Conventions. Natural numbers are read as real numbers through the canonical map ι of The Canonical Map from the Natural Numbers to a Field, as in claim 3 of The Real Numbers and Standard Notation. We use the following transfer facts without further comment. By claims 6 and 7 of Properties of the Canonical Map from the Natural Numbers to an Ordered Field together with the trichotomy of the order on N (claim 3 of Properties of the Order on the Natural Numbers), for k,l∈N one has k<l in N if and only if ι(k)<ι(l) in R, and likewise for ≤. If k<l in N then k+1≤l: otherwise trichotomy gives l<k+1=S(k), so claim 5 of Properties of the Order on the Natural Numbers gives l≤k, which with k≤l forces k=l by claim 2 there, contradicting k<l. Finally ι(k+1)=ι(k)+1 by claim 1 of Properties of the Canonical Map from the Natural Numbers to an Ordered Field, and 1≤ι(k) by claim 2 there. Finally, the order ≤ on R is a total order, because by Ordered Field the order of an ordered field is one.
Fix m∈N and put h=s/m; since 0<s and 0<m (claim 3 of Properties of the Canonical Map from the Natural Numbers to an Ordered Field), h is a positive real number, and mh=s. For j∈[m]n and i∈[n] write
ai(j)=ci+(ji−1)h,bi(j)=ci+jih,bi(j)−ai(j)=h,
so that Qm,j={x∈Rn:ai(j)≤xi<bi(j) for every i∈[n]}.
Proof of claim 1.
(a) [m]n is nonempty and finite. The set [m] has m elements by claim 1 of Basic Properties of Finite Sets, hence is finite, and it is nonempty because 1≤m by claim 4 of Properties of the Order on the Natural Numbers. Claim 3 of Finiteness of Cartesian Products, Tuple Sets, and Permutation Sets now gives that [m]n is nonempty and finite.
(b) Each cell is Borel of measure hn. Fix j∈[m]n and i∈[n], and put Ii={t∈R:ai(j)≤t<bi(j)}. If u≤v≤w with u,w∈Ii, then ai(j)≤u≤v and v≤w<bi(j), so v∈Ii; thus Ii is an interval with endpoints ai(j)≤bi(j), and claim 4 of Existence of Lebesgue Measure on the Real Line shows that Ii is a Borel subset of R with λ(Ii)=bi(j)−ai(j)=h. Since Qm,j=I1×⋯×In, Lebesgue Measure on Rn gives Qm,j∈B(Rn) and
λn(Qm,j)=λ(I1)⋯λ(In)=hn.
(c) Diameter bound. Let x,y∈Qm,j and i∈[n]. From ai(j)≤xi<bi(j) and ai(j)≤yi<bi(j) we get −h<xi−yi<h, hence ∣xi−yi∣≤h and therefore (xi−yi)2=∣xi−yi∣2≤h2 by claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field. Summing over i and using claim 1 of Elementary Properties of the Euclidean Norm on Rn,
∥x−y∥2=i=1∑n(xi−yi)2≤nh2=σn2h2=(σnh)2.
As ∥x−y∥ and σnh are nonnegative, claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field gives ∥x−y∥≤σnh.
(d) Pairwise disjointness. Let j,j′∈[m]n with j=j′, and suppose x∈Qm,j∩Qm,j′. Choose i∈[n] with ji=ji′; by trichotomy we may assume ji<ji′, and then ji+1≤ji′. Since 0<h,
bi(j)=ci+((ji+1)−1)h≤ci+(ji′−1)h=ai(j′),
so xi<bi(j)≤ai(j′)≤xi, which is impossible.
(e) The union of the cells is B. If x∈Qm,j with j∈[m]n, then for each i we have 1≤ji, so 0≤ji−1 and ci≤ai(j)≤xi; and ji≤m, so xi<bi(j)≤ci+mh=ci+s. Hence x∈B.
Conversely let x∈B and fix i∈[n]. Put Ai={k∈[m]:ci+(k−1)h≤xi}. As ci≤xi we have 1∈Ai, so Ai is nonempty, and Ai⊆[m] is finite by claim 3 of Basic Properties of Finite Sets. Let q be its number of elements and f:[q]→Ai a bijection; regarding ι∘f as a q-tuple in R with its total order, Greatest Element of a Finite Family in a Totally Ordered Set provides t∈[q] with f(u)≤f(t) for every u∈[q]. Put ji=f(t), an element of Ai. Since f maps [q] onto Ai, every k∈Ai equals f(u) for some u∈[q], so k≤ji for every k∈Ai; and ji is the unique element of Ai with that property, because two elements of Ai each at least the other are equal by antisymmetry.
By definition of Ai we have ai(j)=ci+(ji−1)h≤xi. Suppose xi<ci+jih failed, that is, ci+jih≤xi. If ji=m this reads ci+s≤xi, contradicting x∈B. Otherwise ji=m, and since ji≤m trichotomy gives ji<m and hence ji+1≤m, so ji+1∈[m]; moreover ci+((ji+1)−1)h=ci+jih≤xi, so ji+1∈Ai and therefore ji+1≤ji, which is false. Hence xi<ci+jih.
Carrying this out for every i∈[n] determines j=(j1,…,jn)∈[m]n with x∈Qm,j. This completes the proof of claim 1.
