Riemann hypothesis

From Wikipedia, the free encyclopedia

https://en.wikipedia.org/wiki/Riemann_hypothesis

This plot of Riemann's zeta (ζ) function (here with argument z) shows trivial zeros where ζ(z) = 0, a pole where ζ(z) = ${\displaystyle \mathrm{\infty }}$, the critical line of nontrivial zeros with Re(z) = 1/2 and slopes of absolute values.

The real part (red) and imaginary part (blue) of the Riemann zeta function ζ(s) along the critical line in the complex plane with real part Re(s) = 1/2. The first nontrivial zeros, where ζ(s) equals zero, occur where both curves touch the horizontal x-axis, for complex numbers with imaginary parts Im(s) equalling ±14.135, ±21.022 and ±25.011.

Riemann zeta function along the critical line with Re(s) = 1/2. Real values are shown on the horizontal axis and imaginary values are on the vertical axis). Re(ζ(1/2 + it), Im(ζ(1/2 + it) is plotted with t ranging between −30 and 30.
The curve starts for t = -30 at ζ(1/2 - 30 i) = -0.12 + 0.58 i, and end symmetrically below the starting point at ζ(1/2 + 30 i) = -0.12 - 0.58 i.
Six zeros of ζ(s) are found along the trajectory when the origin (0,0) is traversed, corresponding to imaginary parts Im(s) = ±14.135, ±21.022 and ±25.011.

In mathematics, the Riemann hypothesis is the conjecture that the Riemann zeta function has its zeros only at the negative even integers and complex numbers with real part 1/2. Many consider it to be the most important unsolved problem in pure mathematics.^[2] It is of great interest in number theory because it implies results about the distribution of prime numbers. It was proposed by Bernhard Riemann (1859), after whom it is named.

The Riemann hypothesis and some of its generalizations, along with Goldbach's conjecture and the twin prime conjecture, make up Hilbert's eighth problem in David Hilbert's list of twenty-three unsolved problems; it is also one of the Clay Mathematics Institute's Millennium Prize Problems, which offers a million dollars to anyone who solves any of them. The name is also used for some closely related analogues, such as the Riemann hypothesis for curves over finite fields.

The Riemann zeta function ζ(s) is a function whose argument s may be any complex number other than 1, and whose values are also complex. It has zeros at the negative even integers; that is, ζ(s) = 0 when s is one of −2, −4, −6, .... These are called its trivial zeros. The zeta function is also zero for other values of s, which are called nontrivial zeros. The Riemann hypothesis is concerned with the locations of these nontrivial zeros, and states that:

The real part of every nontrivial zero of the Riemann zeta function is 1/2.

Thus, if the hypothesis is correct, all the nontrivial zeros lie on the critical line consisting of the complex numbers 1/2 + i t, where t is a real number and i is the imaginary unit.

Riemann zeta function

The Riemann zeta function is defined for complex s with real part greater than 1 by the absolutely convergent infinite series

${\displaystyle \zeta (s)=\sum _{n=1}^{\mathrm{\infty }}\frac{1}{n^s}=\frac{1}{1^s}+\frac{1}{2^s}+\frac{1}{3^s}+\cdots }$

Leonhard Euler already considered this series in the 1730s for real values of s, in conjunction with his solution to the Basel problem. He also proved that it equals the Euler product

${\displaystyle \zeta (s)=\prod _{p\text{ prime}}\frac{1}{1-p^{-s}}=\frac{1}{1-2^{-s}}\cdot \frac{1}{1-3^{-s}}\cdot \frac{1}{1-5^{-s}}\cdot \frac{1}{1-7^{-s}}\cdot \frac{1}{1-{11}^{-s}}\cdots }$

where the infinite product extends over all prime numbers p.^[3]

The Riemann hypothesis discusses zeros outside the region of convergence of this series and Euler product. To make sense of the hypothesis, it is necessary to analytically continue the function to obtain a form that is valid for all complex s. Because the zeta function is meromorphic, all choices of how to perform this analytic continuation will lead to the same result, by the identity theorem. A first step in this continuation observes that the series for the zeta function and the Dirichlet eta function satisfy the relation

${\displaystyle \left(1-\frac{2}{2^s}\right)\zeta (s)=\eta (s)=\sum _{n=1}^{\mathrm{\infty }}\frac{(-1{)}^{n+1}}{n^s}=\frac{1}{1^s}-\frac{1}{2^s}+\frac{1}{3^s}-\cdots ,}$

within the region of convergence for both series. However, the zeta function series on the right converges not just when the real part of s is greater than one, but more generally whenever s has positive real part. Thus, the zeta function can be redefined as ${\displaystyle \eta (s)/(1-2/2^s)}$, extending it from Re(s) > 1 to a larger domain: Re(s) > 0, except for the points where ${\displaystyle 1-2/2^s}$ is zero. These are the points ${\displaystyle s=1+2\pi in/\log 2}$ where ${\displaystyle n}$ can be any nonzero integer; the zeta function can be extended to these values too by taking limits (see Dirichlet eta function § Landau's problem with ζ(s) = η(s)/0 and solutions), giving a finite value for all values of s with positive real part except for the simple pole at s = 1.

In the strip 0 < Re(s) < 1 this extension of the zeta function satisfies the functional equation

${\displaystyle \zeta (s)=2^s{\pi }^{s-1}\sin \left(\frac{\pi s}{2}\right)\mathrm{\Gamma }(1-s)\zeta (1-s).}$

One may then define ζ(s) for all remaining nonzero complex numbers s (Re(s) ≤ 0 and s ≠ 0) by applying this equation outside the strip, and letting ζ(s) equal the right-hand side of the equation whenever s has non-positive real part (and s ≠ 0).

If s is a negative even integer then ζ(s) = 0 because the factor sin(πs/2) vanishes; these are the trivial zeros of the zeta function. (If s is a positive even integer this argument does not apply because the zeros of the sine function are cancelled by the poles of the gamma function as it takes negative integer arguments.)

The value ζ(0) = −1/2 is not determined by the functional equation, but is the limiting value of ζ(s) as s approaches zero. The functional equation also implies that the zeta function has no zeros with negative real part other than the trivial zeros, so all non-trivial zeros lie in the critical strip where s has real part between 0 and 1.

Origin

... es ist sehr wahrscheinlich, dass alle Wurzeln reell sind. Hiervon wäre allerdings ein strenger Beweis zu wünschen; ich habe indess die Aufsuchung desselben nach einigen flüchtigen vergeblichen Versuchen vorläufig bei Seite gelassen, da er für den nächsten Zweck meiner Untersuchung entbehrlich schien.

... it is very probable that all roots are real. Of course one would wish for a rigorous proof here; I have for the time being, after some fleeting vain attempts, provisionally put aside the search for this, as it appears dispensable for the immediate objective of my investigation.

— Riemann's statement of the Riemann hypothesis, from (Riemann 1859). (He was discussing a version of the zeta function, modified so that its roots (zeros) are real rather than on the critical line.)

At the death of Riemann, a note was found among his papers, saying "These properties of ζ(s) (the function in question) are deduced from an expression of it which, however, I did not succeed in simplifying enough to publish it." We still have not the slightest idea of what the expression could be. As to the properties he simply enunciated, some thirty years elapsed before I was able to prove all of them but one [the Riemann Hypothesis itself].

— Jacques Hadamard, The Mathematician's Mind, VIII. Paradoxical Cases of Intuition

Riemann's original motivation for studying the zeta function and its zeros was their occurrence in his explicit formula for the number of primes π(x) less than or equal to a given number x, which he published in his 1859 paper "On the Number of Primes Less Than a Given Magnitude". His formula was given in terms of the related function

${\displaystyle \mathrm{\Pi }(x)=\pi (x)+\frac{1}{2}\pi (x^{\frac{1}{2}})+\frac{1}{3}\pi (x^{\frac{1}{3}})+\frac{1}{4}\pi (x^{\frac{1}{4}})+\frac{1}{5}\pi (x^{\frac{1}{5}})+\frac{1}{6}\pi (x^{\frac{1}{6}})+\cdots }$

which counts the primes and prime powers up to x, counting a prime power pⁿ as 1⁄n. The number of primes can be recovered from this function by using the Möbius inversion formula,

${\displaystyle \begin{array}{rl}\pi (x)& =\sum _{n=1}^{\mathrm{\infty }}\frac{\mu (n)}{n}\mathrm{\Pi }(x^{\frac{1}{n}})\\ & =\mathrm{\Pi }(x)-\frac{1}{2}\mathrm{\Pi }(x^{\frac{1}{2}})-\frac{1}{3}\mathrm{\Pi }(x^{\frac{1}{3}})-\frac{1}{5}\mathrm{\Pi }(x^{\frac{1}{5}})+\frac{1}{6}\mathrm{\Pi }(x^{\frac{1}{6}})-\cdots ,\end{array}}$

where μ is the Möbius function. Riemann's formula is then

${\displaystyle {\mathrm{\Pi }}_0(x)=\mathrm{li}(x)-\sum _{\rho }\mathrm{li}(x^{\rho })-\log 2+{\int }_x^{\mathrm{\infty }}\frac{dt}{t(t^2-1)\log t}}$

where the sum is over the nontrivial zeros of the zeta function and where Π₀ is a slightly modified version of Π that replaces its value at its points of discontinuity by the average of its upper and lower limits:

${\displaystyle {\mathrm{\Pi }}_0(x)=\underset{\epsilon \to 0}{lim}\frac{\mathrm{\Pi }(x-\epsilon )+\mathrm{\Pi }(x+\epsilon )}{2}.}$

The summation in Riemann's formula is not absolutely convergent, but may be evaluated by taking the zeros ρ in order of the absolute value of their imaginary part. The function li occurring in the first term is the (unoffset) logarithmic integral function given by the Cauchy principal value of the divergent integral

${\displaystyle \mathrm{li}(x)={\int }_0^x\frac{dt}{\log t}.}$

The terms li(x^ρ) involving the zeros of the zeta function need some care in their definition as li has branch points at 0 and 1, and are defined (for x > 1) by analytic continuation in the complex variable ρ in the region Re(ρ) > 0, i.e. they should be considered as Ei(ρ log x). The other terms also correspond to zeros: the dominant term li(x) comes from the pole at s = 1, considered as a zero of multiplicity −1, and the remaining small terms come from the trivial zeros. For some graphs of the sums of the first few terms of this series see Riesel & Göhl (1970) or Zagier (1977).

