Symmetry and Centroid Simplification

Why Symmetry Simplifies a Centroid

A centroid is a mass-weighted balance point. In general, finding it requires calculating the total mass and two first moments. Reflection symmetry can simplify that work: contributions on opposite sides of an axis can cancel in one moment, while the other integrals can often be calculated over only half the region.

The symmetry must apply to the mass distribution, not merely to the shape. A geometrically symmetric region with an asymmetric density can have its centroid off the apparent axis of symmetry. The definitions and proof below make both requirements precise.

Mass, Moments, and Centroid

Let \(R\) be a bounded lamina in the plane with a nonnegative density function \(\rho(x,y)\). Assume its total mass is finite and positive and its first moments are finite. The total mass \(m\), moment about the \(x\)-axis \(M_x\), and moment about the \(y\)-axis \(M_y\) are

m=\iint_R \rho(x,y)\,dA,\qquad M_x=\iint_R y\rho(x,y)\,dA,\qquad M_y=\iint_R x\rho(x,y)\,dA.

The centroid is the point \((\bar{x},\bar{y})\) whose coordinates are the first moments divided by the total mass:

\bar{x}=\frac{M_y}{m}, \qquad \bar{y}=\frac{M_x}{m}.

The order of the subscripts can be easy to confuse: the moment about the \(y\)-axis uses the coordinate \(x\), and determines \(\bar{x}\); the moment about the \(x\)-axis uses \(y\), and determines \(\bar{y}\). If \(\rho\) is constant, the lamina is uniform, and the constant density cancels from both centroid ratios.

Definition: A lamina is symmetric about the \(y\)-axis when reflecting each point \((x,y)\) of its region to \((-x,y)\) leaves the region unchanged. It is symmetric about the \(x\)-axis when reflecting \((x,y)\) to \((x,-y)\) leaves the region unchanged. For a nonuniform lamina, symmetry about an axis also requires the density to be unchanged by that reflection.

The Symmetry Theorem

Theorem: If a lamina has finite positive mass and both its region and density are symmetric about the \(y\)-axis, then \(\bar{x}=0\), so its centroid lies on the \(y\)-axis. If both its region and density are symmetric about the \(x\)-axis, then \(\bar{y}=0\), so its centroid lies on the \(x\)-axis.

Here is a proof for the \(y\)-axis. The reflection sends \((x,y)\) to \((-x,y)\). Under that reflection, the region is unchanged and the density has the same value. The integrand for \(M_y\), however, changes sign because its factor \(x\) becomes \(-x\). Changing variables in the integral therefore gives

M_y =\iint_R x\rho(x,y)\,dA =\iint_R (-x)\rho(-x,y)\,dA =-\iint_R x\rho(x,y)\,dA =-M_y.

Thus \(M_y=-M_y\), which implies \(M_y=0\). Since \(m>0\), the centroid formula gives \(\bar{x}=M_y/m=0\). This proves that the centroid lies on the \(y\)-axis. For symmetry about the \(x\)-axis, reflection changes \(y\) to \(-y\), leaves the density and region unchanged, and changes the sign of the integrand \(y\rho(x,y)\) in \(M_x\). The same argument gives \(M_x=0\), hence \(\bar{y}=0\).

The proof also explains exactly what symmetry does not imply. Reflection about the \(y\)-axis cancels the \(x\)-coordinate moment, but does not generally make \(M_x\) vanish. The centroid is on the \(y\)-axis; it need not be at the origin. Similarly, symmetry about the \(x\)-axis determines \(\bar{y}\), not necessarily \(\bar{x}\).

Using a Representative Half-Region

When the region and density are symmetric about an axis, the two reflected halves have equal mass. For an integral whose integrand is unchanged by the reflection, the full-region integral is twice the integral over either half. For the first moment perpendicular to the axis of symmetry, the two halves instead cancel. The axis itself has area zero, so counting it in both halves does not affect a double integral.

Check the reflection.
Confirm that reflecting the region leaves it unchanged. For a nonuniform lamina, also check that the reflected density values agree.

Choose one half.
Describe the part of the region on one side of the symmetry axis, using slices that make its bounds manageable.

Use cancellation or doubling.
The first moment perpendicular to the symmetry axis is zero. For mass and integrands unchanged by reflection, integrate over the chosen half and multiply by two.

Form the centroid coordinates.
Divide the appropriate first moments by the total mass, then check that the coordinate on the symmetry axis agrees with the theorem.

Worked Examples

Worked Example: A Uniform Rectangle Symmetric About the \(y\)-Axis

Let \(R=[-2,2]\times[0,3]\) have constant density \(\rho=5\) units of mass per unit area. Find its centroid using symmetry and a representative half-region.

The rectangle is symmetric about the \(y\)-axis, and its density is constant, so the theorem gives \(\bar{x}=0\). Use the right half, \(0\leq x\leq2\), \(0\leq y\leq3\), and double its mass:

m=2\int_0^2\int_0^3 5\,dy\,dx =2\int_0^2 15\,dx =2(30) =60.

To find \(\bar{y}\), calculate \(M_x\). The integrand \(y\rho\) is unchanged by reflection across the \(y\)-axis, so the same representative half gives

M_x=2\int_0^2\int_0^3 5y\,dy\,dx =2\int_0^2\left[\frac{5y^2}{2}\right]_0^3dx =2\int_0^2\frac{45}{2}\,dx =90.

Therefore \(\bar{y}=M_x/m=90/60=3/2\). Together, \((\bar{x},\bar{y})=(0,3/2)\). As a check, the first moment about the \(y\)-axis vanishes: its integrand \(5x\) is odd under \(x\mapsto -x\), so \(M_y=0\) and \(\bar{x}=0\), as required.

