Chapter 15: Multiple Integrals

Welcome to Last Minute Lecture.

This free chapter overview is designed to help students review and understand key concepts.

These summaries supplement, not replace, the original textbook and may not be redistributed or resold.

For complete coverage, always consult the official text.

You know that feeling when you're just faced with this mountain of information, maybe prepping for a big meeting, or you're trying to get your head around a whole new field, or maybe just deep curiosity pulls you in.

It can be totally overwhelming.

Exactly.

You want that deep understanding.

You want those aha moments, but seriously, who has the time to wade through every single bit of it?

That's really why we do what we do here on The Deep Dive.

We take a stack of sources, could be articles, research papers, even your own notes, and we pull out the absolute key insights.

Kind of a shortcut to genuine understanding.

Getting right to the heart of it.

Precisely.

Today, we are diving deep into something really fundamental.

It underpins so much of the world we see and don't see.

It's chapter 15 on multiple integrals from the big calculus text, Calculus, Early Transcendentals by Stuart, Clegg, and Watson.

Classic.

Multiple integrals, yeah.

Our mission today, we're going to unpack these powerful math tools, try to make them super clear, relatable,

show you some honestly surprising real -world uses, and connect it all back to the bigger calculus journey.

Sounds like a plan.

It's about moving beyond lines and simple areas, right, into, well, multiple dimensions.

Exactly.

How do we measure and understand things when they're complex?

3D shapes, or maybe even abstract distributions.

Okay, let's unpack this.

So maybe let's rewind just a bit.

Single variable calculus, remember the definite integral.

Area under a curve.

Right.

That was our tool.

We sliced the area under a curve into these incredibly thin rectangles, summed them up, and the limit of that, the Riemann sum, gave us the exact area.

The source even has that classic picture, figure one.

And now we're taking that brilliant idea, that core concept, and kind of launching it into another dimension.

From area to volume.

Exactly.

Just like that single integral finds area under a 2D curve, a double integral finds the volume under a 3D surface.

Think of it like, instead of measuring a flat carpet, you're measuring the space trapped under a bumpy blanket draped over a region.

That's a great visual.

So imagine a rectangle on the floor.

That's our base region.

And then this bumpy blanket, that's the surface, defined by some function is laid over it.

We want the volume of air, or whatever, between the floor and that blanket.

So we do the same thing as before.

Slice up that base rectangle into lots and lots of tiny sub -rectangles.

Just like slicing the interval on the x -axis before.

And now, over each of those tiny base rectangles, you can imagine building a thin rectangular column.

Like a little tower or box.

And its height.

The height is just the function's value, how high the blanket is at some sample point within that tiny rectangle.

Figure four in the source shows this really well, these little columns rising up.

Okay, so summing up the volumes of all those tiny columns gives us an approximation of the total volume.

Figure five shows that.

It's kind of like building a Lego model of a mountain, right?

Exactly like that.

The more sub -rectangles you use, the tinier your Lego bricks, the better the approximation gets.

And the double integral itself is the limit of this double Riemann sum, as those little rectangles get infinitesimally small.

The chapter even shows this with an example, estimating volume under an elliptic paraboloid.

Figure eight shows how using four, then 16, then 64, then 256 squares gets you closer and closer to the real volume.

It converges nicely.

But, okay, here's the kicker.

How do we actually calculate these volumes without, you know, summing infinite tiny columns?

That seems hard.

Ah, yes.

That's the practical breakthrough.

Iterated integrals.

This is the real power move for actually doing the calculation.

Iterated, meaning one after another.

Precisely.

Instead of one super complex limit, we solve a double integral by doing two successive single integrations.

The trick is partial integration.

Partial integration, like partial derivatives.

Sort of analogous, yeah.

You essentially treat one variable, say y, as a constant, while you integrate with respect to the other variable, x.

Then you take that result, which will probably still have y in it, and integrate that with respect to y.

Okay, so you break it down, integrate along one direction than the other, almost like taking slices.

