Question: Is there a general relationship between surface area and volume analogous to the below examples?

Example 1. Consider a ball $B$ centered at the origin of a spherical coordinate system. The ball has volume $V(r) = \frac{4}{3}\pi r^3$, where $r$ is the radial distance from the origin to the boundary $\partial B$ of the ball. Now consider the surface area of $\partial V$. Elementary geometry shows that $A(r) = 4 \pi r^2$. Comparing $A$ and $V$, it follows that

$$ \frac{\partial V}{\partial r} = A. $$

Example 2. Another simple example is an open cylinder of length $l$. Place the cylinder at the origin and then define $s$ as the radial distance from the origin (in cylindrical coordinates). The volume of the cylinder is $V(s) = \pi s^2 l$ and its surface area is $A(s) = 2 \pi s l$. Once again, we have a relationship between volume and surface area with reference to the variable of variation (e.g., $s$ is the variable of variation, since both the volume and cylinder share the common length $l$),

$$ \frac{\partial V}{\partial s} = A. $$

Example 3. Although a cube may seem to stray from this relationship, as $\partial_s (s^3) \neq 6s^2$, this is only due to the lack of symmetry. If we center the cube in a Cartesian coordinate system, and define $a$ to be half the side length $s$, we see that $V(a) = (2a)^3 = 8a^3$ and $A(a) = 6(2a)^2 = 24a^2$. Thus,

$$ \frac{\partial V}{\partial a} = A. $$

Now consider a region $Q \subset \mathbb{R}^3$ with boundary $\partial Q$. The "center of mass" is defined as

$$ \textbf{r}_{cm} = \frac{1}{V(Q)} \int_{Q} \textbf{r} dV, $$

where $\textbf{r} = (x, y, z)$ in Cartesian coordinates. Assume that the object is centered in a coordinate system such that $\textbf{r}_{cm} = \textbf{0}$. Now let a coordinate system be described by a basis $\{\textbf{x}_i\}$, where $i = 1, 2, 3$. In general, when will the following relationship hold between $V = V(Q)$ and $A(\partial Q)$,

$$ \frac{\partial V}{\partial x_i} = A? $$

Also, assuming this relationship holds in some coordinate system, is there only one associated region that satisfies the necessary conditions (e.g., a sphere in spherical coordinates, a cube in Cartesian coordinates)?

  • $\begingroup$ You may find this paper very interesting: math.byu.edu/~mdorff/docs/DorffPaper07.pdf $\endgroup$ – fixedp Aug 21 '14 at 9:08
  • $\begingroup$ What do you take $\partial V/\partial x_i$ to mean? Without some fixed pattern for forming different shapes $Q$, I find it hard to see how you would find $V$ in terms of the $x_i$. $\endgroup$ – Jessica B Aug 21 '14 at 9:20
  • $\begingroup$ There is a non-rigorous intuitive argument. If you take $\partial r$ to be a thickness that you add to each face of the object, then certainly $\partial V = A*\partial r$ because the incremental volume is a thin shell that has area $A$ and thickness $\partial r$, therefore volume $A*\partial r$. This seems clear for convex objects, anyway. $\endgroup$ – bubba Aug 21 '14 at 12:37

The answer is: YES - see Stokes Theorem and as special case Greens Theorem.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.