The density of primes \(P\), such that \(-1\) is a residue modulo \(P\) of two consecutive Fibonacci numbers, is \(2/3\) (Q1974414)

Let \(U = (U_n)_{n\geq 0}\) be an \(m^{\text{th}}\) order \((m\geq 2)\) linear recurrence sequence of integers. We say that a prime \(p\) is a divisor of the sequence \(U\) if \(p\) is a divisor of some \(U_n(\neq 0)\). If \(p\) divides some \(m- 1\) consecutive terms, then \(p\) is called a maximal prime divisor of the sequence. Recently many authors investigated the density of the set of prime and maximal prime divisors. E.g. in 1985 J. C. Lagarias proved that the set of primes dividing the Lucas sequence, which is the associate sequence of the Fibonacci sequence, has density 2/3. In the present paper the author proves some similar interesting results concerning linear recursive sequences of order 3. Among others he proves that the density of prime maximal divisors of the sequence \((1 + F_n)\) is 2/3, where \((F_n)\) is the Fibonacci sequence. The following theorem is also proved: ``The pairs (1,1) and \((-1,1)\) are consecutive Fibonacci residues for all primes, while the pairs \((-1, -1)\) and \((1, -1)\) both occur as consecutive Fibonacci residues for the same set of primes; this set has density 2/3.'' We note that the sequence \((1 + F_n)\) is a third order linear recurrence sequence since its characteristic polynomial is \(f(x) = (x - 1)(x^2 -x-1)\).