Triple integrals
A block of snow is packed denser at the bottom than at the top, and its top surface is slanted. Its mass is a limit of sums over tiny boxes. How is that limit defined, and how is it computed one variable at a time?
Where you are 6 in Multiple integrals
Why this exists
Naomi and Andre forecast avalanches. They dig down to a buried weak layer, cut out a block of the snow above it, and judge the load the layer carries: the block’s mass. Snow is packed denser near the bottom, and the surface is not parallel to the layer, so the mass is not density times volume. Cut the block into small boxes, treat the density as constant on each, add the masses, and take the limit.
That limit is a triple integral. A double integral \iint_D h\,dA gives the volume under a graph but counts every point of the solid equally; weighting each point by a density needs an integrand of three variables over a solid.
Definition
Let B = [a_1, b_1] \times [a_2, b_2] \times [a_3, b_3] be a closed box and f : B \to \mathbb{R} bounded. A partition P of B cuts each edge interval,
giving sub-boxes B_{ijk} = [x_{i-1}, x_i] \times [y_{j-1}, y_j] \times [z_{k-1}, z_k] of volume \Delta V_{ijk} = \Delta x_i\,\Delta y_j\,\Delta z_k. The mesh \lVert P \rVert is the largest of all the \Delta x_i, \Delta y_j and \Delta z_k. With a sample point \mathbf{p}_{ijk} in each sub-box, S(P) = \sum_{i, j, k} f(\mathbf{p}_{ijk})\,\Delta V_{ijk} is a triple Riemann sum. The function f is integrable over B, with triple integral I = \iiint_B f(x, y, z)\,dV, if for every \varepsilon > 0 there is a \delta > 0 with \lvert S(P) - I \rvert < \varepsilon for every partition with \lVert P \rVert < \delta and every choice of sample points.
Solids. For a bounded set E in \mathbb{R}^3 and f defined on E, place E in a box B, let f^E be f on E and 0 on the rest of B, and define \iiint_E f\,dV = \iiint_B f^E\,dV when the right side exists. The volume of E is V(E) = \iiint_E 1\,dV, and when V(E) > 0 the average value of f over E is \displaystyle f_{\text{avg}} = \frac{1}{V(E)} \iiint_E f\,dV.
Simple solids. A solid is z-simple if
where D is a closed region of the xy-plane that is type I or type II (Iterated integrals and Fubini's theorem) or radially simple (Double integrals in polar coordinates), and u_1 \le u_2 are continuous on D. Every vertical line through a point of D meets E in one segment, and D is the shadow of E on the xy-plane. An x-simple solid lies between graphs x = u_1(y, z) and x = u_2(y, z) over a region of the yz-plane, and a y-simple solid between graphs y = u_1(x, z) and y = u_2(x, z) over a region of the xz-plane.
Mass. A solid E with continuous density \delta(x, y, z), in mass per unit volume, has mass and first moments
and center of mass \displaystyle (\bar{x}, \bar{y}, \bar{z}) = \left(\frac{M_{yz}}{M}, \frac{M_{xz}}{M}, \frac{M_{xy}}{M}\right). The moment M_{yz} weights each piece by its signed distance x from the yz-plane. With constant density the center of mass is the centroid. The moment of inertia about the z-axis is I_z = \iiint_E (x^2 + y^2)\,\delta\,dV.
Intuition
Group the sub-boxes of a Riemann sum into columns, one over each rectangle R_{ij} = [x_{i-1}, x_i] \times [y_{j-1}, y_j] of the base. Inside a column only z changes, so the column’s terms add to a Riemann sum in z times \Delta A_{ij}, close to \left(\int f\,dz\right) \Delta A_{ij}. Adding the columns is a double Riemann sum of that inner integral: a triple integral is a single integral inside a double one.
For a z-simple solid the column over a point (x, y) of the shadow D runs from the lower surface z = u_1(x, y) to the upper surface z = u_2(x, y) (Figure 1). Units multiply: kilograms per cubic meter times cubic meters gives kilograms.