This formula says that the zeros of the Riemann zeta function control the oscillations of primes around their "expected" positions. Riemann knew that the non-trivial zeros of the zeta function were symmetrically distributed about the line s = 1/2 + it, and he knew that all of its non-trivial zeros must lie in the range 0 ≤ Re(s) ≤ 1. He checked that a few of the zeros lay on the critical line with real part 1/2 and suggested that they all do; this is the Riemann hypothesis.

The result has caught the imagination of most mathematicians because it is so unexpected, connecting two seemingly unrelated areas in mathematics; namely, number theory, which is the study of the discrete, and complex analysis, which deals with continuous processes. (Burton 2006, p. 376)

Consequences

The practical uses of the Riemann hypothesis include many propositions known to be true under the Riemann hypothesis, and some that can be shown to be equivalent to the Riemann hypothesis.

Distribution of prime numbers

Riemann's explicit formula for the number of primes less than a given number in terms of a sum over the zeros of the Riemann zeta function says that the magnitude of the oscillations of primes around their expected position is controlled by the real parts of the zeros of the zeta function. In particular the error term in the prime number theorem is closely related to the position of the zeros. For example, if β is the upper bound of the real parts of the zeros, then ${\displaystyle \pi (x)-\mathrm{li}(x)=O\left(x^{\beta }\log x\right).}$ It is already known that 1/2 ≤ β ≤ 1.

Von Koch (1901) proved that the Riemann hypothesis implies the "best possible" bound for the error of the prime number theorem. A precise version of Koch's result, due to Schoenfeld (1976), says that the Riemann hypothesis implies

${\displaystyle |\pi (x)-\mathrm{li}(x)|<\frac{1}{8\pi }\sqrt{x}\log (x),\text{for all }x\ge 2657,}$

where ${\displaystyle \pi (x)}$ is the prime-counting function, ${\displaystyle \mathrm{li}(x)}$ is the logarithmic integral function, and ${\displaystyle \log (x)}$ is the natural logarithm of x.

Schoenfeld (1976) also showed that the Riemann hypothesis implies

${\displaystyle |\psi (x)-x|<\frac{1}{8\pi }\sqrt{x}{\log }^{2}x,\text{for all }x\ge 73.2,}$

where ${\displaystyle \psi (x)}$ is Chebyshev's second function.

Dudek (2014) proved that the Riemann hypothesis implies that for all ${\displaystyle x\ge 2}$ there is a prime ${\displaystyle p}$ satisfying

${\displaystyle x-\frac{4}{\pi }\sqrt{x}\log x<p\le x}$.

This is an explicit version of a theorem of Cramér.

Growth of arithmetic functions

The Riemann hypothesis implies strong bounds on the growth of many other arithmetic functions, in addition to the primes counting function above.

One example involves the Möbius function μ. The statement that the equation

${\displaystyle \frac{1}{\zeta (s)}=\sum _{n=1}^{\mathrm{\infty }}\frac{\mu (n)}{n^s}}$

is valid for every s with real part greater than 1/2, with the sum on the right hand side converging, is equivalent to the Riemann hypothesis. From this we can also conclude that if the Mertens function is defined by

${\displaystyle M(x)=\sum _{n\le x}\mu (n)}$

then the claim that

${\displaystyle M(x)=O\left(x^{\frac{1}{2}+\epsilon }\right)}$

for every positive ε is equivalent to the Riemann hypothesis (J. E. Littlewood, 1912; see for instance: paragraph 14.25 in Titchmarsh (1986)). (For the meaning of these symbols, see Big O notation.) The determinant of the order n Redheffer matrix is equal to M(n), so the Riemann hypothesis can also be stated as a condition on the growth of these determinants. Littlewood's result has been improved several times since then, by Edmund Landau, Edward Charles Titchmarsh, Helmut Maier and Hugh Montgomery, and Kannan Soundararajan. Soundararajan's result is that

${\displaystyle M(x)=O(x^{1/2}\exp ((\log x{)}^{1/2}(\log \log x{)}^{14}).}$

The Riemann hypothesis puts a rather tight bound on the growth of M, since Odlyzko & te Riele (1985) disproved the slightly stronger Mertens conjecture

${\displaystyle |M(x)|\le \sqrt{x}.}$

Another closely related result is due to Björner (2011), that the Riemann hypothesis is equivalent to the statement that the Euler characteristic of the simplicial complex determined by the lattice of integers under divisibility is ${\displaystyle o(n^{1/2+ϵ})}$ for all ${\displaystyle ϵ>0}$ (see incidence algebra).

The Riemann hypothesis is equivalent to many other conjectures about the rate of growth of other arithmetic functions aside from μ(n). A typical example is Robin's theorem, which states that if σ(n) is the sigma function, given by

${\displaystyle \sigma (n)=\sum _{d\mid n}d}$

then

${\displaystyle \sigma (n)<e^{\gamma }n\log \log n}$

for all n > 5040 if and only if the Riemann hypothesis is true, where γ is the Euler–Mascheroni constant.

A related bound was given by Jeffrey Lagarias in 2002, who proved that the Riemann hypothesis is equivalent to the statement that:

${\displaystyle \sigma (n)<H_n+\log (H_n)e^{H_n}}$

for every natural number n > 1, where ${\displaystyle H_n}$ is the nth harmonic number.

The Riemann hypothesis is also true if and only if the inequality

${\displaystyle \frac{n}{\phi (n)}<e^{\gamma }\log \log n+\frac{e^{\gamma }(4+\gamma -\log 4\pi )}{\sqrt{\log n}}}$

is true for all n ≥ p₁₂₀₅₆₉# where φ(n) is Euler's totient function and p₁₂₀₅₆₉# is the product of the first 120569 primes.

Another example was found by Jérôme Franel, and extended by Landau (see Franel & Landau (1924)). The Riemann hypothesis is equivalent to several statements showing that the terms of the Farey sequence are fairly regular. One such equivalence is as follows: if F_n is the Farey sequence of order n, beginning with 1/n and up to 1/1, then the claim that for all ε > 0

${\displaystyle \sum _{i=1}^m|F_n(i)-\frac{i}{m}|=O\left(n^{\frac{1}{2}+ϵ}\right)}$

is equivalent to the Riemann hypothesis. Here

${\displaystyle m=\sum _{i=1}^n\varphi (i)}$

is the number of terms in the Farey sequence of order n.

For an example from group theory, if g(n) is Landau's function given by the maximal order of elements of the symmetric group S_n of degree n, then Massias, Nicolas & Robin (1988) showed that the Riemann hypothesis is equivalent to the bound

${\displaystyle \log g(n)<\sqrt{{\mathrm{Li}}^{-1}(n)}}$

for all sufficiently large n.

Lindelöf hypothesis and growth of the zeta function

The Riemann hypothesis has various weaker consequences as well; one is the Lindelöf hypothesis on the rate of growth of the zeta function on the critical line, which says that, for any ε > 0,

${\displaystyle \zeta \left(\frac{1}{2}+it\right)=O(t^{\epsilon }),}$

as ${\displaystyle t\to \mathrm{\infty }}$.

The Riemann hypothesis also implies quite sharp bounds for the growth rate of the zeta function in other regions of the critical strip. For example, it implies that

${\displaystyle e^{\gamma }\le \underset{t\to +\mathrm{\infty }}{lim sup}\frac{|\zeta (1+it)|}{\log \log t}\le 2e^{\gamma }}$

${\displaystyle \frac{6}{{\pi }^2}{e}^{\gamma }\le \underset{t\to +\mathrm{\infty }}{lim sup}\frac{1/|\zeta (1+it)|}{\log \log t}\le \frac{12}{{\pi }^2}{e}^{\gamma }}$

so the growth rate of ζ(1+it) and its inverse would be known up to a factor of 2.

