.
Check on the parabola four minus the square
Why: The two lobes of the integrand shown have areas minus 4 and plus 4, which sum to zero, so the first coordinate of that region's centroid is zero without any integration at all.
\[ \int_{-2}^{2} x(4 - x^2)\,dx = -4 + 4 = 0 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 620 — the symmetry principle applied
This is the same principle again, now in the language of integrals, and seeing both versions is what makes it stick. Start from a region symmetric about the vertical axis, which means its height is unchanged when you replace the position by its negative.
Now look at what happens to the integrand of the moment about that axis. The height stays put, but the arm flips sign, so the whole integrand flips sign. An integrand with that property, integrated over an interval symmetric about the origin, gives exactly zero.
The figure is that cancellation drawn. The two lobes have the same shape and opposite signs, so their areas are minus four and plus four, and the moment vanishes. Notice that you are not proving the moment is small, or approximately zero; it is exactly zero, for every symmetric region, which is why you can write the coordinate down with complete confidence.
Worked example
Find the centroid of the region under the downward parabola and above the horizontal axis.
\[ f(x) = 4 - x^2, \qquad -2 \le x \le 2 \]
Apply the symmetry principle
Why: The region is its own mirror image in the vertical axis, so the centroid lies on that axis.
\[ \bar{x} = 0 \]
Integrate for the mass
Why: Only this and one moment are now needed.
\[ m = \int_{-2}^{2}(4 - x^2)\,dx = \left[4x - \tfrac13x^3\right]_{-2}^{2} = \frac{32}{3} \]
Expand the squared height
Why: Before integrating, multiply it out.
\[ \tfrac12(4-x^2)^2 = \tfrac12\left(16 - 8x^2 + x^4\right) \]
Integrate for the x-moment
Why: Three power-rule terms.
\[ M_x = \tfrac12\left[16x - \tfrac83x^3 + \tfrac{x^5}{5}\right]_{-2}^{2} = \frac{256}{15} \]
Divide
Why: Moment over mass.
\[ \bar{y} = \frac{256/15}{32/3} = \frac{8}{5} \]
Figure (svg): The region under the parabola y equals 4 minus x squared above the x-axis, shaded, with a dashed vertical axis of symmetry on the y-axis and a dot at 0 and 1.6 marking the centroid.
Check the height against the region
Why: The region reaches 4 at its peak, and the centroid height 1.6 is well below half of that, which is right: the region is widest near the bottom, so most of its area is low down.
\[ 0 < \tfrac85 < 4, \qquad \tfrac85 < \tfrac42 = 2 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 620 — Example 6.33
Start by looking rather than integrating. The region is a dome over the horizontal axis, unchanged by reflection in the vertical one, so the first coordinate of the centroid is zero and one of the two moment integrals never has to be written down.
For the one that remains, expand the squared height before integrating. Trying to integrate the square directly is where people lose a term.
The check at the end is about proportion rather than arithmetic. The region is four units tall at its peak, so at first glance a balance height of one point six may look low. It is not: the region is at its widest right down at the axis and it narrows to a point at the top, so most of the area is low and the balance point should be well below half the height. Make that comparison a reflex, because it catches the missing factor of one half instantly.
Worked example
\[ f(x) = 1 - x^2, \qquad -1 \le x \le 1 \]
Use symmetry for the first coordinate
Why: The region is symmetric about the vertical axis.
\[ \bar{x} = 0 \]
Integrate for the mass
Why: Two power-rule terms.
\[ m = \int_{-1}^{1}(1 - x^2)\,dx = \left[x - \tfrac13x^3\right]_{-1}^{1} = \frac43 \]
Integrate for the x-moment
Why: Expand the square first, then integrate.
\[ M_x = \tfrac12\int_{-1}^{1}\left(1 - 2x^2 + x^4\right)dx = \tfrac12\left[x - \tfrac23x^3 + \tfrac{x^5}{5}\right]_{-1}^{1} = \frac{8}{15} \]
Divide
Why: Moment over mass.
\[ \bar{y} = \frac{8/15}{4/3} = \frac25 \]
Check against the previous example by scaling
Why: This region is the previous one squashed to a quarter of its height, because one minus a square is a quarter of four minus the square of twice the position; a vertical squash by four divides every height, including the centroid's, by four.
\[ 1 - x^2 = \tfrac14\left[4 - (2x)^2\right] \;\Longrightarrow\; \bar{y} = \tfrac14\cdot\tfrac85 = \tfrac25 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 620 — Checkpoint 6.33
Do this before revealing. It repeats the example on a smaller parabola, so it is quick, and the interesting part is the check rather than the integrals.
