Change-of-variables formula

1 Foundations and Intuition

1.1 Why substitution works

The change-of-variables formula explains how an integral behaves when the variables used to describe a quantity are replaced by new ones. Intuitively, when a transformation re-parameterizes a curve, region, or volume, the integrand must be adjusted to reflect how “small pieces” in the original description correspond to small pieces in the new description. The result is that an integral can be computed in whichever coordinate system makes the integrand and region simplest.

At its core, the formula converts a computation of an accumulated quantity into an equivalent one, preserving the overall total while accounting for local distortions caused by the transformation.

1.2 Scaling factors and “stretching” under mappings

A mapping generally changes lengths, areas, and volumes locally. In one dimension, the derivative measures how the input variable stretches or compresses infinitesimal segments. In higher dimensions, the Jacobian determinant plays a similar role: it captures how the transformation scales oriented volume elements. The absolute value of this scaling factor is what appears when one is concerned with ordinary area or volume rather than orientation.

This “scaling factor” idea is what makes substitution valid: the integrand is evaluated at the new variable, and the contribution from each infinitesimal piece is multiplied by the factor that converts the old infinitesimal measure into the new one.

1.3 Integrating with respect to new variables

In practice, the formula provides a procedure:

1. Express the old variables in terms of new ones (or vice versa).
2. Replace the integrand using the substitution.
3. Multiply by the appropriate Jacobian factor.
4. Transform the region of integration (or limits) accordingly.

When the region description becomes complicated, the formula still works as long as the mapping is well-behaved on the relevant set and the transformed region is correctly identified.

2 One-Variable Change of Variables

2.1 Basic substitution rule

2.1.1 Derivative-based form: \(x=g(u)\)

For an integral of the form \(\int f(x)\,dx\), a typical substitution is \(x=g(u)\). Under appropriate conditions, \[ \int_{a}^{b} f(x)\,dx=\int_{u_1}^{u_2} f(g(u))\,g'(u)\,du, \] where \(u_1\) and \(u_2\) correspond to \(x=a\) and \(x=b\) through the substitution \(x=g(u)\). This expresses that \(dx\) is replaced by \(g'(u)\,du\), reflecting the local stretching/compression of the variable.

2.1.2 Handling limits of integration

Limits transform via the relationship between \(x\) and \(u\). If \(x=g(u)\) is monotone on the interval of interest, the endpoints match cleanly: \(u_1\) is the value such that \(g(u_1)=a\), and \(u_2\) satisfies \(g(u_2)=b\). If the substitution is not monotone, the mapping may cover the target interval multiple times; in such cases, one often partitions the domain into intervals where the substitution behaves nicely.

2.1.3 Orientation and sign considerations

When \(g'(u)\) is negative on the interval, the orientation of the variable substitution reverses. This effect is naturally captured by the factor \(g'(u)\) in the formula. More generally, orientation issues arise when discussing signed integrals; for many geometric applications, one takes absolute values to obtain nonnegative area or volume contributions.

2.2 Typical examples and patterns

2.2.1 Trigonometric substitutions

Trigonometric substitutions are used to simplify expressions involving square roots of quadratic polynomials. A common pattern replaces a square root form with a trigonometric function so that the resulting algebra collapses. For instance, substitutions like \(x=\sin u\) or \(x=\tan u\) convert factors such as \(1-x^2\) or \(1+x^2\) into expressions involving \(\cos u\) or \(\sec u\), making the integral manageable.

2.2.2 Algebraic substitutions

Algebraic substitutions often target integrands where a composite structure appears, such as \(x^2\) inside a function. Typical examples involve setting \(u\) equal to the inner expression, so that \(du\) produces the remaining differential factor. This includes cases where the substitution is motivated by recognizing a derivative sitting inside the integrand, as well as cases requiring a simple rearrangement before substitution is applied.

