Skip to content

Cauchy Sequence in a Metric Space

Definition

A sequence of points (xi)i=1∞(x_i)_{i=1}^\infty in a metric space (X,d)(X,d) is called a Cauchy sequence if, for every ε>0\varepsilon > 0, there exists an integer N≥1N\geq 1 such that the distance d(xj,xk)<εd(x_j,x_k)<\varepsilon for all j,k≥Nj, k \geq N.


Interpretation

A Cauchy sequence is one whose terms eventually all lie close to each other. It is the same idea as a Cauchy sequence of real numbers, but with the absolute value replaced by a general metric. Unlike the definition of convergence, it never mentions a limit, so it can be checked using the terms alone.

Points of a sequence spiralling inwards inside a rectangle representing the space; from some index on all terms lie within a small circle, close to each other.

From some index NN on, all terms in the figure lie inside the small dashed circle, close enough that d(xj,xk)<εd(x_j,x_k)<\varepsilon for all j,k≥Nj,k\ge N. The terms get closer to each other, but a Cauchy sequence need not converge to a point of XX unless the space is complete.

Raise NN until all tail points lie within ε\varepsilon of each other, without using the limit at all.

The condition concerns all pairs j,k≥Nj,k\ge N, not only consecutive terms. The partial sums xk=1+12+⋯+1kx_k=1+\tfrac{1}{2}+\cdots+\tfrac{1}{k} of the harmonic series have ∣xk+1−xk∣=1k+1→0|x_{k+1}-x_k|=\tfrac{1}{k+1}\to0, yet x2k−xk≥k⋅12k=12x_{2k}-x_k\ge k\cdot\tfrac{1}{2k}=\tfrac{1}{2} for every kk, so they are not Cauchy.

Every convergent sequence is Cauchy, by the triangle inequality d(xj,xk)≤d(xj,x)+d(x,xk)d(x_j,x_k)\le d(x_j,x)+d(x,x_k). The converse holds exactly in complete metric spaces, such as Rn\mathbb{R}^n with any pp-metric, and fails in Q\mathbb{Q}, as the Examples show. Being Cauchy depends on the metric, but strongly equivalent metrics have the same Cauchy sequences, because dq(xj,xk)≤b dp(xj,xk)d_q(x_j,x_k)\le b\,d_p(x_j,x_k) and dp(xj,xk)≤1a dq(xj,xk)d_p(x_j,x_k)\le\tfrac{1}{a}\,d_q(x_j,x_k).


Key Results

  • While all convergent sequences are Cauchy, not every Cauchy sequence converges because a sequence can cluster together (Cauchy property) around a limit that does not exist within the current space.

Examples

  • Take the rational numbers Q\mathbb{Q} with the usual metric d(x,y)=∣x−y∣d(x,y)=|x-y|. The sequence an=(1+1n)na_n=\left(1+\dfrac{1}{n}\right)^n consists of rational numbers and converges in R\mathbb{R} to e\mathrm{e}, so it is Cauchy. Since e\mathrm{e} is irrational, it does not converge to any point of Q\mathbb{Q}: the space (Q,d)(\mathbb{Q},d) is not complete.

Prerequisites

Sources

Symbols: epsilon, infinity, rational numbers, real numbers

Last updated: 20 October 2026 03:02