Symmetry again gives the first coordinate for nothing. For the second, expand the squared height into three terms and integrate each one.
Now the check. This region is not simply a shrunken copy of the previous one in both directions: it is squashed to a quarter of the height and to half the width. Only the vertical squashing affects a vertical coordinate, and a quarter of eight fifths is two fifths, which is the answer. Reasoning like that, about how a quantity behaves under scaling, is worth practising; it is often the fastest way to check an answer you have no independent way of computing.
Error analysis
Annotate
On: \( f(x) = x^2 \text{ on } [0,2]: \; \text{symmetric, so } \bar{x} = 1 \)
Look at the claim before opening the annotations. The function really is symmetric about the vertical axis, and the conclusion is still wrong.
The gap is between the function and the region. A symmetric function does not make a symmetric region unless the interval is symmetric too, and this interval runs from zero to two, entirely on one side.
Do the fold test on the picture in your head. Folding this region along the vertical line through one would land the tall right-hand end on top of the thin left-hand end, and they do not match. So there is no symmetry to use, and the true first coordinate is three halves, pulled right by the much wider right-hand part. Whenever you are about to invoke symmetry, fold first.
Sorting
Sort into buckets
For each region, how much of the centroid can be read off from symmetry alone?
Sort these by asking, for each shape, how many mirror lines it has. That count is exactly what decides the bucket.
Two perpendicular mirror lines pin the centroid completely, because it must lie on both and two crossing lines meet in one point. That is the disc and the rectangle, and it is why nobody ever integrates to find the centre of a rectangle.
One mirror line gives you one coordinate and leaves the other to an integral. Watch the last item carefully, because the wedge between a line and its negative is symmetric about the horizontal axis, not the vertical one, so it is the second coordinate that comes out free. And the region between a parabola and a line through the origin has no mirror line at all, however symmetric the two curves look individually.
Estimation
\[ f(x) = 4 - x^2 \quad \text{versus} \quad f(x) = \sqrt{4 - x^2} \]
Predict first
Both regions sit on the interval from minus 2 to 2 and both are symmetric about the vertical axis. Which has its centroid higher above the horizontal axis: the parabola region of Example 6.33, or the half-disc of radius 2?
Correct: the parabola region
Why: The parabola region's centroid is at 1.6; the half-disc's is at four times the radius over three pi, which for radius 2 is about 0.849. But that ranking is mostly about size: the parabola region is 4 units tall and the half-disc only 2. As a fraction of its own height the parabola region sits at 0.40 and the half-disc at 0.42, so the half-disc is very slightly the top-heavier shape. When two regions differ in height, compare the ratio and not the raw number.
Commit to an answer before reading the explanation, because the point of this slide is to catch a habit of thought rather than to test a formula.
The raw numbers say the parabola region wins easily, one point six against about zero point eight five. But that comparison is unfair, because the parabola region is four units tall and the half-disc only two, so of course its centroid is higher off the ground.
Divide each centroid by the height of its own region and the two come out at forty and forty-two percent, almost identical, with the half-disc very slightly the top-heavier. That is the honest comparison. Whenever you are comparing centroids of differently sized objects, express them as a fraction of the object's own extent, or you will be measuring size rather than shape.
Section
Part 5
Concept
Figure (svg): A vertical axis on the left, a shaded region to its right at distance x-bar, a dot at the region's centroid, and a dashed ellipse showing the circular path that dot sweeps as the region turns once about the axis.