Exactly.

You can think of it as finding the area of a cross -section by integrating along, say, the way direction, and then integrating all those cross -sectional areas along the x direction to get the total volume.

And does the order matter?

Like x first, then y, versus y first, then x.

Well, here's something really neat.

For continuous functions over nice rectangular regions, Fubini's theorem, that's theorem 10 in the text, tells us, no, the order doesn't matter, you get the exact same result.

Wow, okay.

That seems important.

It's incredibly powerful.

It gives us flexibility.

Think about slicing that 3D solid again.

You can slice it vertically or horizontally, some of the slices, you still get the same total volume.

Figures 11 and 12 show this idea of slicing in different directions.

That makes sense.

But even if the answer is the same, is one order sometimes, like, way easier than the other?

Oh, absolutely.

That's the practical art of it.

Choosing the right order can be the difference between a straightforward calculation and, well, a nightmare.

The source gives an example, example six, I think,

where one order leads to a really tricky integration by parts, but flipping the order makes it simple.

Huh.

So you gotta look ahead a bit, plan your attack.

Definitely pays off.

And this isn't just about volume, is it?

I remember something about average value, too.

Right.

We can find the average value of a function over a region.

Think of the function surface again, that bumpy blanket or mountainous terrain.

The average value is like finding a single height, a level, such that if you sliced off all the mountain peaks above that level, you'd have exactly enough material to fill in all the valleys below it, making the whole landscape perfectly flat at that average height.

Figure 17 gives a nice picture of this leveling idea.

So it's like finding the average altitude of a mountain range.

Precisely.

And this is super useful.

Example nine uses this with the midpoint rule to estimate average snowfall over Colorado from a contour map.

Think about water resource management, predicting runoff.

It's taking this abstract math and applying it to really complex, large -scale natural phenomena.

OK.

So iterated integrals handle rectangles.

But what about, you know,

weird shapes, triangles,

areas bounded by curves?

The real world isn't always square.

That's the next step, section 15 .2.

We need ways to handle a general region.

General regions.

OK.

We classify them mainly into two types.

Type I regions are where, for a given range of X values, the Y value goes from some lower function, say G1X, up to an upper function, G2X.

So Y is trapped between two curves that depend on X.

Exactly.

And type II regions are the reverse.

For a given range of Y, the X value goes from a left function, H1Y, to a right function, H2Y.

Figures five and seven show these pretty clearly.

And setting up the limits of integration is key here.

I like the source's arrow analogy.

Yeah, that's helpful.

For type I, draw a vertical arrow.

It enters the region at the bottom curve, YG1X, and exits at the top curve, YG2X.

Those become your inner integral limits.

And for type II, a horizontal arrow.

Enters at X, H1Y, exits at X, H2Y.

Those are the inner limits, then.

Precisely.

It helps visualize how the variable you're integrating first is moving across the region.

Now, can a region be both type I and type II?

Often yes.

And that's where strategy comes in again.

One description might be much simpler.

Maybe one boundary curve is defined piecewise in terms of X, but it's a single, simple function in terms of Y.

Choosing the type II description could save you from splitting the integral into multiple pieces.

So drawing the region first is probably essential.

Almost always the best first step.

Example III really drives this home.

And I guess changing the order of integration becomes even more powerful for these general regions.

Maybe not just for convenience, but sometimes making an impossible integral possible.

Absolutely.

This is where you get those real aha moments in calculus class.

You're staring at an integral, say the integral of sin Y2 first, which you just can't do directly.

It's non -elementary.

Right.

No simple antiderivative.

But then you look at the region, describe it the other way, say is type pic instead of type Y, reverse the order of integration.

And suddenly the integral becomes maybe integrating something like X and Y2 dx first, which is easy.

Then the second integration works out beautifully.

Example V is the classic case study for this.

It feels like magic, but it's a smart setup.

That flexibility is amazing.

And just quickly, the basic properties still hold, right?

Like integrating sums or constant multiples.

