Jacobian (Theorem # 34)
Theorem
Let the transformation from n-dimensional spherical coordinates $(r, \theta^1, \dots, \theta^{n-1})$ to n-dimensional Cartesian coordinates $(x^1, \dots, x^n)$, for $n \ge 2$, be given by the equations:
\begin{align*}
x^1 &= r\cos\theta^1 \\
x^2 &= r\sin\theta^1\cos\theta^2 \\
&\vdots \\
x^{n-1} &= r\left(\prod_{j=1}^{n-2}\sin\theta^j\right)\cos\theta^{n-1} \\
x^n &= r\prod_{j=1}^{n-1}\sin\theta^j
\end{align*}
The Jacobian determinant of this transformation, $J_n$, is given by the formula:
\begin{align*} J_n(r, \theta^1, \dots, \theta^{n-1}) = r^{n-1}\sin^{n-2}(\theta^1)\sin^{n-3}(\theta^2)\cdots\sin(\theta^{n-2})\end{align*}
This can be written more compactly as:
\begin{align*}J_n = r^{n-1}\prod_{j=1}^{n-2}\sin^{n-j-1}(\theta^j)\end{align*}
Calculus
Multivariable Calculus
Discussion
The Jacobian of the change of variables from $n$-dimensional spherical coordinates $(r, \theta_1, \dots, \theta_{n-1})$ to Cartesian coordinates $(x_1, \dots, x_n)$ takes the closed form
\begin{align*}
J_n = r^{n-1}\prod_{j=1}^{n-2}\sin^{n-j-1}(\theta_j).
\end{align*}
This formula generalises the familiar area element $r\,dr\,d\theta$ in the plane and the volume element $r^2 \sin\phi\,dr\,d\theta\,d\phi$ in three dimensions. It is essential for evaluating integrals over spheres and balls in $\mathbb{R}^n$, and appears throughout harmonic analysis, potential theory, and probability (e.g. in computing the volume of the $n$-ball or the surface area of $S^{n-1}$). The proof, due to Stephan Kolassa, proceeds by induction: the [determinant](/page/Derivative) is expanded by cofactors, common factors of $\sin\theta_1$ are extracted, and the Pythagorean identity collapses the result to a clean recurrence $J_n = r\sin^{n-2}(\theta_1)\,J_{n-1}$.
Proof
[proofplan]
We prove by induction on $n$ that the [Jacobian](/page/Derivative) determinant $J_n$ of the transformation from $n$-dimensional spherical coordinates to Cartesian coordinates satisfies $J_n = r^{n-1}\prod_{j=1}^{n-2}\sin^{n-j-1}(\theta_j)$. The base cases $n = 2$ and $n = 3$ are computed directly. For the inductive step, we expand the $n \times n$ determinant by cofactors along the first row, factor common terms from the resulting $(n-1) \times (n-1)$ minors, recognise each minor as $J_{n-1}$, and simplify using the Pythagorean identity $\cos^2\theta_1 + \sin^2\theta_1 = 1$.
[/proofplan]
[step:Set up notation and write the Jacobian matrix]
For brevity, write $c_j := \cos\theta_j$ and $s_j := \sin\theta_j$. The transformation from [spherical coordinates](/page/Spherical%20Integrals%20and%20Radial%20Functions) $(r, \theta_1, \dots, \theta_{n-1})$ to Cartesian coordinates $(x_1, \dots, x_n)$ is given by:
\begin{align*}
x_1 &= r\,c_1, \\
x_2 &= r\,s_1\,c_2, \\
&\;\;\vdots \\
x_{n-1} &= r\left(\prod_{j=1}^{n-2}s_j\right)c_{n-1}, \\
x_n &= r\prod_{j=1}^{n-1}s_j.
\end{align*}
The Jacobian [determinant](/page/Derivative) is:
\begin{align*}
J_n(r, \theta_1, \dots, \theta_{n-1}) := \det\left(\frac{\partial x_i}{\partial r} \;\bigg|\; \frac{\partial x_i}{\partial \theta_j}\right)_{\substack{1 \le i \le n \\ 1 \le j \le n-1}}.
\end{align*}
[/step]
[step:Expand the determinant by cofactors along the first row]
The first row of the Jacobian matrix is $(c_1, -rs_1, 0, \dots, 0)$, since $\partial x_1/\partial r = c_1$, $\partial x_1/\partial \theta_1 = -rs_1$, and $\partial x_1/\partial \theta_j = 0$ for $j > 1$. Expanding by cofactors along the first row:
\begin{align*}
J_n = c_1 \det A + r\,s_1 \det B,
\end{align*}
where $A$ is the $(n-1) \times (n-1)$ minor obtained by deleting row 1 and column 1, and $B$ is the $(n-1) \times (n-1)$ minor obtained by deleting row 1 and column 2 (with the appropriate sign from cofactor expansion absorbed).