Theorem of Pappus for volume — If a region is revolved about a line in its plane that does not meet it, the volume swept out is the area of the region multiplied by the distance its centroid travels.
\[ V = A\cdot d, \qquad d = 2\pi\bar{x} \;\text{ for revolution about the }y\text{-axis} \]
The line must miss the region entirely. If it cut through, part of the region would sweep backwards over ground another part had already covered, and the volume would be counted twice.
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 623 — Theorem 6.14
This theorem is the pay-off of the whole section, and the statement is short enough to memorise: the volume swept is the area times the distance the centroid travels.
Look at the picture and notice what is not in the statement. The shape of the region does not appear, nor the way its area is distributed, nor anything about how far the nearest or the furthest point of the region stands from the axis. One area and one distance are enough.
The condition at the end is not a technicality. If the axis passed through the region, the two halves would sweep over the same ground and the volume would be double-counted, so the theorem would simply be false. Whenever you reach for it, check first that the whole region lies on one side of the line.
Concept
Take the region between two curves, revolved about the vertical axis, and compare the centroid formula with the shell formula from Section 6.3.
Write the distance the centroid travels
Why: A full circle of radius equal to the first coordinate of the centroid.
\[ d = 2\pi\bar{x} = 2\pi\,\frac{M_y}{m} \]
Put in the two integrals
Why: The density appears above and below.
\[ d = 2\pi\,\frac{\rho\int_a^b x[f(x)-g(x)]\,dx}{\rho\int_a^b [f(x)-g(x)]\,dx} \]
Recognise the denominator as the area
Why: The density cancels and what is left below is the area.
\[ d = \frac{2\pi\int_a^b x[f(x)-g(x)]\,dx}{A} \]
Multiply through by the area
Why: Clear the fraction.
\[ d\cdot A = 2\pi\int_a^b x\left[f(x)-g(x)\right]dx \]
Compare with the shell formula
Why: The right-hand side is exactly the volume by cylindrical shells.
\[ V = 2\pi\int_a^b x\left[f(x)-g(x)\right]dx \]
Check what has been shown
Why: The two right-hand sides are the same integral, so the volume really is the area times the centroid's journey, with no approximation anywhere.
\[ V = d\cdot A \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 623 — the proof of Theorem 6.14
The proof is short because most of the work was done in Section 6.3, when you derived the shell formula. All that remains is to recognise that you have already met the same integral twice.
Follow the algebra with that in mind. Write the centroid's journey as a full circle of radius equal to the first coordinate of the centroid, substitute the two integrals that define it, and watch the density cancel. Multiplying through by the area clears the fraction and leaves a single integral.
That integral is the shell formula, exactly. So the two expressions are not approximately equal or equal up to a constant: they are the same object written two ways. That is why the theorem is exact, and why it feels like cheating. All it really does is read a formula you already had from right to left.
Notation
Annotate
On: \( V = A \cdot d, \qquad d = 2\pi\bar{x} \)
Four short annotations, and each one heads off a specific error.
The third one is the one that catches people in exams. The distance used is measured from the axis of revolution, which is often not a coordinate axis at all. If the axis is the vertical line one unit left of the origin, the radius is the centroid's coordinate plus one, and a problem will often choose such a line deliberately.
The fourth repeats the condition, and it repeats it because it is the only hypothesis in the theorem. If you find yourself getting a volume of zero, or a volume that is obviously too small, check whether the axis cuts the region. There is a trap slide on exactly that a few screens from now.
Worked example
A disc of radius 2 centred at the point 4 on the horizontal axis is revolved about the vertical axis. Find the volume of the ring it sweeps.
Find the area
Why: A disc of radius 2.
\[ A = \pi(2)^2 = 4\pi \]
Find the centroid
Why: By the symmetry principle the centroid of a disc is its centre; it has two perpendicular axes of symmetry.
\[ (\bar{x}, \bar{y}) = (4, 0) \]
Find the distance travelled
Why: One circle of radius 4.
\[ d = 2\pi(4) = 8\pi \]
Multiply
Why: Area times distance.
\[ V = A\cdot d = 4\pi\cdot 8\pi = 32\pi^2 \approx 315.83 \]
Figure (svg): On the left, a circle of radius 2 centred at 4 on the x-axis, with an arrow of length 4 from the y-axis to its centre. On the right, the torus seen from above as two concentric ellipses of radii 2 and 6 with the centroid's circle of radius 4 dashed between them.