Linearity still works.

And a handy one, if you just integrate the constant function one over any region D.

You just get the area of D.

Exactly.

Property nine.

A nice link back to basic geometry.

Okay.

Makes sense.

Now, what about circles, discs, rings, sectors, things with, you know, radial symmetry?

Trying to describe a disc with X and Y bounds seems awkward.

Extremely awkward.

That's where polar coordinates shine in integration, just like they do in graphing.

Section 15 .3 covers this.

So we switch from X, Y to R theta.

Using the standard transformations.

X equals R cos theta, Y equals R sin theta.

Okay.

Standard stuff.

But the tricky part, the bit everyone forgets, is the area element dA.

It's not just drd theta.

No, and this is absolutely critical.

dA becomes rdrd theta.

You must include that extra factor of r.

Why the r again?

I know it's important, but the intuition.

Think about those tiny polar rectangles we use for the Riemann sum and polar coordinates.

They aren't uniform squares.

A small change in r and theta carves out a shape that's wider the further you are from the origin.

Larger r.

It's like a tiny piece of wedge or an annulus sector.

That r factor precisely accounts for the fact that the area of this little piece depends on how far out it is.

I love you.

Your five shows this geometric origin.

Forgetting the r gives the wrong answer.

Every time.

Got it.

dA equals rdrd theta.

Draw that in.

Please do.

And the payoff is huge.

Integrals that involve y2 plus y2 become integrals involving just r2, which is often much nicer.

And integrating over regions like disks or sectors becomes incredibly simple, usually with constant bounds for r and theta.

So calculating the volume of, say, a paraboloid over a circular base is much cleaner in polar.

Way cleaner.

Example three shows that.

It also makes finding areas of things like polar curves, remember those rose curves,

Example four finds the area of one loop.

Yeah, those were cool.

And even volumes bounded by more complex shapes, like finding the volume under a paraboloid but inside a cylinder whose equation is maybe by 2 plus y2 plus y2 equals 2x.

Polar handles that beautifully.

That cylinder equation just becomes r2 cos theta in polar, giving you your r bounds directly.

Example five tackles that kind of problem.

It turns potentially messy Cartesian limits into elegant polar ones.

Okay, this is powerful stuff, but let's bring it back.

What does this all mean for you, the listener?

How do these calculations actually, you know, solve real problems?

Oh, the applications are everywhere.

Think about physics and engineering.

If you have a thin plate, a lamina, and its density isn't uniform.

Maybe it's heavier in some parts than others.

How do you find its total mass?

Double integral.

Double integral.

You integrate the density function over the area of the lamina, basically mass integral of density dA.

It's summing up density times tiny area pieces.

The same idea applies to finding total electric charge.

If you know the charge density over a region.

Example one calculates charge on a triangle.

Makes sense.

And what about balance, like finding the center of mass?

Crucial concept.

The center of mass is the point where the lamina would perfectly balance if you could put it on a pinpoint.

Think balancing in a regular tray.

Double integrals are how we find the coordinates of this balance point, especially if the density varies.

Example two finds it for a triangular plate.

Is there ever a shortcut for finding the center of mass?

Sometimes.

Symmetry is your friend.

If a lamina has an axis of symmetry, and its density is also symmetric about that axis, then the center of mass must lie on that axis.

This can often tell you one coordinate immediately, saving a lot of calculation.

Example three, with a semi -circular lamina, uses this.

Nice.

What else?

Rotation?

Yes, moment of inertia.

That's basically the rotational equivalent of mass.

It measures how resistant an object is to being spun up or slowed down.

Like how a heavy flywheel is hard to start spinning, but also hard to stop?

Exactly that idea.

Double integrals let us calculate the moment of inertia for 2D shapes.

For example, calculating it for a simple homogenous disk gives that classic formula I equals 12 meters A2, which pops up all over physics.

A gamble four works through that.

Okay, physics, engineering, what about chance?

Probability?

Yep.

