The -tuple
conjecture, also called the first of the Hardy-Littlewood
conjectures, states that the asymptotic number of prime
constellations can be computed explicitly. In particular, unless there is a trivial
divisibility condition that stops
,
, ...,
from consisting of primes
infinitely often, then such prime constellations
will occur with an asymptotic density which is computable in terms of
, ...,
. Let
be integers.
For each odd prime number
, let
be the number of distinct residues
of 0,
, ...,
modulo
. If
for every such
, the
-tuple conjecture predicts that the number of primes
such that
,
, ...,
are all prime is
|
(1)
|
where
|
(2)
|
and the product is over the odd primes
(Halberstam and Richert 1974, Odlyzko et al. 1999). If
for some
, then any prime tuple must contain
, so only finitely many can occur. For
the constant reduces to
|
(3)
|
The conjecture remains open. If it holds, then the second Hardy-Littlewood conjecture fails (Hensley and Richards 1974).
|
(4)
|
is the case . Here
counts twin primes whose
smaller prime number is at most
, and
is the twin primes constant.
The following special case of the conjecture is sometimes known as the prime patterns conjecture. Let
be a finite set of integers.
Then it is conjectured that there exist infinitely many
for which
are all prime iff
does not include all the residues
of any prime. This conjecture also implies that there
are arbitrarily long arithmetic progressions
of primes.