Results
Result 1 (integrability, stated without proof). A function continuous on a closed box is integrable over it. If E is an x-, y- or z-simple solid and f is continuous on E, then f^E is integrable over every box containing E (its discontinuities lie on the boundary of E, a set of zero volume), and V(E) exists. The proofs belong to Real Analysis, like Result 1 of Double integrals as volume.
Result 2 (properties). Linearity, monotonicity, additivity over two solids that share only boundary points, and the bounds c_1\,V(E) \le \iiint_E f\,dV \le c_2\,V(E) when c_1 \le f \le c_2, hold with the proofs of Result 3 of Double integrals as volume, read with boxes for rectangles.
Theorem 3 (Fubini’s theorem on a box). Let B = R \times [a_3, b_3] with R = [a_1, b_1] \times [a_2, b_2].
(a) Let g be integrable over B. If for every (x, y) in R the function z \mapsto g(x, y, z) is integrable on [a_3, b_3], and G(x, y) = \int_{a_3}^{b_3} g(x, y, z)\,dz is integrable over R, then \iiint_B g\,dV = \iint_R G\,dA.
(b) If f is continuous on B, then \iiint_B f\,dV equals each of the six iterated integrals, one for each order of x, y and z; for example
Theorem 4 (simple solids). Let E be z-simple as in the definition and let f be continuous on E. Then
and the double integral is computed by Theorem 4 of Iterated integrals and Fubini's theorem or Theorem 2 of Double integrals in polar coordinates. The same holds for x-simple and y-simple solids with the letters exchanged, and a solid simple in two directions gives the same value both ways.
Corollary 5 (volume). For a z-simple solid, V(E) = \iint_D \left(u_2 - u_1\right) dA. With u_1 = 0 and u_2 = h \ge 0 this is the volume under the graph of h in Result 2 of Double integrals as volume, so the two meanings of volume agree.
Why it is true
As in Iterated integrals and Fubini's theorem, two facts from Real Analysis are used: Result 1, and uniform continuity: a function continuous on a closed bounded set K has, for every \varepsilon > 0, a \delta > 0 with \lvert f(\mathbf{p}) - f(\mathbf{q}) \rvert < \varepsilon whenever \mathbf{p}, \mathbf{q} lie in K and \lVert \mathbf{p} - \mathbf{q} \rVert < \delta.
Theorem 3 (a). Let I = \iiint_B g\,dV. For a partition P let m_{ijk} and M_{ijk} be the infimum and supremum of g on B_{ijk}, L(P) = \sum m_{ijk}\,\Delta V_{ijk} and U(P) = \sum M_{ijk}\,\Delta V_{ijk}. Given \varepsilon > 0, the first step of the proof of Theorem 1 of Iterated integrals and Fubini's theorem, read with sub-boxes for subrectangles, gives \delta > 0 with I - \varepsilon \le L(P) \le U(P) \le I + \varepsilon whenever \lVert P \rVert < \delta.
Fix (x, y) in R_{ij}. On [z_{k-1}, z_k], m_{ijk} \le g(x, y, z) \le M_{ijk}, so integrating and adding over k,
Multiply by \Delta A_{ij} = \Delta x_i\,\Delta y_j and add over i and j: for any sample points \mathbf{q}_{ij} in R_{ij}, the double Riemann sum \sum_{i,j} G(\mathbf{q}_{ij})\,\Delta A_{ij} lies between L(P) and U(P), so within \varepsilon of I. As \lVert P \rVert \to 0 the partition of R has mesh at most \lVert P \rVert, and since G is integrable these sums tend to \iint_R G\,dA. Hence \left\lvert \iint_R G\,dA - I \right\rvert \le \varepsilon for every \varepsilon > 0. \blacksquare
Theorem 3 (b). By Result 1, f is integrable, and each z \mapsto f(x, y, z) is continuous. Let F(x, y) = \int_{a_3}^{b_3} f(x, y, z)\,dz. Given \varepsilon > 0, take \delta from uniform continuity on B. If (x, y) and (x', y') are within \delta, so are (x, y, z) and (x', y', z) for every z, and
So F is continuous, hence integrable, and (a) gives \iiint_B f\,dV = \iint_R F\,dA, which Theorem 3 of Iterated integrals and Fubini's theorem writes in either order: two of the six. The same argument with x or y innermost gives the other four. \blacksquare