Proof of claim 2. Let m∈N and j′∈[2m]n, and write h=s/m and h′=s/(2m), so that h=2h′ and 0<h′. Fix i∈[n] and put Ai′={k∈[m]:2k−1≤ji′}. Since 1≤ji′ and 2⋅1−1=1, we have 1∈Ai′, so Ai′ is a nonempty subset of the finite set [m]; exactly as in part (e) above, let ji be its greatest element. Then
2ji−1≤ji′.
We check that also ji′≤2ji. If ji=m this is ji′≤2m, which holds since ji′∈[2m]. Otherwise ji<m, so ji+1∈[m], and maximality gives ji+1∈/Ai′, that is, ji′<2(ji+1)−1=2ji+1. Both ji′ and 2ji are integers by claims 1 and 2 of Arithmetic, Order and Discreteness of the Integers, so 2ji<ji′ would give 2ji+1≤ji′ by claim 3 there, contradicting the previous inequality; hence ji′≤2ji.
Now let y∈Q2m,j′ and i∈[n]. From 2ji−2≤ji′−1 and 0<h′,
ci+(ji−1)h=ci+(2ji−2)h′≤ci+(ji′−1)h′≤yi,
and from ji′≤2ji,
yi<ci+ji′h′≤ci+2jih′=ci+jih.
Hence y∈Qm,j, so Q2m,j′⊆Qm,j with j=(j1,…,jn)∈[m]n.
For uniqueness, note that Q2m,j′ is nonempty: the point whose ith coordinate is ci+(ji′−1)h′ lies in it, because 0<h′. Since by claim 1 the cells Qm,j, j∈[m]n, are pairwise disjoint, at most one of them can contain the nonempty set Q2m,j′.
Proof of claim 3. The set Jm is a subset of the finite set [m]n, hence finite by claim 3 of Basic Properties of Finite Sets. It is nonempty: choosing x∈K⊆B, claim 1 gives j∈[m]n with x∈Qm,j, and then x∈Qm,j∩K, so j∈Jm. The same argument applied to an arbitrary x∈K shows K⊆Em, and Em⊆B because every cell is contained in B by claim 1.
Let p=∣Jm∣ and let g:[p]→Jm be a bijection. The sets Qm,g(1),…,Qm,g(p) are pairwise disjoint members of B(Rn) with union Em, so Em∈B(Rn) because a σ-algebra is closed under finite unions, and claim 1 of Basic Properties of a Measure together with part (b) above gives
λn(Em)=k=1∑pλn(Qm,g(k))=phn=∣Jm∣(ms)n,
a real number and in particular finite.
Proof of claim 4. Write Ek∗=Emk and Jk∗=Jmk.
(a) Monotonicity. Fix k∈N; then mk+1=2mk. Let x∈Ek+1∗, say x∈Qmk+1,j′ with j′∈Jk+1∗. By claim 2 there is j∈[mk]n with Qmk+1,j′⊆Qmk,j, and then ∅=Qmk+1,j′∩K⊆Qmk,j∩K, so j∈Jk∗ and x∈Ek∗. With K⊆Ek+1∗ from claim 3 this gives K⊆Ek+1∗⊆Ek∗.
(b) The mesh indices grow. We have k≤mk for every k∈N, by induction: m1=1; and if k≤mk then, since 1≤mk, mk+1=2mk=mk+mk≥mk+1≥k+1.
(c) The intersection is K. The inclusion K⊆⋂kEk∗ is part (a). Conversely let x∈⋂kEk∗ and suppose x∈/K. As K is compact it is closed in Rn by Compact Subset of Rn is Closed, so its complement is open and there is a real η>0 with BdE(x,η)⊆Rn∖K, where BdE is the open ball. By The Archimedean Property of the Real Numbers there is k∈N with σns/η<k, and then σns/η<k≤mk by (b), so σns/mk<η. Since x∈Ek∗ there is j∈Jk∗ with x∈Qmk,j, and there is y∈Qmk,j∩K. By claim 1, ∥x−y∥≤σns/mk<η, so dE(x,y)<η and y∈BdE(x,η)∩K, a contradiction. Hence x∈K.
(d) Convergence of the measures. Being compact, K belongs to B(Rn) and λn(K)<∞ by claim 3 of Balls Have Positive Lebesgue Measure and Bounded Sets Have Finite Lebesgue Measure. Put Dk=E1∗∖Ek∗. By (a) the sequence (Dk)k∈N is increasing, its members lie in B(Rn), and by (c)
k∈N⋃Dk=E1∗∖k∈N⋂Ek∗=E1∗∖K.
By claim 3, λn(E1∗)<∞, and Ek∗⊆E1∗, K⊆E1∗; so claim 3 of Basic Properties of a Measure gives that all the quantities below are real and
λn(Dk)=λn(E1∗)−λn(Ek∗),λn(E1∗∖K)=λn(E1∗)−λn(K).
Every λn(Dk) is real and bounded above by λn(E1∗) by claim 2 of Basic Properties of a Measure, so claim 5 there shows that (λn(Dk))k∈N converges to λn(E1∗∖K). Subtracting the two displayed identities,
λn(Dk)−λn(E1∗∖K)=λn(K)−λn(Ek∗).
Let ε∈R with 0<ε. By convergence there is k∈N with ∣λn(K)−λn(Ek∗)∣<ε, and in particular λn(Ek∗)≤λn(K)+ε. This completes the proof of claim 4.