[guided]
The key observation is that the first row has only two non-zero entries: $c_1$ in column 1 and $-rs_1$ in column 2. All remaining entries in row 1 are zero because $x_1 = r\,c_1$ depends only on $r$ and $\theta_1$, not on $\theta_2, \dots, \theta_{n-1}$.
Cofactor expansion along row 1 gives:
\begin{align*}
J_n = c_1 \cdot (-1)^{1+1} \det A + (-rs_1) \cdot (-1)^{1+2} \det B = c_1 \det A + rs_1 \det B,
\end{align*}
where $A$ is the $(n-1) \times (n-1)$ submatrix obtained by deleting row 1 and column 1, and $B$ is the submatrix obtained by deleting row 1 and column 2. The sign on the second term simplifies: $(-rs_1)(-1) = rs_1$.
[/guided]
[/step]
[step:Factor common terms from $\det A$ and $\det B$ to recognise $J_{n-1}$]
In the minor $A$, every entry in column 1 contains a factor of $rc_1$, and every entry in columns $2, \dots, n-1$ contains a factor of $s_1$. Factoring these out:
\begin{align*}
\det A = rc_1\,s_1^{n-2} \det A',
\end{align*}
where $A'$ is the $(n-1) \times (n-1)$ matrix with the same structure as the Jacobian matrix for $n-1$ dimensions with variables $(r, \theta_2, \dots, \theta_{n-1})$. That is, $\det A' = J_{n-1}(r, \theta_2, \dots, \theta_{n-1})$.
Similarly, in the minor $B$, every column contains a factor of $s_1$. Factoring:
\begin{align*}
\det B = s_1^{n-1} \det B',
\end{align*}
where $\det B' = J_{n-1}(r, \theta_2, \dots, \theta_{n-1})$ by the same structural recognition.
[guided]
Consider the minor $A$. Its rows correspond to $x_2, \dots, x_n$ and its columns correspond to $\theta_1, \dots, \theta_{n-1}$. Each $x_i$ for $i \ge 2$ has the form $x_i = r \cdot s_1 \cdot (\text{product of } c_j, s_j \text{ for } j \ge 2)$. The derivative $\partial x_i / \partial r$ always carries the factor $c_1 \cdot (\text{angular part})$ from differentiating the original formula, while $\partial x_i / \partial \theta_j$ for $j \ge 2$ carries a factor $s_1$.
More precisely, examining column 1 of $A$ (derivatives with respect to $r$): every entry has a factor $c_1$ since $\partial x_i/\partial r$ evaluated at $r$ gives the angular part with $c_1$ from the original Jacobian structure. Columns $2, \dots, n-1$ of $A$ (derivatives with respect to $\theta_2, \dots, \theta_{n-1}$): every entry has a factor $s_1$ since each $x_i$ for $i \ge 2$ contains $s_1$ as a factor, which persists under differentiation with respect to $\theta_j$ for $j \ge 2$.
Factoring $rc_1$ from column 1 and $s_1$ from each of the remaining $n-2$ columns:
\begin{align*}
\det A = rc_1 \cdot s_1^{n-2} \cdot J_{n-1}(r, \theta_2, \dots, \theta_{n-1}),
\end{align*}
where $J_{n-1}$ is the Jacobian determinant for $(n-1)$-dimensional spherical coordinates. The remaining matrix after factoring has the same structure as the Jacobian for the transformation $(r, \theta_2, \dots, \theta_{n-1}) \mapsto (x_2', \dots, x_n')$ where $x_i' = x_i / (s_1)$.
For $\det B$: every column of $B$ contains a factor of $s_1$ (column 1 because each $x_i$ for $i \ge 2$ contains $s_1$ as a factor of $\partial x_i / \partial r$, and columns $2, \dots, n-1$ for the same reason as above). Factoring:
\begin{align*}
\det B = s_1^{n-1} \cdot J_{n-1}(r, \theta_2, \dots, \theta_{n-1}).
