HomeArticlesMathematics

Wallpaper Symmetry Groups: The 17 Ways to Repeat a Pattern Forever

Translations, rotations, reflections and glide reflections combine into exactly 17 possible symmetry classes for any repeating 2D pattern, proven by Fedorov in 1891.

mysimulator teamUpdated June 2026≈ 8 min read▶ Open the simulation

17 ways to tile a plane, no more, no fewer

A wallpaper group classifies a repeating 2D pattern by which symmetry operations map it onto itself: translations (sliding), rotations, reflections (mirrors) and glide reflections (a reflection combined with a slide along the mirror line). In 1891 the Russian crystallographer Evgraf Fedorov proved that, despite the apparently infinite variety of possible repeating patterns, every one of them belongs to exactly one of 17 distinct symmetry groups — not 16, not 18. This simulation renders nine of them by tiling one deliberately asymmetric motif with the specific symmetry operations each group allows, so the pattern's regularity comes entirely from the symmetry, not from any hidden symmetry in the motif itself.

live demo · tiling an asymmetric motif under a chosen wallpaper group● LIVE

Why using an asymmetric motif matters

If the repeating unit itself already has mirror symmetry, a pattern built from it can look like it has more symmetry operations than the underlying group actually guarantees, hiding the classification. Using a motif with no symmetry of its own — an irregular comma-like or flag-like shape is traditional in crystallography textbooks — makes every symmetry visible in the final tiling attributable only to the wallpaper group's operations. That's why each of the nine groups here looks visually distinct even though they all repeat the same single motif.

The four operations and how they combine

translation        slide by a fixed vector — every wallpaper group has 2 independent ones
rotation           turn by 360°/n about a fixed point (n = 2, 3, 4 or 6 only — see below)
reflection         mirror across a fixed line
glide reflection   mirror across a line, THEN slide along that same line

Every wallpaper group needs two independent translations to fill the plane (that's what makes it a wallpaper group rather than a border/frieze pattern, which only repeats in one direction). Group p1 has translations only — no rotation, no mirrors, the plainest possible repeat. Groups like pm and pg add a single mirror or a single glide reflection on top of the translations. Denser groups like p4m or p6m stack rotation centres, mirror lines and glide axes all at once, with strict geometric constraints on where each type of symmetry element can sit relative to the others.

The crystallographic restriction: why n = 5 is forbidden

A striking constraint limits rotational symmetry in any pattern with two independent translations to orders n = 2, 3, 4 or 6 only — 5-fold and any n ≥ 7 rotational symmetry is mathematically impossible in a wallpaper pattern (as opposed to a single decorative rosette, which can have any rotational symmetry it likes). The proof is a short argument by contradiction: assume an n-fold centre exists alongside a shortest translation vector, apply the rotation to that vector from both ends, and for n = 5 or n ≥ 7 the resulting new translation vector turns out to be shorter than the one you started with — contradicting the assumption that you picked the shortest one. This is the crystallographic restriction theorem, and it's the reason floor tiles, wallpaper, and crystal lattices never show true 5-fold or 7-fold repeating symmetry, even though 5-fold symmetry is perfectly common in single motifs like starfish or flowers that don't need to tile.

Penrose tilings are the famous escape hatch: they display 5-fold symmetry precisely because they are aperiodic — they never repeat by translation at all, so the crystallographic restriction (which assumes two independent translations) simply doesn't apply to them.

Naming: what p4m and p6m are telling you

The standard crystallographic short names encode the group's structure. The leading p (or c) marks a primitive (or centred) lattice cell. The number is the highest-order rotation present (1, 2, 3, 4 or 6). Trailing letters mark reflection axes: m for a mirror, g for a glide reflection only, and a second letter when mirrors run in two distinct directions. So p4m means a primitive cell with 4-fold rotation centres and mirror lines — the symmetry of a simple checkerboard or square grid of squares — while p6m adds 6-fold rotation and mirrors, the symmetry of a hexagonal honeycomb, the densest and most constrained of all 17 groups.

Where the 17 groups actually get used

This isn't just decorative classification. Crystallography uses the 3D analogue (230 space groups) to catalogue how atoms repeat in real crystal lattices, which is foundational to X-ray diffraction analysis and materials science. Islamic geometric art, going back centuries before Fedorov's proof, empirically explored patterns realising most or all 17 groups long before anyone had a formal classification. And the same group theory underlies modern texture synthesis and procedural pattern generation in computer graphics, where picking a wallpaper group up front guarantees a seamlessly tileable result.

Frequently asked questions

Why can't a wallpaper pattern have 5-fold rotational symmetry?

This is the crystallographic restriction theorem: any pattern that repeats via two independent translation directions can only support rotational symmetry of order 2, 3, 4 or 6. A proof by contradiction shows that assuming a 5-fold (or 7-fold-or-higher) rotation center alongside the shortest translation vector always produces an even shorter translation vector, which is impossible if you started with the shortest one.

How is a wallpaper group different from just picking a repeating tile?

Any repeating tile has at least translational symmetry, but a wallpaper group also classifies which additional operations - rotations, mirror reflections, glide reflections - map the whole infinite pattern onto itself. Two patterns built from completely different motifs can belong to the same wallpaper group if they share the same underlying symmetry operations.

Are there really only 17 wallpaper groups?

Yes, proven rigorously by Evgraf Fedorov in 1891 (with related work by Fedorov, Schoenflies and Polya around the same period). No repeating 2D pattern with two independent translation directions can fall outside these 17 classes, regardless of how intricate or unusual its motif looks.

Try it live

Everything above runs in your browser — open Wallpaper Symmetry Groups and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.

▶ Open Wallpaper Symmetry Groups simulation

What did you find?

Add reproduction steps (optional)