Check against a crude estimate
Why: Cut the ring and straighten it: a cylinder of radius 2 and length 8 pi, whose volume is pi times 4 times 8 pi, again 32 pi squared — the stretching on the outside exactly compensates the squashing on the inside, which is what Pappus is saying.
\[ \pi(2)^2\cdot 8\pi = 32\pi^2 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, pp. 623-624 — Example 6.34
This is the example that sells the theorem. Finding the volume of a torus by washers or by shells means an awkward integral with a square root in it; here it is three lines of multiplication.
Every ingredient is something you can read off. The area of a disc is standard, the centroid of a disc is its centre by symmetry, and the distance travelled is one circle of the centre's radius.
The check is worth understanding rather than just reading. Imagine cutting the ring and straightening it out into a cylinder. That cylinder has the disc as its cross-section and the centroid's journey as its length, and it has the same volume. Straightening stretches the outside and compresses the inside, and the theorem is the statement that those two effects cancel exactly. That is the intuition behind the whole result.
Fill the middle
A disc of radius 1 centred at the point 3 is revolved about the vertical axis. Fill in the centroid's journey.
Fill in the blanks
V = A\cdot d = \pi\cdot 2\pi(3) = 6\pi^2
Why: The area of a disc of radius 1 is pi, and its centroid is its centre, which stands 3 units from the axis, so the journey is 2 pi times 3, that is 6 pi. Multiplying gives 6 pi squared, about 59.22.
Fill in the blank before revealing. The only thing to get right is which distance goes into the journey.
It is the distance from the axis to the centroid, which is the centre of the disc at three units out, not the radius of the disc itself. Mixing those two up is the standard error with this theorem, because both are called radii in ordinary speech.
Compare the answer with the previous one. Halving the radius of the disc quarters its area, and moving from four units out to three shortens the journey by a quarter, so the volume drops from thirty-two times pi squared to six, a factor of more than five. Small changes to the disc matter much more than small changes to the distance, because the area depends on the radius squared.
Worked example
You know the volume of a sphere. Use Pappus in reverse to read off a centroid that would otherwise cost an integral.
\[ \text{half-disc of radius } a, \; x \ge 0, \; \text{revolved about the } y\text{-axis} \]
Name the volume swept
Why: A full sphere of radius a.
\[ V = \tfrac43\pi a^3 \]
Name the area
Why: Half a disc.
\[ A = \tfrac12\pi a^2 \]
Write Pappus with the unknown in it
Why: The journey is one circle of the unknown radius.
\[ \tfrac43\pi a^3 = \tfrac12\pi a^2 \cdot 2\pi\bar{x} \]
Solve for the centroid
Why: Divide by the area and by two pi.
\[ \bar{x} = \frac{\tfrac43\pi a^3}{\pi^2 a^2} = \frac{4a}{3\pi} \]
Figure (svg): A half-disc of radius 1 lying to the right of the y-axis, shaded, with a dot on the x-axis at 0.424 marking its centroid and a label giving four over three pi.
Check the answer is a sensible fraction of the radius
Why: Four over three pi is about 0.4244, so the centroid sits at about 42 percent of the radius: less than half way out, which is right, because a half-disc has far more area near its flat side than near the rim.
\[ \frac{4}{3\pi} \approx 0.4244 < \tfrac12 \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 625 — Exercise 284
This is the theorem used in reverse, and it is a genuinely useful trick. Instead of computing a centroid to find a volume, you use a volume you already know to extract a centroid you would otherwise have to integrate for.
The volume of a sphere is something you have known for years. The area of a half-disc is elementary. Put both into the theorem with the centroid as the unknown and solve the resulting linear equation.
The answer, four times the radius over three pi, is worth memorising, because it turns up constantly in engineering problems and in later calculus courses. Note how modest it is: about forty-two percent of the radius, not fifty. A half-disc has much more area near its flat edge than near its curved one, so its balance point sits closer in than you would guess. The figure makes that believable.
