A triple integral accumulates a quantity throughout a solid. Choosing cylindrical or spherical variables does not change the physical solid, but it changes its description and the volume scale. A mass calculation therefore begins with a density and an actual region; a centroid adds first moments; a moment of inertia adds squared distance to a specified axis. Each has different units and a different integrand.
From a solid to slices
Definition 1.4.1
(Mass and centroid)
For a bounded regular solid and continuous nonnegative density , define
provided .
The first moment in each coordinate uses the density again; dividing by the volume would instead give a geometric average only for an appropriate uniform density. If has units mass per volume, has units mass and each centroid coordinate has units length.
Cartesian integration can be organized by projecting the solid onto a coordinate plane and writing the interval of the remaining coordinate above each projected point. For example, a solid between and in the first octant has triangular projection , . Different orders require different slice bounds, and a slice may split into more than one interval. An iterated expression is correct only if its bounds cover the solid once.
Cylindrical and spherical scales
Cylindrical coordinates use
The factor is the planar polar area scale; changing adds an independent vertical displacement. The cylindrical radius is distance to the axis, not distance to the origin.
For spherical variables take
Here is distance to the origin, is measured from the positive axis, and is the azimuth. At a regular point, the radial, polar and azimuthal coordinate directions have lengths , and per unit coordinate increment; their orthogonality gives the stated absolute Jacobian. Different angle conventions use different symbols and ranges, so define them before substituting.
The origin, polar axis and angular seam are coordinate degeneracies. For a bounded continuous integrand on the solids below, they can be omitted and treated by a regular-coordinate exhaustion: the excluded neighborhoods have volume tending to zero, so boundedness makes their contributions tend to zero. This statement does not justify dropping an unbounded singularity without its own estimate. A full angular interval must cover almost every point once; a double traversal doubles a scalar integral.
Quick check
(Use spherical coordinates to find the volume of the solid ball .)
Write the bounds first, then multiply by the volume element .
Solution · Answer
The bounds are , , , so the volume is .
A cone with height-dependent density
Worked example 1.4.2
(Mass, centroid and inertia of a cone)
Let and
Find the mass, centroid and moment of inertia about the axis. Then find the moment about the parallel line .
At height , the circular cross-section has radius , so cylindrical bounds are , , . Multiplying density by the volume element gives
Symmetry under and gives . For the vertical first moment, use an additional factor :
The center lies inside the cone and above the uniform cone's height, consistent with density increasing with . These comparisons check plausibility; they are not substitutes for the moment calculation.
The squared perpendicular distance to the axis is , so
The centroid lies on the axis. By the parallel-axis identity proved below, the moment about the line is . If length has unit and mass unit , has units because ; then , and have units , and , respectively.
The parallel-axis identity, not a memorized correction
Definition 1.4.3
(Moment about a vertical line)
The moment of inertia about the line parallel to the axis through is
Proposition 1.4.4
(Parallel-axis identity)
Let and let be the horizontal centroid. Then
Proof
(Expanding about the centroid)
Let be the horizontal centroid. Write and similarly for . Expanding produces cross terms proportional to
These vanish by the definition of the centroid, not by an assumption of symmetry. The squared terms that remain are and the constant integrated against , which gives . Therefore
Proof X-ray
Where the cross terms go
The only non-routine move is the choice of expansion point: split at the centroid , not at the origin or at . Each cross term is then , and it vanishes because by the definition of . This needs and the same density in every integral. The height never enters, since the distance to a vertical line ignores it. Splitting at another point leaves the cross term , which does not vanish, so the identity holds only relative to the centroidal axis.
The axes must be parallel. The reference axis is the one through the centroid; adding to an arbitrary initial-axis moment is not generally correct. In the cone, , so the original axis is already the centroidal parallel axis.
Concept lensAnalytic
Moment of inertia as a mean square
Dividing the density by gives a probability density on . Then for a random point with that density, and is its mean. The parallel-axis identity is the variance decomposition , applied to each horizontal coordinate. The centroidal axis minimizes the mean square distance.
A radially weighted hemisphere
Worked example 1.4.5
(Mass, centroid and inertia of a hemisphere)
Let and density , where is spherical radius and . Find , and .
The bounds are , , . Density is , not : the two radii differ away from the equatorial plane. By symmetry , while
For the vertical first moment, , so
Thus . This exceeds the uniform hemisphere's centroid height : radial density weights outer shells more strongly. For the axis moment, , giving
Here . The mass, height and inertia have the correct powers for : , , . The dimensional check detects missing powers but cannot prove the numerical constants.