Large prime gap conjecture

The prime number theorem implies that on average, the gap between the prime p and its successor is log p. However, some gaps between primes may be much larger than the average. Cramér proved that, assuming the Riemann hypothesis, every gap is O(√p log p). This is a case in which even the best bound that can be proved using the Riemann hypothesis is far weaker than what seems true: Cramér's conjecture implies that every gap is O((log p)²), which, while larger than the average gap, is far smaller than the bound implied by the Riemann hypothesis. Numerical evidence supports Cramér's conjecture.

Analytic criteria equivalent to the Riemann hypothesis

Many statements equivalent to the Riemann hypothesis have been found, though so far none of them have led to much progress in proving (or disproving) it. Some typical examples are as follows. (Others involve the divisor function σ(n).)

The Riesz criterion was given by Riesz (1916), to the effect that the bound

${\displaystyle -\sum _{k=1}^{\mathrm{\infty }}\frac{(-x{)}^k}{(k-1)!\zeta (2k)}=O\left(x^{\frac{1}{4}+ϵ}\right)}$

holds for all ε > 0 if and only if the Riemann hypothesis holds. See also the Hardy–Littlewood criterion.

Nyman (1950) proved that the Riemann hypothesis is true if and only if the space of functions of the form

${\displaystyle f(x)=\sum _{\nu =1}^n{c}_{\nu }\rho \left(\frac{{\theta }_{\nu }}{x}\right)}$

where ρ(z) is the fractional part of z, 0 ≤ θ_ν ≤ 1, and

${\displaystyle \sum _{\nu =1}^n{c}_{\nu }{\theta }_{\nu }=0}$,

is dense in the Hilbert space L²(0,1) of square-integrable functions on the unit interval. Beurling (1955) extended this by showing that the zeta function has no zeros with real part greater than 1/p if and only if this function space is dense in L^p(0,1). This Nyman-Beurling criterion was strengthened by Baez-Duarte to the case where ${\displaystyle {\theta }_{\nu }\in \{1/k{\}}_{k\ge 1}}$.

Salem (1953) showed that the Riemann hypothesis is true if and only if the integral equation

${\displaystyle {\int }_0^{\mathrm{\infty }}\frac{z^{-\sigma -1}\varphi (z)}{e^{x/z}+1}dz=0}$

has no non-trivial bounded solutions ${\displaystyle \varphi }$ for ${\displaystyle 1/2<\sigma <1}$.

Weil's criterion is the statement that the positivity of a certain function is equivalent to the Riemann hypothesis. Related is Li's criterion, a statement that the positivity of a certain sequence of numbers is equivalent to the Riemann hypothesis.

Speiser (1934) proved that the Riemann hypothesis is equivalent to the statement that ${\displaystyle {\zeta }^{\prime }(s)}$, the derivative of ${\displaystyle \zeta (s)}$, has no zeros in the strip

${\displaystyle 0<\mathrm{\Re }(s)<\frac{1}{2}.}$

That ${\displaystyle \zeta (s)}$ has only simple zeros on the critical line is equivalent to its derivative having no zeros on the critical line.

The Farey sequence provides two equivalences, due to Jerome Franel and Edmund Landau in 1924.

The de Bruijn–Newman constant denoted by Λ and named after Nicolaas Govert de Bruijn and Charles M. Newman, is defined as the unique real number such that the function

${\displaystyle H(\lambda ,z):={\int }_0^{\mathrm{\infty }}{e}^{\lambda u^2}\mathrm{\Phi }(u)\cos (zu)du}$,

that is parametrised by a real parameter λ, has a complex variable z and is defined using a super-exponentially decaying function

${\displaystyle \mathrm{\Phi }(u)=\sum _{n=1}^{\mathrm{\infty }}(2{\pi }^2{n}^4{e}^{9u}-3\pi n^2{e}^{5u})e^{-\pi n^2{e}^{4u}}}$.

has only real zeros if and only if λ ≥ Λ. Since the Riemann hypothesis is equivalent to the claim that all the zeroes of H(0, z) are real, the Riemann hypothesis is equivalent to the conjecture that ${\displaystyle \mathrm{\Lambda }\le 0}$. Brad Rodgers and Terence Tao discovered the equivalence is actually ${\displaystyle \mathrm{\Lambda }=0}$ by proving zero to be the lower bound of the constant. Proving zero is also the upper bound would therefore prove the Riemann hypothesis. As of April 2020 the upper bound is ${\displaystyle \mathrm{\Lambda }\le 0.2}$.

Consequences of the generalized Riemann hypothesis

Several applications use the generalized Riemann hypothesis for Dirichlet L-series or zeta functions of number fields rather than just the Riemann hypothesis. Many basic properties of the Riemann zeta function can easily be generalized to all Dirichlet L-series, so it is plausible that a method that proves the Riemann hypothesis for the Riemann zeta function would also work for the generalized Riemann hypothesis for Dirichlet L-functions. Several results first proved using the generalized Riemann hypothesis were later given unconditional proofs without using it, though these were usually much harder. Many of the consequences on the following list are taken from Conrad (2010).

• In 1913, Grönwall showed that the generalized Riemann hypothesis implies that Gauss's list of imaginary quadratic fields with class number 1 is complete, though Baker, Stark and Heegner later gave unconditional proofs of this without using the generalized Riemann hypothesis.
• In 1917, Hardy and Littlewood showed that the generalized Riemann hypothesis implies a conjecture of Chebyshev that
  $${\displaystyle \underset{x\to 1^{-}}{lim}\sum _{p>2}(-1{)}^{(p+1)/2}{x}^p=+\mathrm{\infty },}$$

  
  which says that primes 3 mod 4 are more common than primes 1 mod 4 in some sense. (For related results, see Prime number theorem § Prime number race.)
• In 1923, Hardy and Littlewood showed that the generalized Riemann hypothesis implies a weak form of the Goldbach conjecture for odd numbers: that every sufficiently large odd number is the sum of three primes, though in 1937 Vinogradov gave an unconditional proof. In 1997 Deshouillers, Effinger, te Riele, and Zinoviev showed that the generalized Riemann hypothesis implies that every odd number greater than 5 is the sum of three primes. In 2013 Harald Helfgott proved the ternary Goldbach conjecture without the GRH dependence, subject to some extensive calculations completed with the help of David J. Platt.
• In 1934, Chowla showed that the generalized Riemann hypothesis implies that the first prime in the arithmetic progression a mod m is at most Km²log(m)² for some fixed constant K.
• In 1967, Hooley showed that the generalized Riemann hypothesis implies Artin's conjecture on primitive roots.
• In 1973, Weinberger showed that the generalized Riemann hypothesis implies that Euler's list of idoneal numbers is complete.
• Weinberger (1973) showed that the generalized Riemann hypothesis for the zeta functions of all algebraic number fields implies that any number field with class number 1 is either Euclidean or an imaginary quadratic number field of discriminant −19, −43, −67, or −163.
• In 1976, G. Miller showed that the generalized Riemann hypothesis implies that one can test if a number is prime in polynomial time via the Miller test. In 2002, Manindra Agrawal, Neeraj Kayal and Nitin Saxena proved this result unconditionally using the AKS primality test.
• Odlyzko (1990) discussed how the generalized Riemann hypothesis can be used to give sharper estimates for discriminants and class numbers of number fields.
• Ono & Soundararajan (1997) showed that the generalized Riemann hypothesis implies that Ramanujan's integral quadratic form x² + y² + 10z² represents all integers that it represents locally, with exactly 18 exceptions.
• In 2021, Alexander (Alex) Dunn and Maksym Radziwill proved Patterson's conjecture under the assumption of the GRH.

Excluded middle

Some consequences of the RH are also consequences of its negation, and are thus theorems. In their discussion of the Hecke, Deuring, Mordell, Heilbronn theorem, Ireland & Rosen (1990, p. 359) say

The method of proof here is truly amazing. If the generalized Riemann hypothesis is true, then the theorem is true. If the generalized Riemann hypothesis is false, then the theorem is true. Thus, the theorem is true!!     (punctuation in original)

Care should be taken to understand what is meant by saying the generalized Riemann hypothesis is false: one should specify exactly which class of Dirichlet series has a counterexample.

Littlewood's theorem

This concerns the sign of the error in the prime number theorem. It has been computed that π(x) < li(x) for all x ≤ 10²⁵ (see this table), and no value of x is known for which π(x) > li(x).

In 1914 Littlewood proved that there are arbitrarily large values of x for which

${\displaystyle \pi (x)>\mathrm{li}(x)+\frac{1}{3}\frac{\sqrt{x}}{\log x}\log \log \log x,}$

and that there are also arbitrarily large values of x for which

${\displaystyle \pi (x)<\mathrm{li}(x)-\frac{1}{3}\frac{\sqrt{x}}{\log x}\log \log \log x.}$

Thus the difference π(x) − li(x) changes sign infinitely many times. Skewes' number is an estimate of the value of x corresponding to the first sign change.

Littlewood's proof is divided into two cases: the RH is assumed false (about half a page of Ingham 1932, Chapt. V), and the RH is assumed true (about a dozen pages). Stanisław Knapowski (1962) followed this up with a paper on the number of times ${\displaystyle \mathrm{\Delta }(n)}$ changes sign in the interval ${\displaystyle \mathrm{\Delta }(n)}$.