2.2.3 Exponential/logarithmic substitutions

Integrals featuring \(e^{ax}\), \(\ln x\), or expressions like \(a^x\) often simplify using substitutions that linearize the exponential or logarithm. For example, when the integrand contains \(e^{x}\) times another function of \(e^x\), setting \(u=e^x\) turns the problem into an integral in \(u\). Similarly, when \(\ln x\) appears as part of a composition, setting \(u=\ln x\) can replace \(dx/x\) with \(du\), streamlining the integrand.

3 Multivariable Change of Variables

3.1 Mappings between regions

In multiple dimensions, the formula relates integrals over a region in one coordinate system to integrals over a corresponding region in another. One typically considers a transformation \[ (x,y)=(X(u,v),Y(u,v)) \] (or its higher-dimensional analog). The key requirement is that the mapping should be sufficiently smooth and behave regularly enough so that small area/volume elements scale in a controlled manner.

The transformation induces a correspondence between a “source” region in the \((u,v)\)-space and a “target” region in the \((x,y)\)-space. The Jacobian determinant determines how the measure changes under this correspondence.

3.2 The Jacobian determinant

3.2.1 Computing the Jacobian matrix

For a transformation \((u,v)\mapsto (x,y)\), the Jacobian matrix is \[ J=\begin{pmatrix} \frac{\partial x}{\partial u} & \frac{\partial x}{\partial v}\\[6pt] \frac{\partial y}{\partial u} & \frac{\partial y}{\partial v} \end{pmatrix}. \] Its determinant \(\det(J)\) quantifies the infinitesimal area scaling factor. In three dimensions, the Jacobian determinant is computed from the corresponding \(3\times 3\) matrix of partial derivatives.

In computations, one often writes the determinant explicitly, either by direct expansion or using structured formulas (such as triangular forms) that simplify evaluation.

3.2.2 Determinant interpretation

Geometrically, \(	\det(J)	\) represents how an infinitesimal patch in the \((u,v)\)-plane grows or shrinks when mapped to the \((x,y)\)-plane. The sign of \(\det(J)\) records whether the mapping preserves or reverses orientation. When computing ordinary integrals that represent total mass/area/volume, the absolute value is typically used.

3.3 Transforming integrals over regions

3.3.1 Formula for double integrals

Let \(\Omega\) be a region in the \((u,v)\)-plane and suppose \((x,y)=T(u,v)\) maps \(\Omega\) to a region \(R\) in the \((x,y)\)-plane. For a suitable function \(f\), \[

\iint_R f(x,y)\,dA=\iint_{\Omega} f(X(u,v),Y(u,v))\,\left	\det\left(\frac{\partial(x,y)}{\partial(u,v)}\right)\right	\,dudv.

\] This formula converts an integral over \(R\) into one over \(\Omega\), with the scaling factor supplied by the Jacobian determinant.

3.3.2 Formula for triple integrals

Similarly, for a transformation \((u,v,w)\mapsto (x,y,z)\) with mapping \(T\), a typical statement is \[

\iiint_R f(x,y,z)\,dV=\iiint_{\Omega} f(X(u,v,w),Y(u,v,w),Z(u,v,w))\,\left	\det\left(\frac{\partial(x,y,z)}{\partial(u,v,w)}\right)\right	\,dudvdw.

\] The determinant again measures local volume distortion.

3.3.3 Absolute value of the Jacobian

The appearance of \(	\det	\) reflects the fact that standard volume and area elements are nonnegative. If a problem is formulated in terms of oriented integrals, one may use \(\det\) without absolute value, but for most geometric applications—especially those computing physical quantities—the absolute value is the appropriate choice.

4 Coordinate Transformations

4.1 Polar and cylindrical coordinates

4.1.1 Deriving the area element factor

In polar coordinates, the planar transformation is \[ x=r\cos\theta,\quad y=r\sin\theta. \] The Jacobian determinant for \((r,\theta)\mapsto (x,y)\) yields the familiar area element: \[ dA = r\,dr\,d\theta. \] This factor arises because circles in the \((x,y)\)-plane correspond to lines of constant \(r\) and because the mapping stretches area by an amount proportional to the radius.