Worked Example: The Centroid of an Upper Semicircular Lamina

Find the centroid of a uniform lamina occupying the upper half of the disk \(x^2+y^2\leq4\). The region is symmetric about the \(y\)-axis, so \(\bar{x}=0\). To calculate the area and the first moment about the \(x\)-axis, use horizontal slices. At height \(y\), where \(0\leq y\leq2\), the slice runs from \(x=-\sqrt{4-y^2}\) to \(x=\sqrt{4-y^2}\), and its width is \(2\sqrt{4-y^2}\). Write the constant density as \(\rho\).

The mass is density times area. The area integral evaluates as follows:

m=\rho\int_0^2 2\sqrt{4-y^2}\,dy =\rho\left(\frac{\pi(2)^2}{2}\right) =2\pi\rho.

The moment about the \(x\)-axis is

M_x =\rho\int_0^2 y\left(2\sqrt{4-y^2}\right)\,dy =2\rho\int_0^2 y\sqrt{4-y^2}\,dy.

Set \(u=4-y^2\), so \(du=-2y\,dy\). When \(y=0\), \(u=4\); when \(y=2\), \(u=0\). Hence

M_x =-\rho\int_4^0 \sqrt{u}\,du =\rho\int_0^4 u^{1/2}\,du =\rho\left[\frac{2}{3}u^{3/2}\right]_0^4 =\frac{16\rho}{3}.

Thus \(\bar{y}=M_x/m=(16\rho/3)/(2\pi\rho)=8/(3\pi)\), and the centroid is \((0,8/(3\pi))\). The density cancels, as it should for a uniform lamina. The \(x\)-coordinate is zero by reflection symmetry, and the positive \(y\)-coordinate is consistent with the lamina lying above the \(x\)-axis.

Worked Example: A Density Symmetric About the \(x\)-Axis

Let \(R=[0,3]\times[-1,1]\) have density \(\rho(x,y)=2+y^2\). Find its centroid. The rectangle and density are symmetric about the \(x\)-axis, so the theorem gives \(\bar{y}=0\). For mass and \(M_y\), integrate over the upper half and double, since both integrands are unchanged by \(y\mapsto-y\).

The mass is

m=2\int_0^3\int_0^1(2+y^2)\,dy\,dx =2\int_0^3\left[2y+\frac{y^3}{3}\right]_0^1dx =2\int_0^3\frac{7}{3}\,dx =14.

The moment about the \(y\)-axis is

M_y=2\int_0^3\int_0^1x(2+y^2)\,dy\,dx =2\int_0^3\frac{7x}{3}\,dx =\frac{14}{3}\left[\frac{x^2}{2}\right]_0^3 =21.

Therefore \(\bar{x}=M_y/m=21/14=3/2\), and the centroid is \((3/2,0)\). For a direct check of the symmetry result,

M_x=\int_0^3\int_{-1}^1y(2+y^2)\,dy\,dx=0,

because \(y(2+y^2)\) is odd in \(y\): its values at \(y\) and \(-y\) cancel. Consequently \(\bar{y}=M_x/m=0\), as predicted.

A Common Pitfall: The Shape Is Not the Whole Story

A symmetric region does not guarantee a centroid on its symmetry axis unless the density is symmetric too. For a concrete example, take \(R=[-1,1]\times[0,1]\), which is symmetric about the \(y\)-axis, and give it density \(\rho(x,y)=2+x\). This density is strictly positive on the whole region because \(1\leq2+x\leq3\), but it is not symmetric about the \(y\)-axis: \(2+x\) and \(2-x\) differ unless \(x=0\).

The mass and moment about the \(y\)-axis are

m=\int_0^1\int_{-1}^1(2+x)\,dx\,dy =\int_0^1 4\,dy =4, \qquad M_y=\int_0^1\int_{-1}^1x(2+x)\,dx\,dy =\int_0^1\frac{2}{3}\,dy =\frac{2}{3}.

Thus \(\bar{x}=M_y/m=(2/3)/4=1/6\), not zero. The mistaken claim “the region is symmetric, so its centroid lies on the \(y\)-axis” omits a necessary hypothesis. The correction is to check both the region and the density before applying the symmetry theorem.

When the Simplification Matters

Symmetry is useful both for reducing computation and for checking a result. If an integral over the full region has an integrand unchanged by reflection, a representative half can cut the work in half. If the integrand changes sign under reflection, its integral is zero. These conclusions apply to the mass and first moments defined here; before using them, identify how each integrand transforms under the relevant reflection.

A practical check is to calculate the centroid coordinate perpendicular to the claimed symmetry axis in two ways: infer that it is zero from the theorem, and confirm that the corresponding first moment cancels. If a calculation gives a nonzero value, check the reflection bounds, the sign of the moment integrand, and whether the density actually shares the symmetry.

Check Your Understanding

Use the definitions and symmetry conditions in this tutorial to answer the following.

For a lamina symmetric about the \(y\)-axis, which first moment must be zero, and which centroid coordinate does that determine?
What additional condition is needed before geometric symmetry of a region implies the symmetry theorem for a nonuniform lamina?
Why can the moment about the \(x\)-axis of a \(y\)-axis-symmetric lamina be nonzero?
For an integrand unchanged by reflection across an axis, how can a representative half-region be used to calculate its full-region integral?
In the rectangle with density \(2+x\), why does geometric symmetry fail to force \(\bar{x}=0\)?