Gauss's class number conjecture

This is the conjecture (first stated in article 303 of Gauss's Disquisitiones Arithmeticae) that there are only finitely many imaginary quadratic fields with a given class number. One way to prove it would be to show that as the discriminant D → −∞ the class number h(D) → ∞.

The following sequence of theorems involving the Riemann hypothesis is described in Ireland & Rosen 1990, pp. 358–361:

Theorem (Hecke; 1918) — Let D < 0 be the discriminant of an imaginary quadratic number field K. Assume the generalized Riemann hypothesis for L-functions of all imaginary quadratic Dirichlet characters. Then there is an absolute constant C such that

$${\displaystyle h(D)>C\frac{\sqrt{|D|}}{\log |D|}.}$$

Theorem (Deuring; 1933) — If the RH is false then h(D) > 1 if |D| is sufficiently large.

Theorem (Mordell; 1934) — If the RH is false then h(D) → ∞ as D → −∞.

Theorem (Heilbronn; 1934) — If the generalized RH is false for the L-function of some imaginary quadratic Dirichlet character then h(D) → ∞ as D → −∞.

(In the work of Hecke and Heilbronn, the only L-functions that occur are those attached to imaginary quadratic characters, and it is only for those L-functions that GRH is true or GRH is false is intended; a failure of GRH for the L-function of a cubic Dirichlet character would, strictly speaking, mean GRH is false, but that was not the kind of failure of GRH that Heilbronn had in mind, so his assumption was more restricted than simply GRH is false.)

In 1935, Carl Siegel later strengthened the result without using RH or GRH in any way.

Growth of Euler's totient

In 1983 J. L. Nicolas proved that

$${\displaystyle \phi (n)<e^{-\gamma }\frac{n}{\log \log n}}$$

for infinitely many n, where φ(n) is Euler's totient function and γ is Euler's constant. Ribenboim remarks that: "The method of proof is interesting, in that the inequality is shown first under the assumption that the Riemann hypothesis is true, secondly under the contrary assumption."

Generalizations and analogs

Dirichlet L-series and other number fields

The Riemann hypothesis can be generalized by replacing the Riemann zeta function by the formally similar, but much more general, global L-functions. In this broader setting, one expects the non-trivial zeros of the global L-functions to have real part 1/2. It is these conjectures, rather than the classical Riemann hypothesis only for the single Riemann zeta function, which account for the true importance of the Riemann hypothesis in mathematics.

The generalized Riemann hypothesis extends the Riemann hypothesis to all Dirichlet L-functions. In particular it implies the conjecture that Siegel zeros (zeros of L-functions between 1/2 and 1) do not exist.

The extended Riemann hypothesis extends the Riemann hypothesis to all Dedekind zeta functions of algebraic number fields. The extended Riemann hypothesis for abelian extension of the rationals is equivalent to the generalized Riemann hypothesis. The Riemann hypothesis can also be extended to the L-functions of Hecke characters of number fields.

The grand Riemann hypothesis extends it to all automorphic zeta functions, such as Mellin transforms of Hecke eigenforms.

Function fields and zeta functions of varieties over finite fields

Artin (1924) introduced global zeta functions of (quadratic) function fields and conjectured an analogue of the Riemann hypothesis for them, which has been proved by Hasse in the genus 1 case and by Weil (1948) in general. For instance, the fact that the Gauss sum, of the quadratic character of a finite field of size q (with q odd), has absolute value ${\displaystyle \sqrt{q}}$ is actually an instance of the Riemann hypothesis in the function field setting. This led Weil (1949) to conjecture a similar statement for all algebraic varieties; the resulting Weil conjectures were proved by Pierre Deligne (1974, 1980).

Arithmetic zeta functions of arithmetic schemes and their L-factors

Arithmetic zeta functions generalise the Riemann and Dedekind zeta functions as well as the zeta functions of varieties over finite fields to every arithmetic scheme or a scheme of finite type over integers. The arithmetic zeta function of a regular connected equidimensional arithmetic scheme of Kronecker dimension n can be factorized into the product of appropriately defined L-factors and an auxiliary factor Jean-Pierre Serre (1969–1970). Assuming a functional equation and meromorphic continuation, the generalized Riemann hypothesis for the L-factor states that its zeros inside the critical strip ${\displaystyle \mathrm{\Re }(s)\in (0,n)}$ lie on the central line. Correspondingly, the generalized Riemann hypothesis for the arithmetic zeta function of a regular connected equidimensional arithmetic scheme states that its zeros inside the critical strip lie on vertical lines ${\displaystyle \mathrm{\Re }(s)=1/2,3/2,\dots ,n-1/2}$ and its poles inside the critical strip lie on vertical lines ${\displaystyle \mathrm{\Re }(s)=1,2,\dots ,n-1}$. This is known for schemes in positive characteristic and follows from Pierre Deligne (1974, 1980), but remains entirely unknown in characteristic zero.

Selberg zeta functions

Main article: Selberg zeta function

Selberg (1956) introduced the Selberg zeta function of a Riemann surface. These are similar to the Riemann zeta function: they have a functional equation, and an infinite product similar to the Euler product but taken over closed geodesics rather than primes. The Selberg trace formula is the analogue for these functions of the explicit formulas in prime number theory. Selberg proved that the Selberg zeta functions satisfy the analogue of the Riemann hypothesis, with the imaginary parts of their zeros related to the eigenvalues of the Laplacian operator of the Riemann surface.

Ihara zeta functions

The Ihara zeta function of a finite graph is an analogue of the Selberg zeta function, which was first introduced by Yasutaka Ihara in the context of discrete subgroups of the two-by-two p-adic special linear group. A regular finite graph is a Ramanujan graph, a mathematical model of efficient communication networks, if and only if its Ihara zeta function satisfies the analogue of the Riemann hypothesis as was pointed out by T. Sunada.

Montgomery's pair correlation conjecture

Montgomery (1973) suggested the pair correlation conjecture that the correlation functions of the (suitably normalized) zeros of the zeta function should be the same as those of the eigenvalues of a random hermitian matrix. Odlyzko (1987) showed that this is supported by large-scale numerical calculations of these correlation functions.

Montgomery showed that (assuming the Riemann hypothesis) at least 2/3 of all zeros are simple, and a related conjecture is that all zeros of the zeta function are simple (or more generally have no non-trivial integer linear relations between their imaginary parts). Dedekind zeta functions of algebraic number fields, which generalize the Riemann zeta function, often do have multiple complex zeros. This is because the Dedekind zeta functions factorize as a product of powers of Artin L-functions, so zeros of Artin L-functions sometimes give rise to multiple zeros of Dedekind zeta functions. Other examples of zeta functions with multiple zeros are the L-functions of some elliptic curves: these can have multiple zeros at the real point of their critical line; the Birch-Swinnerton-Dyer conjecture predicts that the multiplicity of this zero is the rank of the elliptic curve.

Other zeta functions

There are many other examples of zeta functions with analogues of the Riemann hypothesis, some of which have been proved. Goss zeta functions of function fields have a Riemann hypothesis, proved by Sheats (1998). The main conjecture of Iwasawa theory, proved by Barry Mazur and Andrew Wiles for cyclotomic fields, and Wiles for totally real fields, identifies the zeros of a p-adic L-function with the eigenvalues of an operator, so can be thought of as an analogue of the Hilbert–Pólya conjecture for p-adic L-functions.

Attempted proofs

Several mathematicians have addressed the Riemann hypothesis, but none of their attempts has yet been accepted as a proof. Watkins (2021) lists some incorrect solutions.

Operator theory

Main article: Hilbert–Pólya conjecture

Hilbert and Pólya suggested that one way to derive the Riemann hypothesis would be to find a self-adjoint operator, from the existence of which the statement on the real parts of the zeros of ζ(s) would follow when one applies the criterion on real eigenvalues. Some support for this idea comes from several analogues of the Riemann zeta functions whose zeros correspond to eigenvalues of some operator: the zeros of a zeta function of a variety over a finite field correspond to eigenvalues of a Frobenius element on an étale cohomology group, the zeros of a Selberg zeta function are eigenvalues of a Laplacian operator of a Riemann surface, and the zeros of a p-adic zeta function correspond to eigenvectors of a Galois action on ideal class groups.

Odlyzko (1987) showed that the distribution of the zeros of the Riemann zeta function shares some statistical properties with the eigenvalues of random matrices drawn from the Gaussian unitary ensemble. This gives some support to the Hilbert–Pólya conjecture.