Theorem 4. By the The extreme value theorem, u_1 has a least and u_2 a greatest value on D; choose a_3 below the first, b_3 above the second, a rectangle R containing D, and let B = R \times [a_3, b_3]. By Result 1, g = f^E is integrable over B. For (x, y) in D, z \mapsto g(x, y, z) is f on [u_1, u_2] and 0 elsewhere, so it is integrable with integral
and for (x, y) outside D it is 0. So G = F^D. To see that F is continuous on D, substitute z = u_1 + t(u_2 - u_1), 0 \le t \le 1:
where k(x, y, t) = f\left(x, y, u_1 + t(u_2 - u_1)\right) is continuous on the closed bounded set D \times [0, 1] (both sides are 0 where u_1 = u_2). The estimate of part (b), applied to k, shows the integral in t is continuous, so F is. Then G = F^D is bounded and continuous except on the boundary of D, so it is integrable over R with \iint_R G\,dA = \iint_D F\,dA (Result 1 of Double integrals as volume), and Theorem 3 (a) gives the formula. Both descriptions of a solid simple in two directions compute the one number \iiint_E f\,dV. \blacksquare
Corollary 5. With f = 1 the inner integral is u_2 - u_1. \blacksquare
Worked examples
Choosing the order on a box
Evaluate \iiint_B y z\sin(xz)\,dV over B = [0, 1] \times [0, 2] \times [0, \pi].
SolutionThe integrand is continuous, so by Theorem 3 (b) all six orders give the answer, and the choice is about labor. For fixed z, z\sin(xz) is the derivative in x of -\cos(xz), so integrate in x first; y only rides along:
With z first, integration by parts gives, for x > 0,
It is the derivative of \displaystyle -\frac{\sin \pi x}{x}, which tends to -\pi as x \to 0^+, so its integral over [0, 1] is 0 - (-\pi) = \pi; with the factor 2 from y, the same 2\pi, after more work.
The load on the weak layer
Naomi cuts a block from the snow on a slope. Put the xy-plane on the weak layer, with x up the slope, y across it and z perpendicular to the layer, in meters. The block stands on 0 \le x \le 2, 0 \le y \le 3, and its top is the plane \displaystyle z = 1 + \frac{x}{4}: 1 meter deep at its downhill face, 1.5 meters at its uphill face (Figure 2). From the pit wall Andre measures the density \delta = 300 - 100z kilograms per cubic meter. Find the mass of the block and its average density.
SolutionThe block is z-simple over the rectangle D = [0, 2] \times [0, 3], between u_1 = 0 and \displaystyle u_2 = 1 + \frac{x}{4}, so by Theorem 4
Write \displaystyle h = 1 + \frac{x}{4} for the depth. The inner integral is \Big[300z - 50z^2\Big]_0^h = 300h - 50h^2. In the middle integral substitute dx = 4\,dh, with h from 1 to \displaystyle \frac32:
Nothing depends on y, so the outer integral multiplies by 3: M = 1775 kilograms. By Corollary 5 the volume is \displaystyle 3\int_0^2 \left(1 + \frac{x}{4}\right) dx = 3 \cdot \frac52 = 7.5 cubic meters, so the average density is \displaystyle \frac{1775}{7.5} = \frac{710}{3} \approx 236.7 kilograms per cubic meter, between the 150 at the top of the uphill face and the 300 on the layer, as Result 2 requires.
One solid, two descriptions
Let E be the solid 0 \le x \le 1, 0 \le y \le x, 0 \le z \le 1 - x^2. Evaluate \iiint_E y\,dV with z innermost, then with x innermost.
SolutionAs given, E is z-simple over the triangle with vertices (0, 0), (1, 0) and (1, 1), between the plane z = 0 and the parabolic cylinder z = 1 - x^2 (Figure 1). By Theorem 4,
For x innermost, fix (y, z). The conditions on x are x \ge y and \displaystyle x \le \sqrt{1 - z}, which implies x \le 1. Such x exist exactly when \displaystyle y \le \sqrt{1 - z}, that is z \le 1 - y^2. So E is also x-simple, over 0 \le y \le 1, 0 \le z \le 1 - y^2 in the yz-plane, between x = y and \displaystyle x = \sqrt{1 - z}. Since \displaystyle \int_0^{1 - y^2} \sqrt{1 - z}\,dz = \frac23\left(1 - y^3\right),
as Theorem 4 says it must.
