MULTIVARIABLE CALCULUS • MULTIPLE INTEGRALS

Change of Variables — Choose substitutions and regions for change of variables

Learn how clever substitutions transform complicated integrals into simpler ones by reshaping both the integrand and the region.

Historical Context & Motivation

The idea behind a change of variables is one of the oldest tricks in mathematics: when a problem looks hard in one coordinate system, switch to a different one where it looks easy. This technique has roots stretching back centuries, and it became especially powerful once mathematicians began tackling integrals over complicated two- and three-dimensional regions.

In single-variable calculus, you already know u-substitution — replacing a messy expression with a single variable u to simplify an integral. The multivariable version extends that idea to two or more dimensions, but now you must also track how the shape of the region changes under the substitution, not just the integrand.

🎯 Before You Start
This lesson assumes you can already: take partial derivatives of a multivariable function, evaluate a 2 × 2 determinant, and solve a small system of linear equations for x and y. If any of these feel shaky, it is worth a quick review before continuing — everything else in the lesson builds directly on these three skills.
1687
Newton's Principia
Isaac Newton uses geometric transformations to evaluate areas and volumes, laying the conceptual groundwork for coordinate transformations in calculus.
1770s
Euler & Polar Coordinates
Leonhard Euler systematically applies polar coordinates to simplify integrals over circular regions, demonstrating the power of choosing the right coordinate system.
1841
Jacobi Formalizes the Determinant
Carl Gustav Jacob Jacobi publishes his work on determinants of partial derivatives — the Jacobian — giving mathematicians a universal tool for change of variables in multiple integrals.
1900s
Modern Formulation
The change-of-variables theorem is rigorously proven and becomes a cornerstone of real analysis, differential geometry, physics, and engineering.

The central question this lesson addresses is: given a double (or triple) integral over a complicated region, how do you choose the right substitution and correctly describe the new region of integration so that the integral becomes manageable?

Core Principles & Definitions

Before diving into calculations, you need to understand the four pillars of a successful change of variables. These ideas work together: skip one, and the substitution will fail.

1

The Substitution Map

Define new variables (u, v) as functions of the old ones (x, y), or vice versa. The map must be one-to-one on the interior of the region so that every point in the new region corresponds to exactly one point in the old region.
2

The Jacobian Determinant

The Jacobian measures how much the substitution stretches or compresses area at each point. It replaces the simple du = f′(x) dx from single-variable u-sub.
3

Transformed Integrand

Rewrite the function being integrated entirely in terms of the new variables. The goal is to make the integrand simpler — ideally a constant or a separable function.
4

New Region of Integration

Map every boundary curve of the original region into the (u, v)-plane. The ideal substitution transforms a complicated region into a rectangle or other simple shape.
KEY TAKEAWAY
Think of a change of variables like using a GPS app to reroute around traffic. The destination (the value of the integral) stays the same, but the path you take (coordinate system) is much easier to follow. The Jacobian is the app's adjustment factor — it accounts for whether the new route is wider or narrower than the original road.

Visualizing the Transformation

The diagram below shows how a change of variables works geometrically. On the left is the original xy-region — a tilted parallelogram that would be painful to describe with ordinary x- and y-limits. On the right, after applying the substitution u = x + y and v = x − y, the parallelogram maps to a simple rectangle in the uv-plane. The integral over the rectangle is far easier to set up and evaluate.

Left: the original region R is a parallelogram in the xy-plane with vertices A, B, C, D. Right: under the substitution T(u, v), the region becomes a rectangle S in the uv-plane, making the limits of integration constant.

Notice how the four sides of the parallelogram (labeled A–B, B–C, C–D, D–A) each map to a straight side of the rectangle. This is the hallmark of a well-chosen substitution: curved or slanted boundaries become flat, constant-value boundaries. When each side of the new region is described by u = constant or v = constant, you can write the limits of integration without any algebra. The next section turns this picture into a precise formula, using the same idea of matching boundary equations to new variables.

Mathematical Framework

Building on the four pillars from Section 2 and the geometric picture from Section 3, we can now state the formal change-of-variables theorem for double integrals. If you have an invertible, continuously differentiable transformation from (u, v) to (x, y) — meaning it has a well-defined inverse and its partial derivatives exist and vary smoothly — then the double integral transforms according to a specific rule. Let's break it down piece by piece.