Worked example
Revolve the triangle of Part 3 about the vertical axis. Pappus should reproduce a volume formula you have known since school.
\[ \text{triangle } (0,0), (a,0), (0,b) \text{ about the } y\text{-axis} \]
Write the area
Why: Half base times height.
\[ A = \tfrac12ab \]
Recall the centroid
Why: Computed by integration earlier in this lesson.
\[ \bar{x} = \frac{a}{3} \]
Find the journey
Why: One circle of that radius.
\[ d = 2\pi\cdot\frac{a}{3} = \frac{2\pi a}{3} \]
Multiply
Why: Area times distance.
\[ V = \tfrac12ab\cdot\frac{2\pi a}{3} = \frac{\pi a^2 b}{3} \]
Check against the cone formula
Why: One third of the base area times the height, with base radius a and height b — the standard cone volume, recovered from a centroid and a multiplication.
\[ \tfrac13(\pi a^2)(b) = \frac{\pi a^2b}{3} \quad \checkmark \]
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, p. 625 — Exercise 282
This closes a loop opened in Part 3. You integrated for the centroid of a triangle there; here that answer becomes a volume with a single multiplication.
Write the area, recall the centroid coordinate, multiply by the full circle it traces, and the cone formula appears. Nothing is approximated anywhere, and no integral is computed on this slide at all.
The check is the whole point. One third of the base area times the height is the cone volume you learned long before you met calculus, and here it drops out of a centroid and a multiplication. Where does the one third come from in the classical formula? From the centroid of a triangle sitting a third of the way along its base. That is a satisfying piece of mathematics, and it is the kind of connection worth remembering when the formulas start to feel arbitrary.
Trap
A disc of radius 2 centred at the origin, revolved about the vertical axis:
\[ A = 4\pi, \quad \bar{x} = 0 \]
\[ V = 4\pi\cdot 2\pi(0) = 0\,? \]
Wrong. The solid is a sphere of volume about 33.5.
The axis passes straight through the region, which the theorem forbids. The left half sweeps out exactly the same ball as the right half, so the bookkeeping breaks down entirely.
The repair is always the same: keep only the half on one side of the axis, and use that half's own area and centroid.
\[ A = 2\pi, \quad \bar{x} = \frac{8}{3\pi} \]
\[ d = 2\pi\bar{x} = \frac{16}{3} \]
\[ V = 2\pi\cdot\frac{16}{3} = \frac{32\pi}{3} \]
This is the theorem's one hypothesis being violated, and the failure is spectacular rather than subtle: you get a volume of zero for a solid that plainly has one.
The reason is easy to see once you picture it. The left half of the disc and the right half sweep out exactly the same ball, so the region covers the same space twice. The theorem quietly assumes each point of space is swept once, and the centroid sitting on the axis is the arithmetic signalling that assumption has failed.
Notice which half-disc result the fix leans on. It is the one you extracted two slides ago by running the theorem backwards from the sphere, so the corrected calculation here is that argument read forwards again, and it lands on the sphere volume you expected. Two slides, one relationship, used in both directions.
Sorting
Sort into buckets
For each region and axis, can the theorem be used as it stands, or must the region be cut first?
Sort these by asking one question about each: does the axis cross the region? Draw a quick sketch for any you are unsure of, because this is a geometric question and not an algebraic one.
Four of the six are clear of the axis. Touching the boundary, as the triangle starting at one does not even do, is the limiting case and is fine; what matters is that no part of the region lies on both sides.
The two that need cutting are the disc of radius three centred at one, which straddles the vertical axis, and the parabola region, which runs from minus two to two across it. Notice that the last item is the same parabola region as the fourth but revolved about the other axis, and there it is perfectly fine, because the region lies entirely above the horizontal axis. The region alone never decides; the region together with the axis does.
Section
Part 6
Pattern
Figure (svg): A flow diagram: check for an axis of symmetry first, then choose point masses, a rod or a plane region, then compute the moments and divide by the total mass; a side branch shows Pappus turning an area and a centroid into a volume.
This is the order to work in for any problem in this section, and the diagram is the same order drawn.