A cylinder cut by a slanted plane
Find the centroid of the solid E inside the cylinder x^2 + y^2 = 4, above the plane z = 0 and below the plane z = 4 + x.
SolutionOn the disk D: x^2 + y^2 \le 4, 4 + x \ge 2 > 0, so E is z-simple over D, between 0 and 4 + x. The disk is symmetric under x \mapsto -x, so \iint_D x\,dA = 0, and by Corollary 5
By Theorem 4 and then Theorem 2 of Double integrals in polar coordinates, with \iint_D x^2\,dA = \int_0^{2\pi} \int_0^2 r^3\cos^2\theta\,dr\,d\theta = \pi \cdot 4,
The reflection y \mapsto -y maps E to itself, so M_{xz} = 0. The centroid is \displaystyle \left(\frac{4\pi}{16\pi}, 0, \frac{34\pi}{16\pi}\right) = \left(\frac14, 0, \frac{17}{8}\right). It sits above height 2 and toward x > 0, where the solid is taller.
Reversing a triple iterated integral
Evaluate \displaystyle \int_0^1 \int_0^x \int_x^1 \frac{\sin y}{y}\,dy\,dz\,dx.
SolutionTake \displaystyle \frac{\sin y}{y} to be 1 at y = 0, which makes it continuous. It has no elementary antiderivative, so read the solid off the limits and reorder. The limits say 0 \le x \le 1, 0 \le z \le x and x \le y \le 1, that is
As written, E is y-simple over the triangle 0 \le z \le x \le 1 of the xz-plane. It is also z-simple: over the type II triangle 0 \le x \le y \le 1 of the xy-plane, z runs from 0 to x. By Theorem 4 both ways,
about 0.1506. The two inner integrals gave \displaystyle \frac{y^2}{2}, the area of the slice of E at y: the slicing of Volumes with known cross sections, with an integrand constant on each slice.
The centroid of a tetrahedron
Let a, b, c > 0 and let T be the tetrahedron with vertices (0, 0, 0), (a, 0, 0), (0, b, 0) and (0, 0, c). Prove that \displaystyle V(T) = \frac{abc}{6} and that the centroid of T is \displaystyle \left(\frac{a}{4}, \frac{b}{4}, \frac{c}{4}\right), the average of the four vertices.
SolutionThe face opposite the origin lies in the plane \displaystyle \frac{x}{a} + \frac{y}{b} + \frac{z}{c} = 1, so T is the set where x, y, z \ge 0 and \displaystyle \frac{x}{a} + \frac{y}{b} + \frac{z}{c} \le 1. Write \displaystyle w = 1 - \frac{x}{a}. Then T is z-simple over the triangle 0 \le x \le a, 0 \le y \le bw, between z = 0 and \displaystyle z = c\left(w - \frac{y}{b}\right). Two integrals do the work: with \displaystyle s = w - \frac{y}{b}, for p \ge 0,
and \displaystyle \int_0^a \left(1 - \frac{x}{a}\right)^q dx = \frac{a}{q + 1} for q \ge 0. The inner integral of M_{xy} is \displaystyle \int_0^{c(w - y/b)} z\,dz = \frac{c^2}{2}\left(w - \frac{y}{b}\right)^2, so by Corollary 5 and Theorem 4
So \displaystyle \bar{z} = \frac{abc^2/24}{abc/6} = \frac{c}{4}. The description of T is unchanged when the pairs (x, a) and (z, c) are exchanged, so the same computation with x innermost gives \displaystyle M_{yz} = \frac{a^2 bc}{24} and \displaystyle \bar{x} = \frac{a}{4}; exchanging (y, b) and (z, c) gives \displaystyle \bar{y} = \frac{b}{4}. These are the averages of the vertices’ coordinates. \blacksquare
Boundaries
Not every solid is simple. The shell 1 \le x^2 + y^2 + z^2 \le 4 meets the z-axis in two segments, so it is not z-simple, and by symmetry it is simple in no direction. Its upper half splits into a z-simple piece over the disk x^2 + y^2 \le 1, between \displaystyle z = \sqrt{1 - x^2 - y^2} and \displaystyle z = \sqrt{4 - x^2 - y^2}, and one over the ring 1 \le x^2 + y^2 \le 4, between z = 0 and \displaystyle z = \sqrt{4 - x^2 - y^2}; additivity adds the pieces. In the spherical coordinates of Cylindrical and spherical coordinates the shell is a box.
