cp's OEIS Frontend

D. E. Knuth uses the term 'prime strings' for Lyndon words because of the fundamental theorem stating the unique factorization of strings into nonincreasing prime strings (see Knuth 7.2.1.1). With this terminology T(n,k) is the number of k-ary n-tuples (a_1,...,a_n) such that the string a_1...a_n is prime. - Peter Luschny, Aug 14 2012

Also, for k a power of a prime, the number of monic irreducible polynomials of degree n over GF(k). - Andrew Howroyd, Dec 23 2017

An equivalent description: Array read by antidiagonals: T(n,k) = number of conjugacy classes of primitive words of length k >= 1 over an alphabet of size n >= 1.

There are a few incorrect values in Table 1 in the Perrin-Reutenauer paper (Christophe Reutenauer, personal communication), see A294438. - Lars Blomberg, Dec 05 2017

The fact that T(3,4) = 20 coincides with the number of the amino acids encoded by DNA made Francis Crick, John Griffith and Leslie Orgel conjecture in 1957 that the genetic code is a comma-free code, which later turned out to be false. [Hayes] - Andrey Zabolotskiy, Mar 24 2018

For prime rows, these appear to be evaluations of Moreau's necklace polynomials at the integers with several combinatorial interpretations (see Wikipedia link). - Tom Copeland, Oct 20 2014

From Petros Hadjicostas, Nov 05 2017: (Start)

The g.f. for column k follows easily from I. Gessel's formulas for this sequence. Since S(1,k) = k-1, we have T(1,k) = k + S(1,k) - (k - 1). Thus, Sum_{n >= 1} T(n,k)*x^n = k*x + Sum_{n >= 1} (1/n)*Sum_{d|n} (k - 1)^d*phi(n/d)*x^n - Sum_{s=0} (k-1)*x^{2*s+1}. Letting m = n/d, we get that (column k g.f.) = k*x - (k - 1)*x/(1 -x^2) + Sum_{m >= 1} (phi(m)/m)*Sum_{d >= 1}((k - 1)*x^m)^d/d. But Sum_{d>=1} z^d/d = -log(1 - z), and so (column k g.f.) = k*x - (k - 1)*x/(1 - x^2) - Sum_{m >= 1} (phi(m)/m)*log(1 - (k - 1)*x^m).

The other formula for the g.f. of column k of this sequence follows from the formula Sum_{n >= 1} (phi(n)/n)*log(1 + t^n) = t/(1 - t^2), which in turn follows from the well-known series Sum_{n >= 1} phi(n)*t^n/(1 + t^n) = t*(1 + t^2)/(1 - t^2)^2.

The extra term k*x in the g.f. for column k is due to the fact that we conventionally assume that a necklace with only one bead, colored with one of the k colors available, is such that there are "no adjacent beads having the same color" (even though, strictly speaking, a single bead is adjacent to itself when we go around the circle of the necklace).

One can use the g.f. for column k to derive the so-called "Empirical for row n" formulae that are denoted by a(k) and given in the formula section below (from n = 1 to n = 7). For example, for n = 3, a(k) = a(k, x=0), where a(k, x) = (1/3!)*d^3/dx^3 (column k g.f.). Here, d^3/dx^3 stands for "third derivative w.r.t. x". If we let f(x) = x/(1 - x^2) and g(x, m) = -log(1 - (k - 1)*x^m), then f^{(3)}(0) = 6, while g^{(3)}(0,m) = 2*(k - 1)^3 for m = 1, 0 for m=2, 6*(k - 1) for m = 3, and 0 for m >= 4. Then, a(k) = (1/6)*(-6*(k - 1) + 2*(k - 1)^3 + (2/3)*6*(k - 1)) = (1/3)*k^3 - k^2 + (2/3)*k. Using this method, one can derive an "Empirical for row n" formula for a(k) for any positive integer n. (End)

This sequence would be better defined as a(n) = (n^5-n)/5 with offset 0, which is an integer by Fermat's little theorem. - N. J. A. Sloane, Nov 13 2023

If p is a prime, (n^p-n)/p is an integer by Fermat's Little Theorem. For p = 2, 3, 5, 7, ..., this gives A000217, A007290, A208536, A208537, this sequence, A366489, A366490, A366491.

See A366488 for further information.

See A366488 for further information.

See A366488 for further information.

Row 7 of A208544.