CHANGE OF VARIABLES FORMULA
∬_R f(x, y) dA = ∬_S f(x(u,v), y(u,v)) · |∂(x,y)/∂(u,v)| du dv
R = original region in the xy-plane; S = transformed region in the uv-plane; |∂(x,y)/∂(u,v)| = absolute value of the Jacobian determinant, which scales the area element.
JACOBIAN DETERMINANT (2D)
∂(x,y)/∂(u,v) = | ∂x/∂u ∂x/∂v | = (∂x/∂u)(∂y/∂v) − (∂x/∂v)(∂y/∂u) | ∂y/∂u ∂y/∂v |
This 2 × 2 determinant captures how infinitesimal areas in the (u, v)-plane are stretched or compressed when mapped to the (x, y)-plane. If the Jacobian equals 2, then each tiny square du × dv corresponds to an area of 2 du dv in the xy-plane.
POLAR COORDINATES EXAMPLE
x = r cos θ, y = r sin θ ⟹ |∂(x,y)/∂(r,θ)| = r
This is why the area element in polar coordinates is r dr dθ rather than just dr dθ. The factor of r is the Jacobian.
📋 Three-Step Checklist
When performing a change of variables: (1) Define the substitution — write x and y in terms of u and v. (2) Compute the Jacobian — take the determinant of the matrix of partial derivatives. (3) Transform the region — substitute each boundary equation to find the new limits.

How to Choose the Right Substitution

Choosing the substitution is often the hardest part. There is no single formula that works every time, but experienced mathematicians follow several reliable strategies. The diagram below categorizes the most common situations you will encounter, along with the substitution that fits each one.

This flowchart guides you through choosing a substitution based on the geometry of the region and the form of the integrand. Start by examining the boundary curves, then check whether the integrand suggests natural new variables.

The most common strategy is to let u and v be the expressions that appear in the boundary equations. For example, if the region is bounded by the lines x + y = 1, x + y = 3, x − y = 0, and x − y = 2, then setting u = x + y and v = x − y automatically turns those four boundaries into u = 1, u = 3, v = 0, and v = 2 — a rectangle in the uv-plane.

  1. Strategy 1 — Match the boundaries. Look at the equations of the boundary curves. If they share common expressions (like x + y or x² + y²), let those be your new variables.
  2. Strategy 2 — Match the integrand. If the function being integrated contains a complicated combination like (x − y)/(x + y), setting u = x + y and v = x − y simplifies both the function and (often) the region.
  3. Strategy 3 — Symmetry. If the region has circular symmetry, polar coordinates are almost always the best choice. If it has elliptical symmetry, scale one axis first.
  4. Strategy 4 — Simplify the Jacobian. After choosing u and v, compute the Jacobian. If it turns out to be extremely messy, try a different substitution — a good one usually yields a clean Jacobian.

Worked Example

Let's evaluate the integral ∬R (x − y) dA, where R is the parallelogram with vertices (0, 0), (2, 1), (3, 3), and (1, 2). We will choose a substitution, find the new region, compute the Jacobian, and evaluate the integral.

Evaluating a Double Integral via Change of Variables
1
Step 1 — Find the boundary line equationsThe four sides of the parallelogram lie along the lines: y = x/2 (from (0,0) to (2,1)), y = x/2 + 3/2 (from (1,2) to (3,3)), y = 2x (from (0,0) to (1,2)), and y = 2x − 3 (from (2,1) to (3,3)). Rewriting each equation in the form ax + by = c gives two pairs of parallel boundaries: x − 2y = 0 and x − 2y = −3 (the sides with slope 1/2), and 2x − y = 0 and 2x − y = 3 (the sides with slope 2). Summarizing: 2x − y = 0 and 2x − y = 3; x − 2y = 0 and x − 2y = −3.
2
Step 2 — Choose the substitutionSince the boundaries involve the expressions 2x − y and x − 2y, we set u = 2x − y and v = x − 2y. This maps the parallelogram to the rectangle 0 ≤ u ≤ 3, −3 ≤ v ≤ 0.
New region S: 0 ≤ u ≤ 3, −3 ≤ v ≤ 0
3
Step 3 — Solve for x and y in terms of u and vFrom u = 2x − y and v = x − 2y, we solve the system. Multiply the second equation by 2: 2v = 2x − 4y. Subtract from the first: u − 2v = −y + 4y = 3y, so y = (u − 2v)/3. Substitute back: v = x − 2(u − 2v)/3, so x = v + (2u − 4v)/3 = (3v + 2u − 4v)/3 = (2u − v)/3.
x = (2u − v)/3, y = (u − 2v)/3
4
Step 4 — Compute the Jacobian∂x/∂u = 2/3, ∂x/∂v = −1/3, ∂y/∂u = 1/3, ∂y/∂v = −2/3. The Jacobian is (2/3)(−2/3) − (−1/3)(1/3) = −4/9 + 1/9 = −3/9 = −1/3. We take the absolute value: |J| = 1/3.
|∂(x,y)/∂(u,v)| = 1/3
5
Step 5 — Transform the integrand and evaluateThe integrand x − y = (2u − v)/3 − (u − 2v)/3 = (2u − v − u + 2v)/3 = (u + v)/3. The integral becomes: ∬S [(u + v)/3] × (1/3) dA = (1/9) ∫₀³ ∫₋₃⁰ (u + v) dv du. Inner integral: ∫₋₃⁰ (u + v) dv = [uv + v²/2] from −3 to 0 = 0 − (−3u + 9/2) = 3u − 9/2. Outer integral: (1/9) ∫₀³ (3u − 9/2) du = (1/9)[3u²/2 − 9u/2] from 0 to 3 = (1/9)(27/2 − 27/2) = (1/9)(0) = 0.
∬_R (x − y) dA = 0
💡 Why zero?
The result of 0 makes geometric sense! The parallelogram is symmetric about the line y = x, and the function x − y is positive on one side and equally negative on the other. The positive and negative contributions cancel perfectly.