Bounded integrands. An integrand that blows up, such as (xyz)^{-1/2} near the coordinate planes, gives an improper integral, a limit over solids that avoid the trouble (Exercise 15). Swapping orders there needs the integral of \lvert f \rvert finite (the Fubini and Tonelli theorems of measure theory, beyond this course). Without integrability orders can disagree: the first example in the Boundaries of Iterated integrals and Fubini's theorem, with a third variable the integrand ignores, gives \displaystyle \frac{\pi}{4} and \displaystyle -\frac{\pi}{4}.
More variables. Boxes in \mathbb{R}^n, Riemann sums and Theorem 3 go through for every n, the proof repeated once per variable.
Where this goes next
Cylindrical and spherical coordinates names what Example 4 did, z first and then polar coordinates, with dV = r\,dz\,dr\,d\theta, and adds spherical coordinates. Change of variables and the Jacobian brings in a Jacobian determinant and carries Example 6 to every tetrahedron. The divergence theorem turns a flux through a closed surface into a triple integral. Joint densities of three random variables are integrated over solids exactly as here (Joint distributions).
Exercises
Boxes
- Evaluate \iiint_B (x + 2y + 3z)\,dV over B = [0, 1] \times [0, 2] \times [0, 3].
- Evaluate \iiint_B x y^2 z^3\,dV over B = [0, 2] \times [0, 3] \times [0, 1].
- Find the average value of x^2 + y^2 + z^2 over the cube [0, a] \times [0, a] \times [0, a].
Solids between surfaces
- Evaluate \int_0^1 \int_0^x \int_0^{x + y} x\,dz\,dy\,dx.
- Evaluate \displaystyle \int_1^2 \int_0^z \int_0^y \frac{x}{z^4}\,dx\,dy\,dz.
- Find the volume of the solid 0 \le x \le z \le 4 - y^2.
- Find the volume of the solid bounded by the paraboloid x = y^2 + z^2 and the plane x = 4, integrating in x first.
Order of integration
- Write \int_0^1 \int_0^{1 - x} \int_0^{1 - x - y} f\,dz\,dy\,dx as an iterated integral with x innermost and z outermost.
- Write \int_0^1 \int_0^{x^2} \int_0^{y} f\,dz\,dy\,dx as an iterated integral with y innermost and x outermost.
- Evaluate \displaystyle \int_0^2 \int_0^1 \int_{\sqrt{y}}^{1} z\sqrt{1 + x^3}\,dx\,dy\,dz.
Mass, centroids and averages
- Andre models a wind slab as the box 0 \le x \le 10, 0 \le y \le 6, 0 \le z \le 0.5 in meters, with density \delta = 120 + 160z kilograms per cubic meter. Find its mass and the height \bar{z} of its center of mass.
- Find the centroid of the solid 0 \le z \le 1 - x^2 - y^2.
- Find the average value of xyz over the tetrahedron x, y, z \ge 0, x + y + z \le 1.
- The cube [0, a] \times [0, a] \times [0, a] has constant density and mass M. Find its moment of inertia I_z about the z-axis in terms of M and a.
- Find \displaystyle \lim_{\varepsilon \to 0^+} \iiint_{[\varepsilon, 1]^3} \frac{dV}{\sqrt{xyz}}.
