Reflection symmetry

Fold a square along one of its diagonals and the two triangular halves land on each other exactly — not roughly, exactly, corner for corner. That is reflection symmetry: a shape is cut by a mirror line into two halves that are perfect mirror images of one another. The idea sits inside the Class 8 Constructions chapter alongside ruler-and-compass work on quadrilaterals, because symmetry is the fastest check that a construction came out right. This article gives the precise definition of a line of symmetry, counts the lines of symmetry in every shape ICSE Class 8 expects — triangles, rectangles, rhombi, squares and regular polygons — works through a rhombus and an alphabet-letter example in full, and separates reflection symmetry cleanly from its cousin, rotational symmetry, so the two are never mixed up on paper again.

The core insight

A figure has reflection symmetry about a line $l$ when every point $P$ on it has a mirror image $P'$ (with $l$ as the perpendicular bisector of $PP'$ ) that also lies on the figure, and the count of such lines — $1$ for an isosceles triangle, $3$ for equilateral, $2$ for a rectangle or rhombus, $4$ for a square, and $n$ for a regular $n$ -gon — depends entirely on how many ways the shape's equal sides and angles line up.

The core idea

Defining the mirror line. Take a line $l$ and a point $P$ not on it. The reflection (mirror image) of $P$ in $l$ is the point $P'$ such that $l$ passes exactly halfway between them, at right angles to the segment joining them. In symbols, if $M$ is the point where $l$ meets segment $PP'$ :

PM = P'M, \qquad PP' \perp l.

where: $M$ is the midpoint of $P$ and its image $P'$ , and the two conditions together say $l$ is the perpendicular bisector of $PP'$ . A whole figure has reflection symmetry about $l$ — and $l$ is called its axis of symmetry or line of symmetry — if reflecting every point of the figure in $l$ produces the same figure back again: nothing added, nothing missing.

Counting the lines depends on the shape's own symmetry of sides and angles. A shape with more equal sides and equal angles generally offers more ways to fold it onto itself:

Isosceles triangle (two equal sides, not equilateral): exactly $1$ line — the perpendicular from the apex (the vertex between the two equal sides) to the base, which is also the median and the angle bisector from that vertex.
Equilateral triangle: $3$ lines, one through each vertex and the midpoint of the opposite side.
Rectangle (not a square): $2$ lines, through the midpoints of each pair of opposite sides. The diagonals do not work, since a rectangle's diagonals are not perpendicular to its sides in the way a square's are.
Rhombus (not a square): $2$ lines — its two diagonals, because a rhombus's diagonals bisect each other at right angles and swap the two vertices on either side.
Square: $4$ lines — $2$ through the midpoints of opposite sides plus $2$ along the diagonals, since a square is simultaneously a special rectangle and a special rhombus.
Regular polygon with $n$ sides: exactly $n$ lines of symmetry (proved by the counting argument below).
General (scalene, irregular) triangles and quadrilaterals: usually $0$ lines — no side or angle equality to fold along.

Reflection symmetry is a static property (fold and match); it is entirely different from rotational symmetry (spin about a centre and the figure looks unchanged partway round) — the two must never be assumed to go together.

Figure

A square with two diagonal mirror lines and two midpoint mirror lines all drawn through its centre, four lines of symmetry in total

Derivation AWhy a regular polygon with n sides always has exactly n lines of symmetry, no more and no fewer

The goal is to show that a regular $n$ -gon (all $n$ sides equal, all $n$ interior angles equal) has exactly $n$ distinct lines of symmetry by classifying every mirror line into one of two kinds and counting each kind, using only the fact that the shape has $n$ -fold equal sides and equal angles around a common centre.

Step 1. Set up the centre. A regular polygon has a centre point $O$ equidistant from all $n$ vertices (this is guaranteed by every side and angle being equal). Any line of symmetry of the polygon must pass through $O$ , because reflecting the whole figure has to send the vertex set back onto itself, and the only point every vertex is equidistant from is $O$ — so $O$ cannot move under the reflection, meaning the mirror line passes through it.

Step 2. Classify the case of $n$ odd. Take $n = 5$ (a regular pentagon) as the pattern. Each line of symmetry must pass through one vertex and the midpoint of the opposite side, because reflecting swaps the two neighbouring vertices of that one vertex while fixing it. There are $5$ vertices, so there are exactly $5$ such vertex-to-opposite-midpoint lines — one per vertex — and no other line through $O$ can work, since any other line would fail to map the vertex set exactly onto itself.

Step 3. Classify the case of $n$ even. Take $n = 6$ (a regular hexagon) as the pattern. Now there are two families: $3$ lines joining a pair of opposite vertices (since $6$ vertices pair up into $3$ opposite pairs), and $3$ lines joining the midpoints of a pair of opposite sides (since $6$ sides pair up into $3$ opposite pairs). That gives $3 + 3 = 6$ lines total — again matching $n$ .

