Write ω(n) for the number of distinct prime divisors of n, with
ω(1)=0. The external input is
Erdős–Pomerance–Sárközy Lemma 1:
with modulus m=1, additive function ω and K=1, it gives for X≥2
n≤X∑(ω(n)−A(X))2≪X(A(X)+1),A(X)=p≤X∑p1.(1)
The condition concerning primes dividing the modulus is vacuous.
The function is nonnegative and additive, and ω(pa)=1.
Euler's divergence of the prime reciprocal sum implies A(X)→∞;
the elementary bound A(X)≤1+logX will also suffice below.
For sufficiently large integers j, put
Xj=2j+2,Aj=A(Xj),Hj=⌈Aj3/4⌉,Dj={1≤n≤Xj:∣ω(n)−Aj∣>Hj}.
By (1) and Chebyshev's inequality,
∣Dj∣≪XjHj2Aj+1=o(2j).(2)
For an integer u∈[2j,2j+1), let bj(u) count those n∈Dj
whose image n+ω(n) belongs to (u,u+Hj+1). Each integer image
belongs to at most Hj+1 of these windows. Hence
2j≤u<2j+1∑bj(u)≤(Hj+1)∣Dj∣.
Set Ej={u∈[2j,2j+1)∩Z:bj(u)>Hj+1}.
Then ∣Ej∣≤∣Dj∣=o(2j). Both the union E=⋃jEj and
D=j⋃(Dj∩[2j,2j+1))
have natural density zero. Indeed, for any ε>0 their
dyadic counts are at most ε2j for all sufficiently large
j. Summing complete dyadic intervals and bounding the last partial
one by its whole interval gives at most 2εx+O(1) elements
below x.
If u∈/Ej, consider all positive integers n with
n+ω(n)∈(u,u+Hj+1). They satisfy n≤Xj for sufficiently
large j, since ω(n)≥0 and
2j+1+Hj+1<2j+2. The latter inequality follows from
Hj=O((j+1)3/4+1). Those n outside Dj obey
u−Aj−Hj<n<u+2Hj+1−Aj.
There are at most 3Hj+2 integers in this interval. The exceptional
inputs contribute at most bj(u)≤Hj+1. The whole count is thus
at most 4Hj+3.
For real x≥1, put u=⌊x⌋ and choose the dyadic
index j with 2j≤u<2j+1. Define
F(x)={Hj,2(u+1),u∈/Ej,u∈Ej.(3)
For the finitely many initial u not covered by the construction use
the second alternative. Both alternatives tend to infinity with x.
At a nonexceptional location,
(x,x+F(x))⊂(u,u+Hj+1), so the normalized count is at most
4+3/Hj≤7. At an exceptional location, every counted positive
integer satisfies n<x+F(x)<3(u+1). Its normalized count is therefore
less than 3/2. This proves the asserted bound at every sufficiently
large real location, including nonintegral ones.
Finally, let G be the positive integers outside E∪D, omitting
the finitely many initial values. This set has natural density one.
For u∈G∩[2j,2j+1), eventually
ω(u)≥Aj−Hj≥Aj/2,ω(u)F(u)≤Aj2Hj⟶0.
Thus (3) satisfies the density-one width condition but its normalized
image count cannot tend to infinity along any sequence of locations.
This establishes L34 exactly as stated.