In 1999, Michael Berry and Jonathan Keating conjectured that there is some unknown quantization ${\displaystyle \hat{H}}$ of the classical Hamiltonian H = xp so that

$${\displaystyle \zeta (1/2+i\hat{H})=0}$$

and even more strongly, that the Riemann zeros coincide with the spectrum of the operator ${\displaystyle 1/2+i\hat{H}}$. This is in contrast to canonical quantization, which leads to the Heisenberg uncertainty principle ${\displaystyle {\sigma }_x{\sigma }_p\ge \frac{\hslash }{2}}$ and the natural numbers as spectrum of the quantum harmonic oscillator. The crucial point is that the Hamiltonian should be a self-adjoint operator so that the quantization would be a realization of the Hilbert–Pólya program. In a connection with this quantum mechanical problem Berry and Connes had proposed that the inverse of the potential of the Hamiltonian is connected to the half-derivative of the function

$${\displaystyle N(s)=\frac{1}{\pi }\mathrm{Arg}\xi (1/2+i\sqrt{s})}$$

then, in Berry–Connes approach

$${\displaystyle V^{-1}(x)=\sqrt{4\pi }\frac{d^{1/2}N(x)}{d{x}^{1/2}}.}$$

This yields a Hamiltonian whose eigenvalues are the square of the imaginary part of the Riemann zeros, and also that the functional determinant of this Hamiltonian operator is just the Riemann Xi function. In fact the Riemann Xi function would be proportional to the functional determinant (Hadamard product)

$${\displaystyle det(H+1/4+s(s-1))}$$

as proved by Connes and others, in this approach

$${\displaystyle \frac{\xi (s)}{\xi (0)}=\frac{det(H+s(s-1)+1/4)}{det(H+1/4)}.}$$

The analogy with the Riemann hypothesis over finite fields suggests that the Hilbert space containing eigenvectors corresponding to the zeros might be some sort of first cohomology group of the spectrum Spec (Z) of the integers. Deninger (1998) described some of the attempts to find such a cohomology theory.

Zagier (1981) constructed a natural space of invariant functions on the upper half plane that has eigenvalues under the Laplacian operator that correspond to zeros of the Riemann zeta function—and remarked that in the unlikely event that one could show the existence of a suitable positive definite inner product on this space, the Riemann hypothesis would follow. Cartier (1982) discussed a related example, where due to a bizarre bug a computer program listed zeros of the Riemann zeta function as eigenvalues of the same Laplacian operator.

Schumayer & Hutchinson (2011) surveyed some of the attempts to construct a suitable physical model related to the Riemann zeta function.

Lee–Yang theorem

The Lee–Yang theorem states that the zeros of certain partition functions in statistical mechanics all lie on a "critical line" with their real part equals to 0, and this has led to some speculation about a relationship with the Riemann hypothesis.

Turán's result

Pál Turán (1948) showed that if the functions

$${\displaystyle \sum _{n=1}^N{n}^{-s}}$$

have no zeros when the real part of s is greater than one then

$${\displaystyle T(x)=\sum _{n\le x}\frac{\lambda (n)}{n}\ge 0\text{ for }x>0,}$$

where λ(n) is the Liouville function given by (−1)^r if n has r prime factors. He showed that this in turn would imply that the Riemann hypothesis is true. But Haselgrove (1958) proved that T(x) is negative for infinitely many x (and also disproved the closely related Pólya conjecture), and Borwein, Ferguson & Mossinghoff (2008) showed that the smallest such x is 72185376951205. Spira (1968) showed by numerical calculation that the finite Dirichlet series above for N=19 has a zero with real part greater than 1. Turán also showed that a somewhat weaker assumption, the nonexistence of zeros with real part greater than 1+N^−1/2+ε for large N in the finite Dirichlet series above, would also imply the Riemann hypothesis, but Montgomery (1983) showed that for all sufficiently large N these series have zeros with real part greater than 1 + (log log N)/(4 log N). Therefore, Turán's result is vacuously true and cannot help prove the Riemann hypothesis.

Noncommutative geometry

Connes (1999, 2000) has described a relationship between the Riemann hypothesis and noncommutative geometry, and showed that a suitable analog of the Selberg trace formula for the action of the idèle class group on the adèle class space would imply the Riemann hypothesis. Some of these ideas are elaborated in Lapidus (2008).

Hilbert spaces of entire functions

Louis de Branges (1992) showed that the Riemann hypothesis would follow from a positivity condition on a certain Hilbert space of entire functions. However Conrey & Li (2000) showed that the necessary positivity conditions are not satisfied. Despite this obstacle, de Branges has continued to work on an attempted proof of the Riemann hypothesis along the same lines, but this has not been widely accepted by other mathematicians.

Quasicrystals

The Riemann hypothesis implies that the zeros of the zeta function form a quasicrystal, a distribution with discrete support whose Fourier transform also has discrete support. Dyson (2009) suggested trying to prove the Riemann hypothesis by classifying, or at least studying, 1-dimensional quasicrystals.

Arithmetic zeta functions of models of elliptic curves over number fields

When one goes from geometric dimension one, e.g. an algebraic number field, to geometric dimension two, e.g. a regular model of an elliptic curve over a number field, the two-dimensional part of the generalized Riemann hypothesis for the arithmetic zeta function of the model deals with the poles of the zeta function. In dimension one the study of the zeta integral in Tate's thesis does not lead to new important information on the Riemann hypothesis. Contrary to this, in dimension two work of Ivan Fesenko on two-dimensional generalisation of Tate's thesis includes an integral representation of a zeta integral closely related to the zeta function. In this new situation, not possible in dimension one, the poles of the zeta function can be studied via the zeta integral and associated adele groups. Related conjecture of Fesenko (2010) on the positivity of the fourth derivative of a boundary function associated to the zeta integral essentially implies the pole part of the generalized Riemann hypothesis. Suzuki (2011) proved that the latter, together with some technical assumptions, implies Fesenko's conjecture.

Multiple zeta functions

Deligne's proof of the Riemann hypothesis over finite fields used the zeta functions of product varieties, whose zeros and poles correspond to sums of zeros and poles of the original zeta function, in order to bound the real parts of the zeros of the original zeta function. By analogy, Kurokawa (1992) introduced multiple zeta functions whose zeros and poles correspond to sums of zeros and poles of the Riemann zeta function. To make the series converge he restricted to sums of zeros or poles all with non-negative imaginary part. So far, the known bounds on the zeros and poles of the multiple zeta functions are not strong enough to give useful estimates for the zeros of the Riemann zeta function.

Location of the zeros

Number of zeros

The functional equation combined with the argument principle implies that the number of zeros of the zeta function with imaginary part between 0 and T is given by

${\displaystyle N(T)=\frac{1}{\pi }\mathrm{A}\mathrm{r}\mathrm{g}(\xi (s))=\frac{1}{\pi }\mathrm{A}\mathrm{r}\mathrm{g}(\mathrm{\Gamma }(\frac{s}{2}){\pi }^{-\frac{s}{2}}\zeta (s)s(s-1)/2)}$

for s=1/2+iT, where the argument is defined by varying it continuously along the line with Im(s)=T, starting with argument 0 at ∞+iT. This is the sum of a large but well understood term

${\displaystyle \frac{1}{\pi }\mathrm{A}\mathrm{r}\mathrm{g}(\mathrm{\Gamma }(\frac{s}{2}){\pi }^{-s/2}s(s-1)/2)=\frac{T}{2\pi }\log \frac{T}{2\pi }-\frac{T}{2\pi }+7/8+O(1/T)}$

and a small but rather mysterious term

${\displaystyle S(T)=\frac{1}{\pi }\mathrm{A}\mathrm{r}\mathrm{g}(\zeta (1/2+iT))=O(\log T).}$

So the density of zeros with imaginary part near T is about log(T)/2π, and the function S describes the small deviations from this. The function S(t) jumps by 1 at each zero of the zeta function, and for t ≥ 8 it decreases monotonically between zeros with derivative close to −log t.

Trudgian (2014) proved that, if ${\displaystyle T>e}$, then

${\displaystyle |N(T)-\frac{T}{2\pi }\log \frac{T}{2\pi e}|\le 0.112\log T+0.278\log \log T+3.385+\frac{0.2}{T}}$.

Karatsuba (1996) proved that every interval (T, T+H] for ${\displaystyle H\ge T^{\frac{27}{82}+\epsilon }}$ contains at least

${\displaystyle H(\log T{)}^{\frac{1}{3}}{e}^{-c\sqrt{\log \log T}}}$

points where the function S(t) changes sign.

Selberg (1946) showed that the average moments of even powers of S are given by

${\displaystyle {\int }_0^T|S(t){|}^{2k}dt=\frac{(2k)!}{k!(2\pi {)}^{2k}}T(\log \log T{)}^k+O(T(\log \log T{)}^{k-1/2}).}$

This suggests that S(T)/(log log T)^1/2 resembles a Gaussian random variable with mean 0 and variance 2π² (Ghosh (1983) proved this fact). In particular |S(T)| is usually somewhere around (log log T)^1/2, but occasionally much larger. The exact order of growth of S(T) is not known. There has been no unconditional improvement to Riemann's original bound S(T)=O(log T), though the Riemann hypothesis implies the slightly smaller bound S(T)=O(log T/log log T). The true order of magnitude may be somewhat less than this, as random functions with the same distribution as S(T) tend to have growth of order about log(T)^1/2. In the other direction it cannot be too small: Selberg (1946) showed that S(T) ≠ o((log T)^1/3/(log log T)^7/3), and assuming the Riemann hypothesis Montgomery showed that S(T) ≠ o((log T)^1/2/(log log T)^1/2).

