Landau's problems are four conjectures about prime numbers which were unsolved at the time Edmund Landau presented them at the International Congress of Mathematicians in 1912.

They include:

  1. Goldbach's conjecture
  2. Twin prime conjecture
  3. Legendre's conjecture
  4. Conjecture that there are infinitely many primes of the form p = n^2 + 1.

It is a hundred years later and I think all four problems are still unsolved.

Is this statement true? Does anyone know if there is any recent progress on any of these?

Thanks for any additional insights or pointers.


  • 3
    $\begingroup$ According to Wiki, the first two are unsolved. Per OEIS 002496 the last is unsolved. And Legendre's conjecture is still a conjecture. $\endgroup$ – daniel Nov 26 '12 at 23:53
  • $\begingroup$ I noticed this (arxiv.org/abs/1211.6046) was uploaded to arxiv very recently. Not being an expert I have absolutely no idea as to whether it is correct or not, however it may at least contain references to other recent progress on Legendre's conjecture. $\endgroup$ – Alex J Best Nov 27 '12 at 23:03
  • $\begingroup$ He cites one paper, by Rosser-Schoenfeld, from which I think he uses a bound for Chebyshev's first function. $\endgroup$ – daniel Nov 27 '12 at 23:43
  • $\begingroup$ Yeah I noticed that after I posted, perhaps not so useful after all! Still it seemed like too much of a coincidence not to post after it came in the arxiv email this morning. $\endgroup$ – Alex J Best Nov 27 '12 at 23:56
  • 1
    $\begingroup$ Very useful if it's right. $\endgroup$ – daniel Nov 28 '12 at 0:01

The answer to the first two questions can be found in (for example) Wikipedia by searching under "twin primes" and "Goldbach," respectively. According to the articles there, the first two problems remain unsolved. OEIS sequence A002496 lists the elements of the sequence $\{n^2 +1 | n^2+1 = p\}$ with a comment (in 2001) that the sequence is conjectured but not proven to be infinite.

Specialists in questions 1, 2, and 4 might address the question of "recent progress" with respect to these. The questions themselves pre-date Landau's discussion. According to Pintz, below, the "twin prime" conjecture may date to the time of Euclid. There are hundreds of papers dealing with these questions. They are easy to understand but extremely difficult to prove.

Legendre's conjecture remains unproven. Papers by Ingham (1937) and Cheng (2010) prove there is a prime between $n^3$ and $(n+1)^3,$ and a 1975 paper by Chen proves there is a prime or a near-prime on square intervals. There are dozens of other papers with nice results about primes on (sort of) small intervals but square intervals are currently out of reach.

Questions 3 and 4 are not unrelated, since a sequence of the type in question 4 could not impinge on a (hypothetical) countable sequence of prime-free square intervals.

While I have only started reading it, there is an authoritative review of Landau's problems in: Janos Pintz, Landau's Problems on Primes, Journal de theorie des nombres de Bordeaux, 2009.

[My computer is giving me a "math processing error" so I am leaving the answer here. Edits invited. Will add to this if a reboot solves.]

  • $\begingroup$ "They are easy to understand but extremely difficult." What does that mean?? $\endgroup$ – TonyK Nov 27 '12 at 1:30
  • 5
    $\begingroup$ I think he means they are very easy to understand (meaning), but so far, have proven too difficult to figure out a proof. $\endgroup$ – Amzoti Nov 27 '12 at 1:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.