A useful technique that stumped me on a couple of proofs recently was the idea of “consecutive integers”. A friend had the following proposition: The result of subtracting a number from its cube is divisible by 3.

We can also state the problem in symbols: for any a∈Z, it is true that 3|(a3−a). To quickly check if this makes sense, we also write down the following examples:

−23−(−2)=−6=3⋅−2 −13−(−1)=0=3⋅0 03−0=0=3⋅0 13−1=0=3⋅0 23−2=6=3⋅2 33−3=24=3⋅8 43−4=60=3⋅20

We can use the idea that (a3−a) is divisible by 3, if it is equal to 3⋅b where b∈Z.

Proof. Suppose a∈Z. We start with a3−a.
Pulling out a, we get a(a2−1).
The next step is to spot the difference of squares: a(a2−12)=a(a−1)(a+1).

At this point we get stuck, as there is not much left to do arithmetically. We still don’t have the 3 or the b we’re looking for.

However, there are some hints to help us. The first thing to note is that we have three terms: a,(a−1),(a+1). We can further re-arrange these to get: (a−1)(a+0)(a+1). These are three consecutive integers.

Given n consecutive integers, we can be sure that exactly one of those integers is divisible by n. Thus exactly one of {(a−1),(a+0),(a+1)} is divisible by 3. Which means that we can write the expression in three possible cases, where c∈Z:

(3c)(a+0)(a+1) (a−1)(3c)(a+1) (a−1)(a+0)(3c)

For each of these cases, it is trivial to extract b∈Z such that the expression becomes 3b.

(3c)(a+0)(a+1)=3b, where b=c(a+0)(a+1)

(a−1)(3c)(a+1)=3b, where b=(a−1)c(a+1)

(a−1)(a+0)(3c)=3b, where b=(a−1)(a+0)c

For the three forms of (a3−a), there is a b such that (a3−a)=3b.
Therefore, (a3−3) is divisible by 3. ◻

Here is another problem that relies on a similar appearance of consecutive integers. It can be found in the fifth chapter of the amazing Book of Proof.

If n is odd, then 8|(n2−1). We will of course try to show that n2−1 is equal to 8c where c∈Z.

Proof. Suppose n is odd. We can write n=2a+1, for some a∈Z. The expression n2−1 can be simplified as:

n2−1=(2a+1)2−1=4a2+4a=4a(a+1)

Again we get stuck arithmetically, the required ‘8’ is nowhere to be found in 4a(a+1) . The hint lies in the possibility that we need to pull out a ‘2’ from somewhere such that the ‘4’ can be written as an ‘8’. On careful observation, we notice that the two terms a and (a+1) are consecutive integers. Therefore, exactly one of those must be divisible by 2. This leads to two cases, where b∈Z:

4(2b)(a+1) 4(a)(2b)

In either case, we can extract a c∈Z such that:

4(2b)(a+1)=8b(a+1)=8c, where c=b(a+1)

4(a)(2b)=8ab=8c, where c=ab

Here too, the two forms for (n2−1) are equal to 8c.
Therefore, (n2−1) is divisible by 8. ◻