Numerical calculations confirm that S grows very slowly: |S(T)| < 1 for T < 280, |S(T)| < 2 for T < 6800000, and the largest value of |S(T)| found so far is not much larger than 3.

Riemann's estimate S(T) = O(log T) implies that the gaps between zeros are bounded, and Littlewood improved this slightly, showing that the gaps between their imaginary parts tend to 0.

Theorem of Hadamard and de la Vallée-Poussin

Hadamard (1896) and de la Vallée-Poussin (1896) independently proved that no zeros could lie on the line Re(s) = 1. Together with the functional equation and the fact that there are no zeros with real part greater than 1, this showed that all non-trivial zeros must lie in the interior of the critical strip 0 < Re(s) < 1. This was a key step in their first proofs of the prime number theorem.

Both the original proofs that the zeta function has no zeros with real part 1 are similar, and depend on showing that if ζ(1+it) vanishes, then ζ(1+2it) is singular, which is not possible. One way of doing this is by using the inequality

${\displaystyle |\zeta (\sigma {)}^3\zeta (\sigma +it{)}^4\zeta (\sigma +2it)|\ge 1}$

for σ > 1, t real, and looking at the limit as σ → 1. This inequality follows by taking the real part of the log of the Euler product to see that

${\displaystyle |\zeta (\sigma +it)|=\exp \mathrm{\Re }\sum _{p^n}\frac{p^{-n(\sigma +it)}}{n}=\exp \sum _{p^n}\frac{p^{-n\sigma }\cos (t\log p^n)}{n},}$

where the sum is over all prime powers pⁿ, so that

${\displaystyle |\zeta (\sigma {)}^3\zeta (\sigma +it{)}^4\zeta (\sigma +2it)|=\exp \sum _{p^n}{p}^{-n\sigma }\frac{3+4\cos (t\log p^n)+\cos (2t\log p^n)}{n}}$

which is at least 1 because all the terms in the sum are positive, due to the inequality

${\displaystyle 3+4\cos (\theta )+\cos (2\theta )=2(1+\cos (\theta ){)}^2\ge 0.}$

Zero-free regions

The most extensive computer search by Platt and Trudgian for counter examples of the Riemann hypothesis has verified it for ${\displaystyle |t|\le 3.0001753328\cdot {10}^{12}}$. Beyond that zero-free regions are known as inequalities concerning σ + i t, which can be zeroes. The oldest version is from De la Vallée-Poussin (1899–1900), who proved there is a region without zeroes that satisfies 1 − σ ≥ C/log(t) for some positive constant C. In other words, zeros cannot be too close to the line σ = 1: there is a zero-free region close to this line. This has been enlarged by several authors using methods such as Vinogradov's mean-value theorem.

The most recent paper by Mossinghoff, Trudgian and Yang is from December 2022 and provides four zero-free regions that improved the previous results of Kevin Ford from 2002, Mossinghoff and Trudgian themselves from 2015 and Pace Nielsen's slight improvement of Ford from October 2022:

${\displaystyle \sigma \ge 1-\frac{1}{5.558691\log |t|}}$ whenever ${\displaystyle |t|\ge 2}$,

${\displaystyle \sigma \ge 1-\frac{1}{55.241(\log |t|{)}^{2/3}(\log \log |t|{)}^{1/3}}}$ whenever ${\displaystyle |t|\ge 3}$ (largest known region in the bound ${\displaystyle 3.0001753328\cdot {10}^{12}\le |t|\le \exp (64.1)\approx 6.89\cdot {10}^{27}}$),

${\displaystyle \sigma \ge 1-\frac{0.04962-\frac{0.0196}{1.15+\log 3+\frac{1}{6}\log t+\log \log t}}{0.685+\log 3+\frac{1}{6}\log t+1.155\cdot \log \log t}}$ whenever ${\displaystyle |t|\ge 1.88\cdot {10}^{14}}$ (largest known region in the bound ${\displaystyle \exp (64.1)\le |t|\le \exp (1000)\approx 1.97\cdot {10}^{434}}$) and

${\displaystyle \sigma \ge 1-\frac{0.05035}{\frac{27}{164}(\log |t|)+7.096}+\frac{0.0349}{(\frac{27}{164}(\log |t|)+7.096{)}^2}}$ whenever ${\displaystyle |t|\ge \exp (1000)}$ (largest known region in its own bound)

The paper also has a improvement to the second zero-free region, whose bounds are unknown on account of ${\displaystyle |t|}$ being merely assumed to be "sufficiently large" to fulfill the requirements of the paper's proof. This region is

${\displaystyle \sigma \ge 1-\frac{1}{48.1588(\log |t|{)}^{2/3}(\log \log |t|{)}^{1/3}}}$.

Zeros on the critical line

Hardy (1914) and Hardy & Littlewood (1921) showed there are infinitely many zeros on the critical line, by considering moments of certain functions related to the zeta function. Selberg (1942) proved that at least a (small) positive proportion of zeros lie on the line. Levinson (1974) improved this to one-third of the zeros by relating the zeros of the zeta function to those of its derivative, and Conrey (1989) improved this further to two-fifths. In 2020, this estimate was extended to five-twelfths by Pratt, Robles, Zaharescu and Zeindler by considering extended mollifiers that can accommodate higher order derivatives of the zeta function and their associated Kloosterman sums.

Most zeros lie close to the critical line. More precisely, Bohr & Landau (1914) showed that for any positive ε, the number of zeroes with real part at least 1/2+ε and imaginary part at between -T and T is ${\displaystyle O(T)}$. Combined with the facts that zeroes on the critical strip are symmetric about the critical line and that the total number of zeroes in the critical strip is ${\displaystyle \mathrm{\Theta }(T\log T)}$, almost all non-trivial zeroes are within a distance ε of the critical line. Ivić (1985) gives several more precise versions of this result, called zero density estimates, which bound the number of zeros in regions with imaginary part at most T and real part at least 1/2+ε.

Hardy–Littlewood conjectures

In 1914 Godfrey Harold Hardy proved that ${\displaystyle \zeta \left(\frac{1}{2}+it\right)}$ has infinitely many real zeros.

The next two conjectures of Hardy and John Edensor Littlewood on the distance between real zeros of ${\displaystyle \zeta \left(\frac{1}{2}+it\right)}$ and on the density of zeros of ${\displaystyle \zeta \left(\frac{1}{2}+it\right)}$ on the interval ${\displaystyle (T,T+H]}$ for sufficiently large ${\displaystyle T>0}$, and ${\displaystyle H=T^{a+\epsilon }}$ and with as small as possible value of ${\displaystyle a>0}$, where ${\displaystyle \epsilon >0}$ is an arbitrarily small number, open two new directions in the investigation of the Riemann zeta function:

1. For any ${\displaystyle \epsilon >0}$ there exists a lower bound ${\displaystyle T_0=T_0(\epsilon )>0}$ such that for ${\displaystyle T\ge T_0}$ and ${\displaystyle H=T^{\frac{1}{4}+\epsilon }}$ the interval ${\displaystyle (T,T+H]}$ contains a zero of odd order of the function ${\displaystyle \zeta (\frac{1}{2}+it)}$.

Let ${\displaystyle N(T)}$ be the total number of real zeros, and ${\displaystyle N_0(T)}$ be the total number of zeros of odd order of the function ${\displaystyle \zeta \left(\frac{1}{2}+it\right)}$ lying on the interval ${\displaystyle (0,T]}$.

2. For any ${\displaystyle \epsilon >0}$ there exists ${\displaystyle T_0=T_0(\epsilon )>0}$ and some ${\displaystyle c=c(\epsilon )>0}$, such that for ${\displaystyle T\ge T_0}$ and ${\displaystyle H=T^{\frac{1}{2}+\epsilon }}$ the inequality ${\displaystyle N_0(T+H)-N_0(T)\ge cH}$ is true.

Selberg's zeta function conjecture

Main article: Selberg's zeta function conjecture

Atle Selberg (1942) investigated the problem of Hardy–Littlewood 2 and proved that for any ε > 0 there exists such ${\displaystyle T_0=T_0(\epsilon )>0}$ and c = c(ε) > 0, such that for ${\displaystyle T\ge T_0}$ and ${\displaystyle H=T^{0.5+\epsilon }}$ the inequality ${\displaystyle N(T+H)-N(T)\ge cH\log T}$ is true. Selberg conjectured that this could be tightened to ${\displaystyle H=T^{0.5}}$. A. A. Karatsuba (1984a, 1984b, 1985) proved that for a fixed ε satisfying the condition 0 < ε < 0.001, a sufficiently large T and ${\displaystyle H=T^{a+\epsilon }}$, ${\displaystyle a=\frac{27}{82}=\frac{1}{3}-\frac{1}{246}}$, the interval (T, T+H) contains at least cH log(T) real zeros of the Riemann zeta function ${\displaystyle \zeta \left(\frac{1}{2}+it\right)}$ and therefore confirmed the Selberg conjecture. The estimates of Selberg and Karatsuba can not be improved in respect of the order of growth as T → ∞.