Common Substitutions Compared

Different problems call for different substitutions. The table below compares the most frequently used changes of variables, showing when each one shines and when it falls short.

Comparison of common changes of variables
SubstitutionBest ForJacobianLimitation
Polar (r, θ)Circular or annular regions; x² + y² in the integrandrDoesn't help with non-circular boundaries
Linear (u = ax+by, v = cx+dy)Parallelogram regions with straight-line boundariesConstant (easy!)Only works for linear boundaries
Scaled Polar (x=ar cos θ, y=br sin θ)Elliptical regions; (x/a)² + (y/b)² in the integrandabrOnly helps when the region truly has elliptical symmetry
Custom (u = g(x,y), v = h(x,y))Any region whose boundaries involve specific combinations of x and yVaries — must compute each timeRequires careful algebraic work to invert and compute |J|
KEY TAKEAWAY
A change of variables is like choosing the right wrench from a toolbox. Polar coordinates are the go-to wrench for circular problems, linear substitutions handle parallelogram-shaped regions, and custom substitutions are the adjustable wrench — flexible but requiring more effort. The right choice makes the job easy; the wrong one makes it impossible.

Connection to Triple Integrals & Advanced Theory

Everything you've learned about change of variables in double integrals extends naturally to triple integrals. The Jacobian becomes a 3 × 3 determinant, and the transformed region lives in three dimensions. The two most common 3D coordinate systems — cylindrical and spherical — are just specific changes of variables with their own Jacobians.

Double integrals vs. triple integrals in change of variables
FeatureDouble Integral (2D)Triple Integral (3D)
Variables(x, y) → (u, v)(x, y, z) → (u, v, w)
Jacobian size2 × 2 determinant3 × 3 determinant
Polar → ?Polar: |J| = rCylindrical: |J| = r; Spherical: |J| = ρ² sin φ
Region2D area in the plane3D volume in space
Key skillMap boundary curvesMap boundary surfaces

At a higher level, the change-of-variables theorem is connected to deep ideas in differential geometry and measure theory. The Jacobian determinant is really measuring how volumes transform under smooth maps — a concept that becomes essential in physics (general relativity), probability (multivariate distributions), and computer graphics (texture mapping). Mastering the 2D version now gives you a solid foundation for all of these applications.

Practice Problems

PROBLEM 1CONCEPTUAL
In your own words, explain why you need to multiply by the absolute value of the Jacobian determinant when performing a change of variables. What would go wrong if you forgot it?
PROBLEM 2BASIC CALCULATION
Let u = x + y and v = x − y. Compute the Jacobian ∂(x, y)/∂(u, v) and find |J|.
PROBLEM 3INTERMEDIATE
Evaluate ∬R e(x+y)/(x−y) dA, where R is the region bounded by x + y = 1, x + y = 4, x − y = 1, and x − y = 3. Use the substitution u = x + y, v = x − y.
PROBLEM 4APPLIED
A heat sensor measures temperature T(x, y) = x² + y² over a semicircular plate of radius 3 in the upper half-plane (y ≥ 0). Set up and evaluate the integral ∬ T dA using polar coordinates to find the total heat energy across the plate.
PROBLEM 5CRITICAL THINKING
Suppose you need to integrate f(x, y) = 1 over the region bounded by xy = 1, xy = 4, y/x = 1, and y/x = 3. Propose a substitution, find the Jacobian, transform the region, and set up (but do not evaluate) the resulting integral.

Lesson Summary

A change of variables transforms a difficult double integral into a simpler one by introducing new variables (u, v). The process has three essential steps: (1) choose a substitution that simplifies the region and the integrand — let the boundary equations guide your choice; (2) compute the Jacobian determinant, which tells you how area elements scale under the transformation; and (3) transform the region by substituting each boundary curve's equation into the new variables, ideally producing a rectangle.

Common substitutions include polar coordinates (|J| = r) for circular regions, linear substitutions for parallelogram regions, and scaled polar coordinates for elliptical regions. The best substitution makes both the integrand and the limits of integration simple. This technique extends naturally to triple integrals with 3 × 3 Jacobians and is fundamental to cylindrical and spherical coordinate systems.

Varsity Tutors • Multivariable Calculus • Change of Variables — Choose substitutions and regions for change of variables