Multiple integrals play a big role there too.

Especially with continuous random variables.

If you have two random variables, X and Y, a double integral of their joint probability density function over a certain region tells you the probability that X, Y falls within that region.

So you could calculate, say, the probability that someone waits less than 10 minutes for movie tickets and less than five minutes for popcorn.

If you know the distribution of waiting times, yes.

Example seven does a similar calculation, finding the probability the total wait time is less than 20 minutes.

It lets us model and quantify combined uncertainties.

That's really practical.

Any other examples?

Manufacturing quality control is a big one.

Imagine making roller bearings.

Their diameter and lengths might vary slightly, maybe following a normal distribution.

Double integrals can calculate the probability that a randomly chosen bearing meets specific tolerance requirements for both dimensions simultaneously.

Example eight touches on this, using a bivariate normal distribution shown in figure nine.

Wow, okay.

And surface area too.

Like the actual area of a curved surface, not the volume under it.

Right, section 15 .5, if you have a surface defined by Z, F, X, Y over a region D in the xi plane, there's a specific double integral formula to find its area.

It involves the square root of one plus the squares of the partial derivatives.

Score it one plus dF dS2 plus dF dY2.

That looks kind of familiar.

It's very analogous to the arc length formula from single variable calculus, just bumped up a dimension.

It's how you'd find the area of a piece of a paraboloid or a tilted plane segment.

Okay, so double integral cover areas, volumes under surfaces, mass, center of mass, probability, surface area.

Yeah.

What's next?

Going fully 3D.

Exactly.

We move from double integrals to triple integrals.

Just as we went from lines to areas, now we go from areas to volumes.

But volumes of actual 3D solids calculating properties throughout the solid, not just on a surface or base.

So integrating over a block or a sphere or some weird 3D shape.

Precisely.

We start with integrating over simple rectangular boxes, then generalized to any bounded solid region in 3D space.

And Fubini's theorem still holds.

Can we change the order?

Yes.

Fubini extends to triple integrals.

For well -behaved functions and regions, you have six possible orders of integration.

DZ, DZDX, DZDX, DZD, DZD, DZD, etc.

And they all give the same answer.

Six orders.

Wow.

Choosing the right one must be even more critical then.

It absolutely is.

Defining the limits for a general 3D solid can be tricky.

You might need to think about surfaces bounding the solid from below and above, for Z, then project the solid onto a 2D plane and use type I or type II bounds for Y and X.

Visualizing the solid and choosing the order that gives the simplest limits is often the hardest but most important part of the problem.

Example 3 in 15 .6 shows how changing the description makes a problem feasible.

And what do triple integrals let us calculate?

Well, the most basic thing.

If you integrate the constant function 1 over solid region E.

You get the volume of E.

You got it.

Example 5 calculates a tetrahedron's volume this way.

But more importantly, we can find the total mass of a solid object if its density varies throughout into great density dV.

And we can find the center of mass of a 3D object too.

The concepts extend directly from 2D just with an extra integral.

Example 6 finds the center of mass of a solid bounded by a parabolic cylinder.

Okay, so just like polar coordinates help simplify 2D integrals with circular symmetry, are there special coordinate systems for 3D too?

Yes, thank goodness.

There are two main ones that are incredibly useful.

Cylindrical and spherical coordinates.

Let's start with cylindrical.

What's the idea?

Cylindrical coordinates, r, theta, z, are basically just polar coordinates in the xi plane, with a regular z coordinate added for height.

So r and theta describe the position in the horizontal plane, and z tells you how high up or down you are.

Figure 2 in 15 .7 shows this.

So when would you use these?

They're perfect for solids that have symmetry around an axis, usually the z axis, think cylinders obviously, but also cones, paraboloids, things like that.

Any time by 2 plus y2 shows up in the equations defining the solid or the function you're integrating, that's a big hint.

Cylindrical might be good because by 2 plus y2 just becomes r2.

And the volume element dv becomes, don't tell me there's another r.