Karatsuba (1992) proved that an analog of the Selberg conjecture holds for almost all intervals (T, T+H], ${\displaystyle H=T^{\epsilon }}$, where ε is an arbitrarily small fixed positive number. The Karatsuba method permits to investigate zeros of the Riemann zeta function on "supershort" intervals of the critical line, that is, on the intervals (T, T+H], the length H of which grows slower than any, even arbitrarily small degree T. In particular, he proved that for any given numbers ε, ${\displaystyle {\epsilon }_1}$ satisfying the conditions ${\displaystyle 0<\epsilon ,{\epsilon }_1<1}$ almost all intervals (T, T+H] for ${\displaystyle H\ge \exp \{(\log T{)}^{\epsilon }\}}$ contain at least ${\displaystyle H(\log T{)}^{1-{\epsilon }_1}}$ zeros of the function ${\displaystyle \zeta \left(\frac{1}{2}+it\right)}$. This estimate is quite close to the one that follows from the Riemann hypothesis.

Numerical calculations

The function

${\displaystyle {\pi }^{-\frac{s}{2}}\mathrm{\Gamma }(\frac{s}{2})\zeta (s)}$

has the same zeros as the zeta function in the critical strip, and is real on the critical line because of the functional equation, so one can prove the existence of zeros exactly on the real line between two points by checking numerically that the function has opposite signs at these points. Usually one writes

${\displaystyle \zeta (\frac{1}{2}+it)=Z(t)e^{-i\theta (t)}}$

where Hardy's Z function and the Riemann–Siegel theta function θ are uniquely defined by this and the condition that they are smooth real functions with θ(0)=0. By finding many intervals where the function Z changes sign one can show that there are many zeros on the critical line. To verify the Riemann hypothesis up to a given imaginary part T of the zeros, one also has to check that there are no further zeros off the line in this region. This can be done by calculating the total number of zeros in the region using Turing's method and checking that it is the same as the number of zeros found on the line. This allows one to verify the Riemann hypothesis computationally up to any desired value of T (provided all the zeros of the zeta function in this region are simple and on the critical line).

These calculations can also be used to estimate ${\displaystyle \pi (x)}$ for finite ranges of ${\displaystyle x}$. For example, using the latest result from 2020 (zeros up to height ${\displaystyle 3\times {10}^{12}}$), it has been shown that

${\displaystyle |\pi (x)-\mathrm{li}(x)|<\frac{1}{8\pi }\sqrt{x}\log (x),\text{for }2657\le x\le 1.101\times {10}^{26}.}$

In general, this inequality holds if

${\displaystyle x\ge 2657}$ and ${\displaystyle \frac{9.06}{\log \log x}\sqrt{\frac{x}{\log x}}\le T,}$

where ${\displaystyle T}$ is the largest known value such that the Riemann hypothesis is true for all zeros ${\displaystyle \rho }$ with ${\displaystyle \mathrm{\Im }\left(\rho \right)\in \left(0,T\right]}$.

Some calculations of zeros of the zeta function are listed below, where the "height" of a zero is the magnitude of its imaginary part, and the height of the nth zero is denoted by γ_n. So far all zeros that have been checked are on the critical line and are simple. (A multiple zero would cause problems for the zero finding algorithms, which depend on finding sign changes between zeros.) For tables of the zeros, see Haselgrove & Miller (1960) or Odlyzko.

Year	Number of zeros	Author
1859?	3	B. Riemann used the Riemann–Siegel formula (unpublished, but reported in Siegel 1932).
1903	15	J. P. Gram (1903) used Euler–Maclaurin formula and discovered Gram's law. He showed that all 10 zeros with imaginary part at most 50 range lie on the critical line with real part 1/2 by computing the sum of the inverse 10th powers of the roots he found.
1914	79 (γn ≤ 200)	R. J. Backlund (1914) introduced a better method of checking all the zeros up to that point are on the line, by studying the argument S(T) of the zeta function.
1925	138 (γn ≤ 300)	J. I. Hutchinson (1925) found the first failure of Gram's law, at the Gram point g126.
1935	195	E. C. Titchmarsh (1935) used the recently rediscovered Riemann–Siegel formula, which is much faster than Euler–Maclaurin summation. It takes about O(T3/2+ε) steps to check zeros with imaginary part less than T, while the Euler–Maclaurin method takes about O(T2+ε) steps.
1936	1041	E. C. Titchmarsh (1936) and L. J. Comrie were the last to find zeros by hand.
1953	1104	A. M. Turing (1953) found a more efficient way to check that all zeros up to some point are accounted for by the zeros on the line, by checking that Z has the correct sign at several consecutive Gram points and using the fact that S(T) has average value 0. This requires almost no extra work because the sign of Z at Gram points is already known from finding the zeros, and is still the usual method used. This was the first use of a digital computer to calculate the zeros.
1956	15000	D. H. Lehmer (1956) discovered a few cases where the zeta function has zeros that are "only just" on the line: two zeros of the zeta function are so close together that it is unusually difficult to find a sign change between them. This is called "Lehmer's phenomenon", and first occurs at the zeros with imaginary parts 7005.063 and 7005.101, which differ by only .04 while the average gap between other zeros near this point is about 1.
1956	25000	D. H. Lehmer
1958	35337	N. A. Meller
1966	250000	R. S. Lehman
1968	3500000	Rosser, Yohe & Schoenfeld (1969) stated Rosser's rule (described below).
1977	40000000	R. P. Brent
1979	81000001	R. P. Brent
1982	200000001	R. P. Brent, J. van de Lune, H. J. J. te Riele, D. T. Winter
1983	300000001	J. van de Lune, H. J. J. te Riele
1986	1500000001	van de Lune, te Riele & Winter (1986) gave some statistical data about the zeros and give several graphs of Z at places where it has unusual behavior.
1987	A few of large (~1012) height	A. M. Odlyzko (1987) computed smaller numbers of zeros of much larger height, around 1012, to high precision to check Montgomery's pair correlation conjecture.
1992	A few of large (~1020) height	A. M. Odlyzko (1992) computed a 175 million zeros of heights around 1020 and a few more of heights around 2×1020, and gave an extensive discussion of the results.
1998	10000 of large (~1021) height	A. M. Odlyzko (1998) computed some zeros of height about 1021
2001	10000000000	J. van de Lune (unpublished)
2004	~900000000000	S. Wedeniwski (ZetaGrid distributed computing)
2004	10000000000000 and a few of large (up to ~1024) heights	X. Gourdon (2004) and Patrick Demichel used the Odlyzko–Schönhage algorithm. They also checked two billion zeros around heights 1013, 1014, ..., 1024.
2020	12363153437138 up to height 3000175332800	Platt & Trudgian (2021). They also verified the work of Gourdon (2004) and others.

Gram points

A Gram point is a point on the critical line 1/2 + it where the zeta function is real and non-zero. Using the expression for the zeta function on the critical line, ζ(1/2 + it) = Z(t)e ^− iθ(t), where Hardy's function, Z, is real for real t, and θ is the Riemann–Siegel theta function, we see that zeta is real when sin(θ(t)) = 0. This implies that θ(t) is an integer multiple of π, which allows for the location of Gram points to be calculated fairly easily by inverting the formula for θ. They are usually numbered as g_n for n = 0, 1, ..., where g_n is the unique solution of θ(t) = nπ.

Gram observed that there was often exactly one zero of the zeta function between any two Gram points; Hutchinson called this observation Gram's law. There are several other closely related statements that are also sometimes called Gram's law: for example, (−1)ⁿZ(g_n) is usually positive, or Z(t) usually has opposite sign at consecutive Gram points. The imaginary parts γ_n of the first few zeros (in blue) and the first few Gram points g_n are given in the following table

		g−1	γ1	g0	γ2	g1	γ3	g2	γ4	g3	γ5	g4	γ6	g5
0	3.436	9.667	14.135	17.846	21.022	23.170	25.011	27.670	30.425	31.718	32.935	35.467	37.586	38.999

This is a polar plot of the first 20 non-trivial Riemann zeta function zeros (including Gram points) along the critical line ${\displaystyle \zeta (1/2+it)}$ for real values of ${\displaystyle t}$ running from 0 to 50. The consecutively labeled zeros have 50 red plot points between each, with zeros identified by concentric magenta rings scaled to show the relative distance between their values of t. Gram's law states that the curve usually crosses the real axis once between zeros.

The first failure of Gram's law occurs at the 127th zero and the Gram point g₁₂₆, which are in the "wrong" order.

g124	γ126	g125	g126	γ127	γ128	g127	γ129	g128
279.148	279.229	280.802	282.455	282.465	283.211	284.104	284.836	285.752

A Gram point t is called good if the zeta function is positive at 1/2 + it. The indices of the "bad" Gram points where Z has the "wrong" sign are 126, 134, 195, 211, ... (sequence A114856 in the OEIS). A Gram block is an interval bounded by two good Gram points such that all the Gram points between them are bad. A refinement of Gram's law called Rosser's rule due to Rosser, Yohe & Schoenfeld (1969) says that Gram blocks often have the expected number of zeros in them (the same as the number of Gram intervals), even though some of the individual Gram intervals in the block may not have exactly one zero in them. For example, the interval bounded by g₁₂₅ and g₁₂₇ is a Gram block containing a unique bad Gram point g₁₂₆, and contains the expected number 2 of zeros although neither of its two Gram intervals contains a unique zero. Rosser et al. checked that there were no exceptions to Rosser's rule in the first 3 million zeros, although there are infinitely many exceptions to Rosser's rule over the entire zeta function.

