In number theory, a Wolstenholme prime is a special type of prime number satisfying a stronger version of Wolstenholme's theorem. Wolstenholme's theorem is a congruence relation satisfied by all prime numbers greater than 3. Wolstenholme primes are named after mathematician Joseph Wolstenholme, who first described this theorem in the 19th century.
Named after | Joseph Wolstenholme |
---|---|
Publication year | 1995[1] |
Author of publication | McIntosh, R. J. |
No. of known terms | 2 |
Conjectured no. of terms | Infinite |
Subsequence of | Irregular primes |
First terms | 16843, 2124679 |
Largest known term | 2124679 |
OEIS index |
|
Interest in these primes first arose due to their connection with Fermat's Last Theorem. Wolstenholme primes are also related to other special classes of numbers, studied in the hope to be able to generalize a proof for the truth of the theorem to all positive integers greater than two.
The only two known Wolstenholme primes are 16843 and 2124679 (sequence A088164 in the OEIS). There are no other Wolstenholme primes less than 109.[2]
Are there any Wolstenholme primes other than 16843 and 2124679?
Wolstenholme prime can be defined in a number of equivalent ways.
A Wolstenholme prime is a prime number p > 7 that satisfies the congruence
where the expression in left-hand side denotes a binomial coefficient.[3] In comparison, Wolstenholme's theorem states that for every prime p > 3 the following congruence holds:
A Wolstenholme prime is a prime p that divides the numerator of the Bernoulli number Bp−3.[4][5][6] The Wolstenholme primes therefore form a subset of the irregular primes.
A Wolstenholme prime is a prime p such that (p, p–3) is an irregular pair.[7][8]
A Wolstenholme prime is a prime p such that[9]
i.e. the numerator of the harmonic number expressed in lowest terms is divisible by p3.
The search for Wolstenholme primes began in the 1960s and continued over the following decades, with the latest results published in 2007. The first Wolstenholme prime 16843 was found in 1964, although it was not explicitly reported at that time.[10] The 1964 discovery was later independently confirmed in the 1970s. This remained the only known example of such a prime for almost 20 years, until the discovery announcement of the second Wolstenholme prime 2124679 in 1993.[11] Up to 1.2×107, no further Wolstenholme primes were found.[12] This was later extended to 2×108 by McIntosh in 1995 [5] and Trevisan & Weber were able to reach 2.5×108.[13] The latest result as of 2007 is that there are only those two Wolstenholme primes up to 109.[14]
It is conjectured that infinitely many Wolstenholme primes exist. It is conjectured that the number of Wolstenholme primes ≤ x is about ln ln x, where ln denotes the natural logarithm. For each prime p ≥ 5, the Wolstenholme quotient is defined as
Clearly, p is a Wolstenholme prime if and only if Wp ≡ 0 (mod p). Empirically one may assume that the remainders of Wp modulo p are uniformly distributed in the set {0, 1, ..., p–1}. By this reasoning, the probability that the remainder takes on a particular value (e.g., 0) is about 1/p.[5]