Common mistake
(Using the distance to the origin as the distance to the axis)
The cylindrical radius is the distance to the axis, whereas is the distance to the origin. A moment about the axis needs , not . In the hemisphere above, using would give , while the correct value is .
Quick check
(Find of the uniform solid ball of density in terms of its mass .)
Use and the volume element .
Solution · Answer
. Since , this is .
Symmetry with its hypotheses visible
Theorem 1.4.6
(Odd moments vanish under symmetry)
Let be a bounded regular solid and continuous with . If and are both invariant under , then , so .
Proof
(Reflection reverses the sign)
The reflection maps to itself and preserves volume and . Substituting it into the integral gives , so the integral equals its own negative and is zero.
Counterexample mode
A symmetric shape with an asymmetric density
The hypothesis on cannot be dropped. Take the hemisphere with , which is nonnegative on . The shape is invariant under , so , but . With and , hence , we get and . Therefore : the symmetry argument needs both the region and the density to be invariant.
Similarly, symmetry can show two centroid coordinates agree without determining their common value. After using the symmetry, calculate the remaining scalar integral and check that the result lies within the coordinate range of the solid. These checks prevent a correct-looking formula from being applied to a different physical model.
Cones, spheres and the order of limits
A cone with is a family of disks whose radii grow with height. A sphere with is a family of rays whose endpoint is fixed in spherical variables. Either can be described in more than one coordinate system, but the bounds must transform with the variables. For the cone, integrating in the order would require and ; keeping the original range together with the old upper bound after swapping the differentials would not describe a valid iterated integral.
All integrals in the worked examples are proper and have bounded continuous densities. If density is singular, a coordinate Jacobian may increase or decrease the apparent power of the radius, and convergence at the singularity must be proved before exchanging limits or using a formula designed for a bounded integrand. The measure factor is part of the convergence test, not an optional decoration.
Revision check. Write the physical distance to the selected axis before choosing coordinates. State the mass-density units and the nonzero mass condition for the centroid. Identify the region by inequalities, then calculate the determinant and choose a once-only angular range. Finally compare dimensions, feasible-coordinate ranges and any symmetry claim with the density actually given.
Reversing the cone's integration order
The first worked cone permits an independent calculation of its mass using vertical slices. For each , the inequality is equivalent to . Hence
This agrees with the horizontal-slice calculation. For the first moment, the same new bounds give
The lower vertical bound is a function of ; carrying over after reversing would add material outside the cone. Both orders describe a compact solid with continuous integrands, so the relevant proper integration theorem applies. Agreement of the two answers is a helpful independent check, but the algebraic equivalence of the bounds is what establishes that they describe the same solid.
Common mistake
(Keeping the old height range after reversing)
With and the region is the full cylinder, not the cone. The same mass integral then gives , exactly twice the correct mass .
A quick cylinder check for parallel axes
Worked example 1.4.7
(A uniform cylinder and a parallel axis)
For a uniform solid cylinder of radius , height and density , symmetry puts its centroidal axis on the axis. Its mass is , and direct integration yields
An axis parallel to it at horizontal distance has moment . The correction depends on the perpendicular distance between axes, not their distance from an arbitrary origin. Setting recovers , and the extra term is nonnegative, as expected for a shift away from a centroidal axis. These limiting checks are quick diagnostics, not replacements for the expansion proving the theorem.
Counterexample mode
Shifting from a non-centroidal axis
Suppose one moves from the axis , which touches the cylinder's edge, to the axis on the opposite side by adding with . The identity gives , and the reflection gives the same value for . The shortcut would instead give , which is wrong. The identity compares a centroidal axis with a parallel one; here both axes are non-centroidal.
Bounds that every nonnegative model must obey
In the cone, and . Multiplying each inequality by the nonnegative density and integrating gives and , provided . The computed ratio is , so it passes the second check; passes the first. Positive mass follows directly from the computed formula, or because the density is strictly positive on a nonempty interior region.
These bounds depend on nonnegative density. A signed algebraic weight can give a zero or negative total and need not obey the centroid bounds after division. It can be useful in other mathematical models, but it is not automatically a physical mass density. Check positivity and the axis before interpreting the numerical result.
Quick check
(A solid has density , with lengths in meters and mass in kilograms. What are the units of ?)
Compare the units of with those of .
Solution · Answer
A density has units and has units , so has units .