\end{align*}
[/guided]
[/step]
[step:Combine using the Pythagorean identity to obtain the recurrence]
Substituting the factored expressions:
\begin{align*}
J_n &= c_1 \cdot rc_1\,s_1^{n-2}\,J_{n-1} + rs_1 \cdot s_1^{n-1}\,J_{n-1} \\
&= rc_1^2\,s_1^{n-2}\,J_{n-1} + rs_1^n\,J_{n-1} \\
&= rs_1^{n-2}\,J_{n-1}\left(c_1^2 + s_1^2\right) \\
&= rs_1^{n-2}\,J_{n-1},
\end{align*}
where we used the Pythagorean identity $c_1^2 + s_1^2 = \cos^2\theta_1 + \sin^2\theta_1 = 1$.
[/step]
[step:Verify the base cases and complete the induction]
**Base case $n = 2$:** The transformation is $x_1 = r\cos\theta_1$, $x_2 = r\sin\theta_1$. The Jacobian matrix is:
\begin{align*}
\begin{pmatrix} \cos\theta_1 & -r\sin\theta_1 \\ \sin\theta_1 & r\cos\theta_1 \end{pmatrix},
\end{align*}
with determinant $J_2 = r(\cos^2\theta_1 + \sin^2\theta_1) = r$. This matches $r^{2-1} \cdot (\text{empty product}) = r$.
**Base case $n = 3$:** Direct computation gives $J_3 = r^2 \sin\theta_1$, matching $r^{3-1}\sin^{3-2}\theta_1 = r^2\sin\theta_1$.
**Inductive step:** Assume $J_{n-1}(r, \theta_2, \dots, \theta_{n-1}) = r^{n-2}\prod_{j=2}^{n-2}\sin^{n-j-1}(\theta_j)$. By the recurrence:
\begin{align*}
J_n &= r\,\sin^{n-2}(\theta_1)\,J_{n-1}(r, \theta_2, \dots, \theta_{n-1}) \\
&= r \cdot \sin^{n-2}(\theta_1) \cdot r^{n-2}\prod_{j=2}^{n-2}\sin^{n-j-1}(\theta_j) \\
&= r^{n-1}\sin^{n-2}(\theta_1)\prod_{j=2}^{n-2}\sin^{n-j-1}(\theta_j) \\
&= r^{n-1}\prod_{j=1}^{n-2}\sin^{n-j-1}(\theta_j).
\end{align*}
This completes the induction.
[guided]
Let us verify that the recurrence $J_n = r\,s_1^{n-2}\,J_{n-1}$ together with the base cases yields the claimed closed-form formula.
**Base case $n = 2$:** The Jacobian matrix is $2 \times 2$:
\begin{align*}
\begin{pmatrix} \cos\theta_1 & -r\sin\theta_1 \\ \sin\theta_1 & r\cos\theta_1 \end{pmatrix}.
\end{align*}
The determinant is $r\cos^2\theta_1 + r\sin^2\theta_1 = r$. The claimed formula gives $r^{2-1} \cdot \prod_{j=1}^{0}(\cdots) = r \cdot 1 = r$, since the empty product equals $1$. The base case holds.
**Base case $n = 3$:** Using the recurrence, $J_3 = r\,\sin^1(\theta_1)\,J_2 = r\sin\theta_1 \cdot r = r^2\sin\theta_1$. The claimed formula gives $r^2\sin^{3-1-1}(\theta_1) = r^2\sin\theta_1$. The base case holds.
**Inductive step:** Suppose the formula holds for $n-1$, i.e., $J_{n-1}(r, \theta_2, \dots, \theta_{n-1}) = r^{n-2}\prod_{j=2}^{n-2}\sin^{n-j-1}(\theta_j)$. Note that the product starts at $j = 2$ because the $(n-1)$-dimensional spherical coordinate system uses angles $\theta_2, \dots, \theta_{n-1}$, and the exponent of $\sin\theta_j$ in $J_{n-1}$ is $(n-1) - (j-1) - 1 = n - j - 1$. Applying the recurrence:
\begin{align*}
J_n &= r \cdot \sin^{n-2}(\theta_1) \cdot r^{n-2}\prod_{j=2}^{n-2}\sin^{n-j-1}(\theta_j) = r^{n-1}\prod_{j=1}^{n-2}\sin^{n-j-1}(\theta_j),
\end{align*}
where we merged $\sin^{n-2}(\theta_1) = \sin^{n-1-1}(\theta_1)$ into the product as the $j = 1$ term. This completes the induction and proves the formula for all $n \ge 2$.
[/guided]
[/step]
Prerequisites (0/2 completed)
Prerequisites Graph
Interactive dependency map showing how this theorem builds on foundational concepts
Loading dependency graph...
Theorem
Definition
Current
Requires
Definitions & Concepts