Symmetry comes first because it is free and because it can eliminate one of the two integrals before you have written anything. Then identify what kind of object you have, because that decides whether you are adding or integrating, but not what you are computing: it is always mass, then moments, then one division per coordinate.
The last two lines are the ones that save marks. The range check costs nothing and catches nearly every arithmetic slip, and the reminder about density explains why so many worked solutions begin by setting it to one. The final line is the return trip: with the centroid in hand, the Theorem of Pappus turns it into a volume, and with a volume in hand it gives the centroid back.
Ranking
Put in order
Order these steps for finding the centroid of the region between two curves.
Why: The limits come first, because you cannot test for symmetry or integrate without them. Symmetry next, since it may delete one of the two moment integrals. Then the area, the moments, the two divisions, and the range check that costs nothing and catches the commonest slips.
Order the steps before checking. There is one genuine subtlety here and one thing that is not negotiable.
The subtlety is where symmetry goes. It has to come after the limits, because until you know how far the region extends you cannot say whether it is symmetric, as the error-analysis slide showed. But it must come before the moments, or you will integrate something you did not need.
What is not negotiable is that the divisions come after all the integrals and the check comes last. If you divide as you go you will lose track of which moment belongs to which coordinate, and if you skip the check you will hand in a centroid that lies outside its own region more often than you would like.
Real world
Discussion prompt
A cantilevered observation deck extends over a canyon. Engineers model it as a lamina and track its centre of mass as visitors walk out onto the far end. Why is that one point the quantity they watch, rather than the total weight?
OpenStax Calculus Volume 1, §6.6 Moments and Centers of Mass §6.6, pp. 621-622 — the Grand Canyon Skywalk student project
Write your answer before revealing. The question is about engineering judgement rather than about calculation, and the reasoning is the same reasoning you used for the seesaw on the first slide of this lesson.
The key idea is in the second line of the reveal. Once each part of the system is replaced by its own centre of mass, the whole structure is just two point masses, and Part One's formula applies directly. That reduction is why engineers track the centre of mass of a complicated structure rather than modelling every beam.
Think about why the visitors at the far edge matter so much. Their share of the total weight is modest, but their arm is the longest in the sum, and the moment multiplies the two. That product is what tips a structure, and it is why an observation deck has a posted limit on how many people may stand at the far end, not merely on how many may be on it at all.
Check
Check your understanding
Masses of 4 kg and 6 kg sit at positions 1 and 6 on a number line. Which statement is correct?
Answer: A
Why: The moment is 4 times 1 plus 6 times 6, which is 4 plus 36, or 40 kilogram-metres. The total mass is 10 kilograms, so the centre of mass is 4 metres, which lies between 1 and 6 and nearer the heavier mass at 6.
This question is about telling apart two quantities that come from the same calculation one line apart. Compute both before choosing.
The moment is four times one plus six times six, which is forty kilogram-metres. The centre of mass is that divided by the total mass of ten kilograms, so four metres. Notice that the answer lies between the two positions and closer to the heavier mass, exactly as a weighted average must.
The wrong options are each a specific error: swapping the two, forgetting the division, and taking the plain midpoint of the positions with no weighting at all. The last of those is the one to watch for in your own work, because a midpoint feels like an answer.
Check
Check your understanding
For the region under y = x³ on the interval from 0 to 1, which integral gives the moment about the x-axis?
Answer: C
Why: The moment about the x-axis integrates half the square of the height. Squaring x cubed gives x to the sixth, and the arm supplies the divisor of 2, so the integrand is x to the sixth over 2 and the moment is 1 over 14.
Work out what each of the four integrands would be before looking at the options. Three of them are real quantities for this region and only one is the moment being asked for.
The moment about the horizontal axis integrates half the square of the height. Squaring the cube gives the sixth power, and the arm supplies the divisor, so the integrand is the sixth power halved.
The two tempting wrong answers are the area, which carries no arm at all, and the moment about the other axis, which has one factor of the position and no squaring. Getting into the habit of naming the arm before writing the integrand will keep these apart.
Check
Check your understanding
A square of side 2 has its centre at the point (5, 0) and is revolved about the y-axis. What volume does it sweep?