In cylindrical coordinates, dv equals r, dz, dr, d, theta.

That r comes to the polar coordinate part in the base plane.

You're essentially stacking up those polar area elements r, dr, d, theta,

along the z axis, visualizing a tiny cylindrical wedge helps see where it comes from.

Okay, dv equals r, dz, dr, d, theta, got it.

And this simplifies things like finding the mass of a cylinder or volume inside a paraboloid.

Dramatically simplifies them, often turning complicated bounds in x, y, z into simple constant bounds in r, theta, z.

Examples 3, 4, and 5 in 15 .7 really showcase this power.

All right, so cylindrical for axis symmetry, what about spherical coordinates?

Spherical coordinates, rho, theta, theta, are for problems with symmetry about a single point at the origin.

Think spheres or parts of spheres or cones centered at the origin.

Okay, what were rho, theta, and phi?

Rho is the distance directly from the origin to the point.

Theta is the same angle as in cylindrical polar coordinates, the angle in the xi plane from the positive x axis, and phi is the angle down from the positive z axis.

Rho is always non -negative, theta goes 0 to 2p pi, and phi goes from 0 straight up to pi straight down.

Figure 1 in 15 .8 lays these out.

That phi angle seems a bit different, and the volume element dv must be interesting here.

It's the most complex but also the most powerful for the right problems.

And spherical coordinates dv rho 2 sin phi dr o o d theta.

Whoa, rho 2 sin phi, where does all that come from?

It comes with a geometry of a tiny spherical box or wedge.

As you move further from the origin, rho increases.

The volume element expands, that's the rho 2 part.

And as you move away from the z axis towards the xi plane, phi goes from 0 or pi towards pi 2.

The wedge also gets wider, that's the sin phi part.

Sin phi is largest at the equator, phi pi in small is near the poles, phi 0 pi.

This factor correctly weights the volume contribution.

Figure 8 gives a sense of this.

That's quite a factor to remember.

Rho 2 sin phi dr o o d theta d theta phi.

But it makes integrals over spheres or parts of spheres incredibly easy.

Imagine trying to integrate something like e by 2 plus y2 plus z2 32 over the unit ball in rectangular coordinates.

Nightmare.

By 2 plus y2 plus z2, that's just rho 2.

Exactly.

So that integrand becomes e, rho 2 equals e o 3, and the bounds for the unit ball are simply 0 2 0 equals 1, 0 equals theta 2 pi, 0 equals pi.

The integral becomes beautiful.

Example 3 in 15 .8 is the classic illustration of this ultimate simplification.

It's all about choosing the right coordinate system for the geometry of the problem.

Incredible.

It really shows the power of having the right tool.

It really does.

So wow, we've really covered some ground here.

From just area under a curve, we went to volume under a surface using double integrals, learned how to handle weird shapes with type I and type II regions,

tamed circles with polar coordinates, then jumped into real world applications,

mass, center of mass, probability, surface area, then pushed into three dimensions with triple integrals, and finally found these specialized tools, cylindrical and spherical coordinates, to handle symmetries in 3D space.

It's quite a journey through dimensions.

It really is.

And hopefully for you listening, this has felt like a massive shortcut to getting a real handle on multiple integrals, those little connections, those aha moments.

That's where information turns into real insight, right?

And just think about the elegance here.

These mathematical tools let us precisely describe and quantify shapes, densities, probabilities, even in dimensions or situations we can't easily picture,

like finding the volume of an irregular tumor mentioned at the start of the chapter, or modeling airflow, or understanding population distributions.

The framework is incredibly versatile.

It makes you wonder what other unseen phenomena out there, maybe in fields we haven't even thought of, could these same principles help us understand and model something to ponder?

Definitely food for thought.

Well, we really hope this deep dive has sparked your curiosity, maybe answered some questions, and given you some of those aha moments about multiple integrals.

Thank you for joining us on the deep dive.

From the Last Minute Lecture Team, thank you for listening.