Step 4. Generalise and conclude. In both cases the count of distinct valid mirror lines through $O$ equals the number of vertices, $n$ , because each vertex (odd case) or each vertex-pair/side-pair (even case) contributes exactly one line, and equal sides plus equal angles guarantee no additional stray line can also work. Hence a regular $n$ -gon has exactly $n$ lines of symmetry — matching the square ( $n=4$ : $2$ diagonal $+$ $2$ midpoint $= 4$ ) and equilateral triangle ( $n=3$ , the odd case, $3$ vertex-to-midpoint lines) already listed above.

Worked example

Problem: A rhombus $ABCD$ is constructed with side $5\text{ cm}$ and diagonal $AC = 8\text{ cm}$ . State its lines of symmetry, explain why each one works, and then say how many lines of symmetry a regular hexagon tile of the same side length would have.

Step 1. Recall the rhombus property. A rhombus has all four sides equal ( $AB = BC = CD = DA = 5\text{ cm}$ ), and its diagonals bisect each other at right angles. That is exactly the setup needed for a diagonal to be a mirror line.

Step 2. Check diagonal $AC$ . Reflecting the rhombus across $AC$ must send $B$ to $D$ . Since $AB = AD$ and $CB = CD$ (all sides equal), and $BD$ is perpendicular to $AC$ and bisected by it, $B$ and $D$ are indeed mirror images of each other across $AC$ — so $AC$ is a line of symmetry.

Step 3. Check diagonal $BD$ . By the identical reasoning with the roles of the vertices swapped, reflecting across $BD$ sends $A$ onto $C$ . So $BD$ is also a line of symmetry.

Step 4. Count and confirm. No other line through the centre works here, because a plain rhombus (unlike a square) has unequal angles, so the midpoint-to-midpoint lines that work for a square do not work for it. The rhombus has exactly 2 lines of symmetry: its diagonals $AC$ and $BD$ .

Competitive stretch

By the $n$ -lines rule proved above, a regular hexagon tile has $n = 6$ lines of symmetry: $3$ through opposite vertex pairs and $3$ through opposite side-midpoint pairs — three times as many as the rhombus, even though both shapes have four or six equal sides respectively, because the hexagon's equal angles add the extra midpoint family that the rhombus's unequal angles rule out.

Common traps

Confusing reflection symmetry with rotational symmetry. A parallelogram that is not a rhombus or rectangle looks unchanged after a $180^\circ$ rotation, but it has zero lines of symmetry — the two properties are independent and must be checked separately.
Assuming every quadrilateral has a line of symmetry. A general irregular quadrilateral, or a scalene triangle, typically has 0 lines — the mirror-image condition ( $PM = P'M$ with $PP' \perp l$ for every point) must actually be verified, never assumed from "it looks balanced."
Miscounting a rhombus's lines as 4. Students borrow the square's answer by mistake. A plain rhombus has only its $2$ diagonals as mirror lines; the midpoint-to-midpoint lines only work when the angles are also all equal, i.e. when the rhombus is a square.
Drawing the fold line by eye. A true line of symmetry is located precisely — as a perpendicular bisector or angle bisector via compass construction — not sketched freehand, which can hide a small asymmetry that fails the $PM = P'M$ test.

Ask the tutor: Want to work out the lines of symmetry for a regular octagon or check whether a shape you've just constructed is truly symmetric — ask your AI tutor to walk through it with you.

Does a rhombus ever have 4 lines of symmetry like a square

Yes, but only when it is a square. A general rhombus has unequal angles (two acute, two obtuse), so only its two diagonals work as mirror lines — the two midpoint-to-midpoint lines that work for a square fail because reflecting across them would need the angles at each end to match, and in a non-square rhombus they don't. Once all four angles become $90^\circ$ , the rhombus turns into a square and gains those two extra lines, for $4$ total.

Why does N have no line of symmetry even though it looks balanced

"Balanced" is exactly the trap. N does look unchanged if you rotate it $180^\circ$ about its centre — that is rotational symmetry, a different property. But no straight fold line maps N onto itself: fold it vertically and the top-left and bottom-right strokes land in empty space, not on each other; fold it horizontally, same problem. Always test with the actual mirror-image rule ( $PM = P'M$ , $PP' \perp l$ ) rather than a general sense of "looks even."

How do I quickly find how many lines of symmetry a regular polygon has without drawing it

Use the rule proved above: a regular polygon with $n$ sides has exactly $n$ lines of symmetry, split as $n$ vertex-to-opposite-midpoint lines when $n$ is odd, or $n/2$ vertex-pair lines plus $n/2$ side-midpoint-pair lines when $n$ is even. So a regular pentagon has $5$ , a regular octagon has $8$ ( $4$ vertex-pair $+$ $4$ midpoint-pair), and so on — you never need to sketch every fold to get the count right.

Reflection symmetry

The core idea

Worked example

Competitive stretch

Common traps

Before this

Where this leads