4.1.2 Common integral setups

Typical applications include integrals over disks and annular regions. A disk \(x^2+y^2\le R^2\) becomes \(0\le r\le R\) with \(0\le\theta\le 2\pi\). Rectangular regions in \((x,y)\) generally become more complicated in polar coordinates, so the choice depends on which description makes the bounds and integrand simpler.

4.2 Spherical coordinates

4.2.1 Deriving the volume element factor

Spherical coordinates use a radius and two angles to describe a point: \[ x=\rho\sin\phi\cos\theta,\quad y=\rho\sin\phi\sin\theta,\quad z=\rho\cos\phi. \] The Jacobian determinant yields the volume element \[ dV=\rho^2\sin\phi\,d\rho\,d\phi\,d\theta. \] The \(\rho^2\) factor reflects the scaling of spherical shells, while \(\sin\phi\) corresponds to how angular spacing changes with latitude.

4.2.2 Typical spherical-region integrals

Spherical balls \(\rho\le R\) are straightforward in these coordinates: \(0\le\rho\le R\), \(0\le\phi\le\pi\), \(0\le\theta\le 2\pi\). Integrals over regions like shells \(R_1\le \rho\le R_2\) or sectors defined by angular constraints also become direct once the bounds are translated into \(\rho,\phi,\theta\).

4.3 General curvilinear coordinates

4.3.1 Jacobians in nonstandard systems

For coordinate systems that do not align with Cartesian axes, the Jacobian still provides the correct scaling. The procedure is to compute the determinant of the derivative matrix for the coordinate transformation and insert its absolute value into the integral. In many curvilinear systems, intermediate expressions can be simplified by exploiting symmetries or by choosing coordinates that separate variables.

Even when the transformation is complicated, the formula remains conceptually consistent: integrals change by multiplying by the local volume/area scaling factor.

4.3.2 Practical workflow for transforming regions

A common workflow is:

1. Identify the mapping between old and new coordinates.
2. Compute the Jacobian determinant.
3. Translate the region boundaries into inequalities in the new variables.
4. Confirm that the bounds cover the region once (or account for multiple coverage by partitioning).
5. Evaluate the transformed integral.

This workflow emphasizes that the bounds are as important as the Jacobian; an incorrect region description invalidates the result even if the determinant is computed correctly.

5 Formal Conditions and Pitfalls

5.1 Regularity requirements (differentiability)

The change-of-variables formula relies on differentiability of the transformation to ensure the Jacobian determinant is defined and behaves predictably. Typically, one assumes at least continuous differentiability on the region of interest, or conditions that guarantee the mapping is smooth enough for the derivative-based scaling argument to hold.

If the transformation has points where derivatives fail to exist or where the mapping becomes singular, the formula may still apply in a generalized sense, but additional care is needed.

5.2 One-to-one vs. many-to-one mappings

A mapping may fail to be injective, meaning different points in the source region map to the same point in the target region. In such situations, the naive “single coverage” interpretation can break down. The integral may need to be expressed with careful partitioning so that each target element is accounted for correctly, or one uses more advanced versions of the theorem that incorporate multiplicity.

In elementary calculus applications, students usually restrict to transformations that are one-to-one (or use partitions) to avoid these complications.

5.3 Image of a region and coverage issues

Even when a transformation is smooth, the image of a region might not match the intended target region without overlap or gaps. For instance, a parametric description could trace part of a curve multiple times or miss boundary segments. Correctly identifying the set of points actually covered by the mapping is essential.

A practical method is to test the transformation on boundary points and verify that the transformed inequalities correspond to the desired geometric region.

5.4 Boundary behavior and measure-zero sets

Some subtleties involve sets where the Jacobian determinant becomes zero or where the mapping behaves irregularly. Often, such problematic sets can have measure zero (they occupy “no volume” in the integration sense), and they do not affect the value of the integral for many standard settings. Nevertheless, understanding whether singularities matter in a given problem is part of applying the theorem responsibly.

