Skip to content

Accumulation Point in a Metric Space

Definition

Let (X,d)(X,d) be a metric space and let E⊆XE\subseteq X. A point x∈Xx\in X is called an accumulation point (or cluster point, limit point) of EE if every metric ball around xx contains a point of EE different from xx, i.e.,

∀ε>0,(Bd(x,ε)∖{x})∩E≠∅.\forall \varepsilon>0,\quad \left(B_d(x,\varepsilon)\setminus\{x\}\right)\cap E\neq\varnothing.

Let (X,d)(X,d) be a metric space and let (xn)n∈N(x_n)_{n\in\mathbb{N}} be a sequence in XX. A point x∈Xx\in X is called an accumulation point (or cluster point) of (xn)(x_n) if, for every ε>0\varepsilon>0, there exist infinitely many n∈Nn\in\mathbb{N} such that xn∈Bd(x,ε)x_n\in B_d(x,\varepsilon), i.e., for every ε>0\varepsilon>0,

∣{n∈N ⁣:xn∈Bd(x,ε)}∣=∞.\left|\{n\in\mathbb{N}\colon x_n\in B_d(x,\varepsilon)\}\right|=\infty.

Interpretation

An accumulation point of a set is a location you can approach arbitrarily closely using other points from the set. For a sequence, it is a point to which the sequence returns again and again: however small the ball around it, infinitely many terms lie in that ball.

Both definitions ask something of every ε>0\varepsilon>0, so only small balls matter. For a set, the words "different from xx" rule out counting xx itself, and xx need not belong to EE: for E=(0,1)E=(0,1) in R\mathbb{R}, every point of [0,1][0,1] is an accumulation point, including 00 and 11. For a sequence, indices nn are counted, not distinct values. The constant sequence xn=cx_n=c has cc as an accumulation point, although the set {c}\{c\} of its terms has none.

The accumulation points of a sequence in a metric space are exactly the limits of its convergent subsequences, which is the form in which they appear in the Bolzano-Weierstrass theorem. For a bounded real sequence, the limit superior and limit inferior are its largest and smallest accumulation points. The accumulation points of a set are what a closed set must contain: a set is closed if and only if it contains all its accumulation points, and its closure is the set together with them.

A common slip is to treat a single accumulation point as a limit. The sequence 0,1,0,2,0,3,…0,1,0,2,0,3,\dots has 00 as its only accumulation point, yet it does not converge, because it is unbounded.


Key Results

  • A limit point can be characterized as an adherent point that is not an isolated point (a point x∈Ex\in E such that some ball Bd(x,ε)B_d(x,\varepsilon) meets EE only in xx).

Examples

  • For E={1, 1/2, 1/3, … }E=\{1,\,1/2,\,1/3,\,\dots\}, the only accumulation point is 00. Each point 1/n1/n is isolated.
  • The sequence an=(−1)na_n=(-1)^n has accumulation points 1 and -1. It does not converge.
  • Consider the sequence (an)n=1∞(a_n)_{n=1}^\infty that is defined by an=sin⁡na_n=\sin n. We can show that any number A∈[−1,1]A\in[-1,1] is an accumulation point of this sequence.

    The terms sin n plotted for n up to 50, with a thin horizontal band around 0.5; a few terms fall inside the band, and infinitely many fall in every such band, so 0.5 is an accumulation point.

    Among the first 5050 terms, four lie in the band [0.4,0.6][0.4,0.6] (red); infinitely many terms fall in every band around 0.50.5, so 0.50.5 is an accumulation point.


Prerequisites

Sources

Aliases: cluster point, limit point, subsequential limit Symbols: element of, epsilon, for all, infinity, intersection, natural numbers, real numbers, subset

Last updated: 20 October 2026 03:02