Proofs
- Let g, h, k be continuous on [a_1, b_1], [a_2, b_2], [a_3, b_3], and B the box they form. Let G = \int_{a_1}^{b_1} g\,dx, H = \int_{a_2}^{b_2} h\,dy and K = \int_{a_3}^{b_3} k\,dz. Prove that \iiint_B g(x)h(y)k(z)\,dV = GHK.
- Prove that the ball x^2 + y^2 + z^2 \le a^2 has volume \displaystyle \frac43\pi a^3, using Corollary 5 and polar coordinates.
- Let E be z-simple over D with u_1 = -u_2, and let f be continuous on E with f(x, y, -z) = -f(x, y, z). Prove that \iiint_E f\,dV = 0.
- Let f be continuous on B = [a_1, b_1] \times [a_2, b_2] \times [a_3, b_3]. For a_1 < t \le b_1 let \Phi(t) be the integral of f over [a_1, t] \times [a_2, b_2] \times [a_3, b_3], and let \Phi(a_1) = 0. Prove that \Phi'(t) = \int_{a_2}^{b_2} \int_{a_3}^{b_3} f(t, y, z)\,dz\,dy.
Answers
- 42. Integrating in z, then y, then x gives \displaystyle 3x + 6y + \frac{27}{2}, then 6x + 39, then 3 + 39.
- \displaystyle \frac92. The integrand separates (Exercise 16), so the integral is the product of \int_0^2 x\,dx = 2, \int_0^3 y^2\,dy = 9 and \displaystyle \int_0^1 z^3\,dz = \frac14.
- a^2. The volume is a^3, and \displaystyle \iiint x^2\,dV = \frac{a^3}{3} \cdot a^2, the same for y^2 and z^2, so the integral is a^5.
- \displaystyle \frac38. The inner integral is x(x + y), then \displaystyle \int_0^x (x^2 + xy)\,dy = \frac32 x^3, and \displaystyle \int_0^1 \frac32 x^3\,dx = \frac38.
- \displaystyle \frac{\ln 2}{6}. The inner integral is \displaystyle \frac{y^2}{2z^4}, the middle one \displaystyle \frac{1}{6z}, and \displaystyle \int_1^2 \frac{dz}{6z} = \frac{\ln 2}{6}.
- \displaystyle \frac{256}{15}. The solid is x-simple, 0 \le x \le z, over -2 \le y \le 2, 0 \le z \le 4 - y^2. So \displaystyle V = \frac12\int_{-2}^{2} (4 - y^2)^2\,dy, which is \displaystyle \frac12\left(64 - \frac{128}{3} + \frac{64}{5}\right).
- 8\pi. Over the disk y^2 + z^2 \le 4, x runs from y^2 + z^2 to 4. In polar coordinates of the yz-plane, V = 2\pi\int_0^2 (4 - r^2)\,r\,dr, which is 2\pi(8 - 4).
- \int_0^1 \int_0^{1 - z} \int_0^{1 - y - z} f\,dx\,dy\,dz. The solid is the tetrahedron x, y, z \ge 0, x + y + z \le 1; for fixed y, z \ge 0 with y + z \le 1, x runs from 0 to 1 - y - z.
- \int_0^1 \int_0^{x^2} \int_z^{x^2} f\,dy\,dz\,dx. The solid is 0 \le x \le 1, 0 \le z \le y \le x^2. For fixed x and z, y runs from z to x^2, and such y exist exactly when 0 \le z \le x^2.
- \displaystyle \frac49\left(2\sqrt{2} - 1\right) \approx 0.8126. For each z, the region 0 \le y \le 1, \displaystyle \sqrt{y} \le x \le 1 is also 0 \le x \le 1, 0 \le y \le x^2. Integrating in y first leaves \displaystyle \int_0^1 x^2\sqrt{1 + x^3}\,dx, which is \displaystyle \frac29\left(2\sqrt{2} - 1\right) by u = 1 + x^3, and \int_0^2 z\,dz = 2.
- M = 4800 kilograms and \displaystyle \bar{z} = \frac{13}{48} \approx 0.271 meters. The base has area 60, so M = 60\int_0^{0.5} (120 + 160z)\,dz, which is 60(60 + 20). Likewise \displaystyle M_{xy} = 60\left(15 + \frac{20}{3}\right) = 1300.