Boundary behavior also matters when limits are transformed; if the boundary contributes nontrivially (for example, in line integrals rather than area/volume integrals), the analysis must be adjusted accordingly.

6 Change of Variables in Probability (Optional Connection)

6.1 Distributions via transformations

Probability theory frequently uses the same scaling ideas to describe how random variables change under transformations. If a random variable is mapped through a function to produce a new variable, the distribution of outcomes transforms according to how the mapping stretches or compresses sets of values.

This approach is widely used because it provides a systematic way to derive probability densities for transformed variables.

6.2 Density transformation using Jacobians

For continuous random variables, if \((U,V)\) has a known joint density and \((X,Y)\) is obtained by a differentiable transformation, the joint density of \((X,Y)\) can be expressed in terms of the original density multiplied by the absolute Jacobian determinant (again, under conditions ensuring correct coverage). In one dimension, the derivative appears analogously.

Thus, the Jacobian determinant functions as a change-of-measure factor: it converts probabilities computed via infinitesimal intervals/areas in one variable system into those in another.

7 Applications and Techniques

7.1 Area and volume computation

A central application is computing geometric quantities defined by integrals. By choosing coordinates aligned with the symmetry of the region—such as polar for circular regions or spherical for radial ones—one can transform a difficult integral into a simpler expression with straightforward bounds and an appropriate Jacobian factor.

This method also extends to computing quantities like mass distributions in mechanics when the density is expressed in terms of coordinates.

7.2 Simplifying integrals by symmetry

Change of variables is often used as a symmetry tool. If an integrand depends on radial distance, angular differences, or other invariant combinations, coordinate systems that reflect those invariances can reduce the complexity of the expression. The Jacobian factor, while adding an extra term, may be outweighed by the simplification gained in the integrand.

This technique is common in evaluating integrals over disks, spheres, or regions invariant under rotations and reflections.

7.3 Evaluating integrals with complex integrands

Integrands that appear complicated in Cartesian form may become tractable after transformation. Examples include functions involving products like \(x^2+y^2\), which become a single variable in polar coordinates. Similarly, expressions involving \(\sqrt{x^2+y^2+z^2}\) often simplify in spherical coordinates.

The goal is to convert the integrand into a form that separates variables or aligns with known integral templates.

7.4 Numerical and computational considerations (conceptual)

While change-of-variables is primarily a symbolic tool, it also influences computation. In numerical integration, transforming variables can improve convergence or reduce oscillation by mapping irregular regions into simpler domains. Conceptually, the Jacobian still corrects for the change in measure, but computational methods may use it to generate more stable sampling strategies.

Care is required because numerical errors can be amplified if the Jacobian becomes very large or very small within the region.

8 Exercises and Worked Examples

8.1 Step-by-step substitution practice

A typical exercise provides an integral with an embedded composite function, asking for identification of the substitution and careful transformation of differentials and bounds. A good practice problem walks through:

• choosing \(u=g(x)\),
• rewriting the integrand in terms of \(u\),
• replacing \(dx\) using \(du=g'(x)\,dx\) (or the inverse),
• converting the limits consistently,
• and simplifying the resulting integral.

The objective is to build reliability in mapping every component of the original integral to its counterpart.

8.2 Jacobian computation drills

Jacobians are often tested through transformations such as

• linear transformations in two or three dimensions,
• nonlinear coordinate changes that produce structured determinants,
• and curvilinear systems like polar or cylindrical variants.

Drills usually emphasize determinant computation and interpretation: students must compute the determinant correctly, then decide whether to use its absolute value based on the integral’s geometric meaning.

8.3 Region-mapping and limit-setting problems

Region-mapping problems focus on translating inequalities under a transformation. Typical tasks include:

• determining the transformed bounds for a disk or annulus,
• describing how a rectangular box maps into a curvilinear region,
• finding the image of a region defined by inequalities in one coordinate system.

These problems train the skill of verifying coverage and ensuring boundaries are transformed without omission or duplication.