How this instrument works
An ellipse has exactly one center point, and both its long and short axes pass through it — so locating one axis's two opposite vertices and averaging their coordinates pins the point down exactly, cx = (x1+x2)/2 and cy = (y1+y2)/2. This is the same idea as finding a circle's middle from the two ends of a diameter: an ellipse is a circle stretched unevenly in one direction, and stretching a segment does not move its midpoint.
The reason the averaging works is that x and y are independent coordinates, so an ellipse's reflective symmetry through that point applies to each one separately. Reflecting the vertex (x1, y1) through it lands exactly on (x2, y2); averaging is the arithmetic way of finding what sits equidistant from both, without needing the semi-axis lengths, the rotation angle, or where the two foci lie inside the curve.
The method only holds if the two entered points are genuinely opposite vertices of the same line — swap one for an arbitrary point on the boundary, or pair a major-axis vertex with a minor-axis one, and the average lands somewhere with no particular geometric meaning. A circle is the tidy limiting case, where every diameter plays that role, which is why this trick still holds when an ellipse's two semi-axis lengths happen to match.
- Enter one vertex into Major-axis endpoint 1: x and Major-axis endpoint 1: y.
- Enter the opposite vertex — the far end of that same segment — into Major-axis endpoint 2: x and Major-axis endpoint 2: y.
- Read Center: x and Center: y for the coordinates of the ellipse's single center point.
- Confirm both points sit on the major axis rather than an arbitrary point elsewhere on the curve, since the averaging only holds for true opposite vertices.
Worked example — a major axis from (−3, 2) to (7, 2)
Suppose an ellipse's major axis runs from the vertex (−3, 2) to the vertex (7, 2); the matching second coordinate on both ends already shows the axis is horizontal. Averaging the first coordinates gives cx = (−3 + 7) ⁄ 2 = 2. Because the two known y-values are already equal, their average cy is trivially that same value, 2 — so the center sits at (2, 2), precisely midway between the two vertices.
A quick cross-check confirms it: the distance from (2, 2) to each vertex should match, since both sit the same semi-major length from the middle. To (−3, 2) that distance is 5 units, and to (7, 2) it is also 5 units — confirming the center really is the midpoint, sitting exactly halfway along the 10-unit major axis.
Questions
What's the difference between an ellipse's center and its foci?
The center is the midpoint of both axes, equidistant from the two vertices you enter here. The foci are two different points inside the ellipse, offset from the center along the major axis by c = √(a² − b²); they coincide with the center only in the limiting case of a circle, where a = b and c = 0.
Can I use the minor axis's endpoints instead of the major axis's?
Yes — an ellipse has exactly one center, and it's the midpoint of both axes through it. This calculator labels its fields for the major axis because that pair of vertices is usually the one measured or known first, but averaging the minor-axis endpoints returns the identical point.
What mistake does this formula catch people out on?
Entering a vertex and an unrelated boundary point instead of two true opposite vertices. The midpoint formula only returns the center when both inputs sit at opposite ends of the same straight axis; average a vertex with some other point on the curve and the result has no particular geometric meaning.
Does the order I enter the two endpoints matter?
No. Addition is commutative, so (x1 + x2) equals (x2 + x1), and the same goes for the y-coordinates. Label either vertex 'endpoint 1' and the calculation returns exactly the same center either way.
How is this related to the midpoint formula for a line segment?
It is that formula, applied to the segment formed by the two axis vertices. Finding a line segment's midpoint, a circle's center from a diameter, and an ellipse's center from major-axis vertices are all the same computation: average the x-coordinates, then average the y-coordinates.
Does this work for a tilted or rotated ellipse?
Yes. Because the averaging happens independently on the x-coordinates and the y-coordinates supplied, the method needs no information about rotation angle, semi-axis lengths, or orientation — only the coordinates of the two major-axis vertices, however the ellipse sits on the plane.