- \displaystyle \left(0, 0, \frac13\right). Over the disk r \le 1, \displaystyle V = 2\pi\int_0^1 (1 - r^2)\,r\,dr = \frac{\pi}{2} and \displaystyle M_{xy} = \pi\int_0^1 (1 - r^2)^2\,r\,dr = \frac{\pi}{6}. The solid is symmetric under x \mapsto -x and y \mapsto -y, so \bar{x} = \bar{y} = 0.
- \displaystyle \frac{1}{120}. With w = 1 - x, the z-integral is \displaystyle \frac12 xy(w - y)^2, and \displaystyle \int_0^w y(w - y)^2\,dy = \frac{w^4}{12}, so the integral of xyz is \displaystyle \frac{1}{24}\int_0^1 x(1 - x)^4\,dx, which is \displaystyle \frac{1}{24} \cdot \frac{1}{30} = \frac{1}{720}. The volume is \displaystyle \frac16 (Example 6).
- \displaystyle I_z = \frac23 M a^2. With density k, \displaystyle \iiint x^2\,dV = \frac{a^5}{3} and the same for y^2, so \displaystyle I_z = \frac23 k a^5, and M = k a^3.
- 8. On [\varepsilon, 1]^3 the integrand is continuous and separates (Exercise 16), giving \left(\int_{\varepsilon}^1 x^{-1/2}\,dx\right)^3, which is \displaystyle \left(2 - 2\sqrt{\varepsilon}\right)^3 and tends to 8.
- The integrand is continuous on B, so by Theorem 3 (b) the integral is \int_{a_1}^{b_1} \int_{a_2}^{b_2} \int_{a_3}^{b_3} g(x)h(y)k(z)\,dz\,dy\,dx. In the inner integral g(x)h(y) is a constant factor, so it equals g(x)h(y)K. In the middle integral g(x)K is constant, giving g(x)HK, and the outer integral gives GHK. \blacksquare
- The ball is z-simple over the closed disk D: x^2 + y^2 \le a^2, between \displaystyle \mp\sqrt{a^2 - x^2 - y^2}, both continuous on D. By Corollary 5, \displaystyle V = \iint_D 2\sqrt{a^2 - x^2 - y^2}\,dA. The disk is the polar rectangle 0 \le r \le a, 0 \le \theta \le 2\pi, so by Theorem 2 of Double integrals in polar coordinates \displaystyle V = 2\pi\int_0^a 2\sqrt{a^2 - r^2}\,r\,dr. An antiderivative of \displaystyle 2r\sqrt{a^2 - r^2} is \displaystyle -\frac23\left(a^2 - r^2\right)^{3/2}, so \displaystyle V = 2\pi \cdot \frac23 a^3 = \frac43\pi a^3. \blacksquare
- Since u_1 \le u_2 and u_1 = -u_2, u_2 \ge 0. By Theorem 4 the integral is \iint_D F\,dA with F(x, y) = \int_{-u_2}^{u_2} f(x, y, z)\,dz. Split at 0 and put z = -s in the left piece: it becomes \int_0^{u_2} f(x, y, -s)\,ds, which is -\int_0^{u_2} f(x, y, s)\,ds. The pieces cancel, so F = 0 on D and the integral is 0. \blacksquare
- Let G(x) = \int_{a_2}^{b_2} \int_{a_3}^{b_3} f(x, y, z)\,dz\,dy. Theorem 3 (b) on [a_1, t] \times [a_2, b_2] \times [a_3, b_3] with x outermost gives \Phi(t) = \int_{a_1}^{t} G(x)\,dx, also at t = a_1. G is continuous: given \varepsilon > 0, take \delta from uniform continuity of f on B. If \lvert x - x' \rvert < \delta, then \lvert f(x, y, z) - f(x', y, z) \rvert < \varepsilon for all y, z. Integrating in z and then y gives \lvert G(x) - G(x') \rvert \le \varepsilon\,(b_2 - a_2)(b_3 - a_3). By the fundamental theorem of calculus, \Phi'(t) = G(t), one-sided at the endpoints. \blacksquare