3.7. CAUCHY SEQUENCES 93

Proof: From the above corollary, if the set is closed and bounded, then it is sequentiallycompact. Suppose now that K is sequentially compact. Why is it closed and bounded? Ifit is not bounded, then you could pick {kn}∞

n=1 such that |kn| ≥ n. Since K is sequentiallycompact, it follows that there is a subsequence,

{kn j

}which satisfies lim j→∞ kn j = k ∈ K.

But then this sub sequence would be contained in some interval which is impossible fromthe construction. Thus K is bounded. Why must K be closed? Suppose K fails to containp where p = limn→∞ pn, pn ∈ K. A subsequence

{pnk

}must converge to a point of K.

But this subsequence must converge to p by Theorem 3.3.14 which is a contradiction. ByTheorem 3.6.11 K is closed.

3.7 Cauchy SequencesDefinition 3.7.1 A sequence {xk}∞

k=1 is called a Cauchy sequence if for any ε > 0there exists nε such that whenever m,n ≥ nε , it follows that |xn − xm|< ε . In other words,the terms of the sequence “bunch up”.

I will be vague about the context of the following fundamental proposition because itapplies in far greater generality than R. You can think of the sequence being in R becausethis is the main example of interest here. Part 1. is especially useful in more generalcontexts.

Proposition 3.7.2 If {xn} is a Cauchy sequence, then

1. If a subsequence{

xnk

}∞

k=1 converges to x, it follows that limn→∞ xn = x.

2. If limn→∞ xn = x, then {xn} must be a Cauchy sequence.

3. Every Cauchy sequence is bounded.

Proof: Consider 1.There exists nε such that if n,m > nε , then |xn − xm| < ε/3. Therealso exists kε such that if k > kε , then

∣∣x− xnk

∣∣ < ε/3. Now let k > max(kε ,nε) . Thennk ≥ k > max(kε ,nε) and so |x− xk| ≤

∣∣x− xnk

∣∣+ ∣∣xnk − xk∣∣< ε

3 +ε

3 < ε so limn→∞ xn = x.2. As to the next claim, there is N such that if m ≥ N, then |x− xm|< ε/2. If m,n > N,

then |xm − xn| ≤ |xm − x|+ |x− xn|< ε

2 +ε

2 = ε and so any convergent sequence is a Cauchysequence.

3. Finally, if {xn} is a Cauchy sequence, then there exists N such that if m,n ≥ N,then |xm − xn| ≤ 1. In particular, |xn − xN | ≤ 1 and so |xn| ≤ |xN |+ 1. Now for any k,|xk| ≤ max(|xN |+1, |xi| , i = 1,2, · · · ,N) .

Theorem 3.7.3 Let {xn} be a Cauchy sequence in R. Then it converges. Conversely,if a sequence {xn} converges, then the sequence is a Cauchy sequence.

Proof: Since {xn} is a Cauchy sequence, it is bounded by Proposition 3.7.2 so is con-tained in some closed interval [−a,a] , a sequentially compact set from the above Theorem3.6.2. Therefore, there is a subsequence

{xnk

}such that limk→∞ xnk = x ∈ [−a,a]. By

Proposition 3.7.2, the original Cauchy sequence converges to x. The second claim is fromProposition 3.7.2.

3.7. CAUCHY SEQUENCES 93Proof: From the above corollary, if the set is closed and bounded, then it is sequentiallycompact. Suppose now that K is sequentially compact. Why is it closed and bounded? Ifit is not bounded, then you could pick {k,}*_, such that |k,| >. Since K is sequentiallycompact, it follows that there is a subsequence, {kn; } which satisfies lim j,.. kn =keK.But then this sub sequence would be contained in some interval which is impossible fromthe construction. Thus K is bounded. Why must K be closed? Suppose K fails to containPp where p = limp Pn; Pn € K. A subsequence { Pn, must converge to a point of K.But this subsequence must converge to p by Theorem 3.3.14 which is a contradiction. ByTheorem 3.6.11 K is closed. J3.7. Cauchy SequencesDefinition 3.7.1 4 sequence {x,};,_, is called a Cauchy sequence if for any € > 0there exists ng such that whenever m,n > ne, it follows that |x, —Xm| < €. In other words,the terms of the sequence “bunch up”.I will be vague about the context of the following fundamental proposition because itapplies in far greater generality than IR. You can think of the sequence being in R becausethis is the main example of interest here. Part 1. is especially useful in more generalcontexts.Proposition 3.7.2 /f {x,} is a Cauchy sequence, then1. Ifa subsequence {Xn yet converges to x, it follows that limy.0Xy) = Xx.2. If impy—yooXn = x, then {x,} must be a Cauchy sequence.3. Every Cauchy sequence is bounded.Proof: Consider 1.There exists ng such that if n,m > ne, then |x, —x»| < €/3. Therealso exists kg such that if k > ke, then |x — Xn, | < €/3. Now let k > max (ke,ne). Thenny > k > max (ke, ne) and so |x — x;| < |x — Xn, | + [ng — Xz <§4+§ <€ so lim, 40% =X.2. As to the next claim, there is N such that if m > N, then |x —x,,| < €/2. If m,n >N,then |Xm —Xn| < |Xm —x|+|x—xn| < 5 +5 = € and so any convergent sequence is a Cauchysequence.3. Finally, if {x,} is a Cauchy sequence, then there exists N such that if m,n > N,then |x —2x,| < 1. In particular, |x, —xy| <1 and so |x,| < |xy| +1. Now for any k,xg] < max (lxy|+1, |x], =1,2,---,N).Theorem 3.7.3 Let {x,} be a Cauchy sequence in R. Then it converges. Conversely,if a sequence {x,} converges, then the sequence is a Cauchy sequence.Proof: Since {x,} is a Cauchy sequence, it is bounded by Proposition 3.7.2 so is con-tained in some closed interval [—a, a] , a sequentially compact set from the above Theorem3.6.2. Therefore, there is a subsequence {x,, } such that limy_,..%,, =x € [—a,a]. ByProposition 3.7.2, the original Cauchy sequence converges to x. The second claim is fromProposition 3.7.2. §j