Answer: B
Why: The area is 4 and the centroid, by symmetry, is the centre of the square at a distance 5 from the axis, so the journey is 10 pi. Multiplying gives 40 pi, about 125.7.
Compute the area and the journey separately before choosing, and be careful about which number is which.
The area of the square is four, and its centroid is its own centre, five units from the axis by symmetry, so the journey is two pi times five. Multiplying the two gives forty pi.
Each wrong option is a specific slip: using the radius instead of the full circumference, reporting the journey on its own as if it were a volume, and multiplying two lengths together instead of an area by a length. That last one is the most useful to notice, because a units check kills it immediately. Pappus always multiplies something in square units by something in linear units.
Explain it to yourself
Discussion prompt
In two sentences, explain why the moment about the y-axis integrates the position times the height, while the moment about the x-axis integrates half the square of the height. Do not quote the formulas; explain the arms.
Write your two sentences before revealing. If you can explain this, you never have to memorise either moment formula again, because you can rebuild both from the picture of a single strip.
The whole explanation is about arms. Horizontally, a vertical strip is concentrated at one position, so it acts there and the arm is that position. Vertically, the same strip is spread out from the axis to the curve, so it acts at its own middle, and the arm is half the height.
One extra thing to say to yourself as you write: the square in the second formula has two different sources. One factor of the height is the strip's mass, the other is the arm. Being able to say which is which is the difference between knowing the formula and remembering it.
Exit ticket
\[ R: \; 0 \le x \le 3, \; 0 \le y \le 9 - x^2 \]
Discussion prompt
Find the centroid of R, say whether symmetry helped, and then use the Theorem of Pappus to give the volume swept when R is revolved about the y-axis.
This pulls the whole lesson into one problem, so take it slowly. Find the area, both moments, both coordinates, and then convert the first coordinate into a volume.
The first thing to check is symmetry, and here it gives you nothing: the region is only the right half of the parabola, so there is no vertical line you could fold it along. Both coordinates have to be earned.
The last line is the one to think about. The volume by Pappus is the area times the circle traced by the first coordinate, and if you write it out you will find it is two pi times the same integral that produced the moment about the vertical axis. That is not a coincidence, it is the proof from earlier in this part, and being able to see it in your own numbers is a good sign that the theorem has landed.
Recap
| object | total mass | moments | centre of mass |
|---|---|---|---|
| point masses on a line | the sum of the masses | the sum of mass times position | the moment over the total mass |
| point masses in a plane | the sum of the masses | mass times y, and mass times x | each moment over the total mass, crossed over |
| a rod with density ρ | the integral of ρ | the integral of x times ρ | the moment over the mass |
| a region under f | ρ times the integral of f | ρ times the integral of f²/2, and of x f | each moment over the mass; ρ cancels |
| a region between f and g | ρ times the integral of f − g | ρ times the integral of (f² − g²)/2, and of x(f − g) | each moment over the mass |
\[ V = A\cdot d, \qquad d = 2\pi\bar{x} \]
Two shortcuts are worth more than the formulas: symmetry, which hands you a coordinate for nothing, and Pappus, which turns a centroid into a volume or a volume back into a centroid. Chapter 6 ends here; the same moment integrals return in multivariable calculus with a density that varies across the region as well as along it.
Stewart, Calculus: Early Transcendentals 8e, §8.3 Applications to Physics and Engineering §8.3, pp. 558-568 — the same material in Stewart
One recipe, five settings. Total mass, moments, divide. The only thing that changes across the rows of this table is how the two totals are computed, and even that is just a sum becoming an integral.
Two things are worth having at your fingertips. The arm to the vertical axis is always the position; the arm to the horizontal axis is half the height, or the average of the two heights for a region between curves. And a constant density always cancels, which is why a centroid is a fact about a shape.
If you keep only three results from this lesson, keep these. The centroid of a triangle is a third of the way along each leg. The centroid of a half-disc is four times the radius over three pi from the flat side. And the volume of a solid of revolution is the area times the distance the centroid travels. Between them they answer a surprising share of the exercises in this section and of the problems you will meet later.
Want this taught 1-on-1? Alexander tutors Calculus I — $55/session, free consultation.