Gram's rule and Rosser's rule both say that in some sense zeros do not stray too far from their expected positions. The distance of a zero from its expected position is controlled by the function S defined above, which grows extremely slowly: its average value is of the order of (log log T)^1/2, which only reaches 2 for T around 10²⁴. This means that both rules hold most of the time for small T but eventually break down often. Indeed, Trudgian (2011) showed that both Gram's law and Rosser's rule fail in a positive proportion of cases. To be specific, it is expected that in about 66% one zero is enclosed by two successive Gram points, but in 17% no zero and in 17% two zeros are in such a Gram-interval on the long run Hanga (2020).

Arguments for and against the Riemann hypothesis

Mathematical papers about the Riemann hypothesis tend to be cautiously noncommittal about its truth. Of authors who express an opinion, most of them, such as Riemann (1859) and Bombieri (2000), imply that they expect (or at least hope) that it is true. The few authors who express serious doubt about it include Ivić (2008), who lists some reasons for skepticism, and Littlewood (1962), who flatly states that he believes it false, that there is no evidence for it and no imaginable reason it would be true. The consensus of the survey articles (Bombieri 2000, Conrey 2003, and Sarnak 2005) is that the evidence for it is strong but not overwhelming, so that while it is probably true there is reasonable doubt.

Some of the arguments for and against the Riemann hypothesis are listed by Sarnak (2005), Conrey (2003), and Ivić (2008), and include the following:

• Several analogues of the Riemann hypothesis have already been proved. The proof of the Riemann hypothesis for varieties over finite fields by Deligne (1974) is possibly the single strongest theoretical reason in favor of the Riemann hypothesis. This provides some evidence for the more general conjecture that all zeta functions associated with automorphic forms satisfy a Riemann hypothesis, which includes the classical Riemann hypothesis as a special case. Similarly Selberg zeta functions satisfy the analogue of the Riemann hypothesis, and are in some ways similar to the Riemann zeta function, having a functional equation and an infinite product expansion analogous to the Euler product expansion. But there are also some major differences; for example, they are not given by Dirichlet series. The Riemann hypothesis for the Goss zeta function was proved by Sheats (1998). In contrast to these positive examples, some Epstein zeta functions do not satisfy the Riemann hypothesis even though they have an infinite number of zeros on the critical line. These functions are quite similar to the Riemann zeta function, and have a Dirichlet series expansion and a functional equation, but the ones known to fail the Riemann hypothesis do not have an Euler product and are not directly related to automorphic representations.
• At first, the numerical verification that many zeros lie on the line seems strong evidence for it. But analytic number theory has had many conjectures supported by substantial numerical evidence that turned out to be false. See Skewes number for a notorious example, where the first exception to a plausible conjecture related to the Riemann hypothesis probably occurs around 10³¹⁶; a counterexample to the Riemann hypothesis with imaginary part this size would be far beyond anything that can currently be computed using a direct approach. The problem is that the behavior is often influenced by very slowly increasing functions such as log log T, that tend to infinity, but do so so slowly that this cannot be detected by computation. Such functions occur in the theory of the zeta function controlling the behavior of its zeros; for example the function S(T) above has average size around (log log T)^1/2. As S(T) jumps by at least 2 at any counterexample to the Riemann hypothesis, one might expect any counterexamples to the Riemann hypothesis to start appearing only when S(T) becomes large. It is never much more than 3 as far as it has been calculated, but is known to be unbounded, suggesting that calculations may not have yet reached the region of typical behavior of the zeta function.
• Denjoy's probabilistic argument for the Riemann hypothesis is based on the observation that if μ(x) is a random sequence of "1"s and "−1"s then, for every ε > 0, the partial sums
  $${\displaystyle M(x)=\sum _{n\le x}\mu (n)}$$

  
  (the values of which are positions in a simple random walk) satisfy the bound
  $${\displaystyle M(x)=O(x^{1/2+\epsilon })}$$

  
  with probability 1. The Riemann hypothesis is equivalent to this bound for the Möbius function μ and the Mertens function M derived in the same way from it. In other words, the Riemann hypothesis is in some sense equivalent to saying that μ(x) behaves like a random sequence of coin tosses. When μ(x) is nonzero its sign gives the parity of the number of prime factors of x, so informally the Riemann hypothesis says that the parity of the number of prime factors of an integer behaves randomly. Such probabilistic arguments in number theory often give the right answer, but tend to be very hard to make rigorous, and occasionally give the wrong answer for some results, such as Maier's theorem.
• The calculations in Odlyzko (1987) show that the zeros of the zeta function behave very much like the eigenvalues of a random Hermitian matrix, suggesting that they are the eigenvalues of some self-adjoint operator, which would imply the Riemann hypothesis. All attempts to find such an operator have failed.
• There are several theorems, such as Goldbach's weak conjecture for sufficiently large odd numbers, that were first proved using the generalized Riemann hypothesis, and later shown to be true unconditionally. This could be considered as weak evidence for the generalized Riemann hypothesis, as several of its "predictions" are true.
• Lehmer's phenomenon, where two zeros are sometimes very close, is sometimes given as a reason to disbelieve the Riemann hypothesis. But one would expect this to happen occasionally by chance even if the Riemann hypothesis is true, and Odlyzko's calculations suggest that nearby pairs of zeros occur just as often as predicted by Montgomery's conjecture.
• Patterson suggests that the most compelling reason for the Riemann hypothesis for most mathematicians is the hope that primes are distributed as regularly as possible.

Astronomical interferometer

From Wikipedia, the free encyclopedia
https://en.wikipedia.org/wiki/Astronomical_interferometer

An astronomical interferometer or telescope array is a set of separate telescopes, mirror segments, or radio telescope antennas that work together as a single telescope to provide higher resolution images of astronomical objects such as stars, nebulas and galaxies by means of interferometry. The advantage of this technique is that it can theoretically produce images with the angular resolution of a huge telescope with an aperture equal to the separation, called baseline, between the component telescopes. The main drawback is that it does not collect as much light as the complete instrument's mirror. Thus it is mainly useful for fine resolution of more luminous astronomical objects, such as close binary stars. Another drawback is that the maximum angular size of a detectable emission source is limited by the minimum gap between detectors in the collector array.

Interferometry is most widely used in radio astronomy, in which signals from separate radio telescopes are combined. A mathematical signal processing technique called aperture synthesis is used to combine the separate signals to create high-resolution images. In Very Long Baseline Interferometry (VLBI) radio telescopes separated by thousands of kilometers are combined to form a radio interferometer with a resolution which would be given by a hypothetical single dish with an aperture thousands of kilometers in diameter. At the shorter wavelengths used in infrared astronomy and optical astronomy it is more difficult to combine the light from separate telescopes, because the light must be kept coherent within a fraction of a wavelength over long optical paths, requiring very precise optics. Practical infrared and optical astronomical interferometers have only recently been developed, and are at the cutting edge of astronomical research. At optical wavelengths, aperture synthesis allows the atmospheric seeing resolution limit to be overcome, allowing the angular resolution to reach the diffraction limit of the optics.

ESO's VLT interferometer took the first detailed image of a disc around a young star.

Astronomical interferometers can produce higher resolution astronomical images than any other type of telescope. At radio wavelengths, image resolutions of a few micro-arcseconds have been obtained, and image resolutions of a fractional milliarcsecond have been achieved at visible and infrared wavelengths.

One simple layout of an astronomical interferometer is a parabolic arrangement of mirror pieces, giving a partially complete reflecting telescope but with a "sparse" or "dilute" aperture. In fact the parabolic arrangement of the mirrors is not important, as long as the optical path lengths from the astronomical object to the beam combiner (focus) are the same as would be given by the complete mirror case. Instead, most existing arrays use a planar geometry, and Labeyrie's hypertelescope will use a spherical geometry.

History

A 20-foot Michelson interferometer mounted on the frame of the 100-inch Hooker Telescope, 1920.

One of the first uses of optical interferometry was applied by the Michelson stellar interferometer on the Mount Wilson Observatory's reflector telescope to measure the diameters of stars. The red giant star Betelgeuse was the first to have its diameter determined in this way on December 13, 1920. In the 1940s radio interferometry was used to perform the first high resolution radio astronomy observations. For the next three decades astronomical interferometry research was dominated by research at radio wavelengths, leading to the development of large instruments such as the Very Large Array and the Atacama Large Millimeter Array.

Optical/infrared interferometry was extended to measurements using separated telescopes by Johnson, Betz and Townes (1974) in the infrared and by Labeyrie (1975) in the visible. In the late 1970s improvements in computer processing allowed for the first "fringe-tracking" interferometer, which operates fast enough to follow the blurring effects of astronomical seeing, leading to the Mk I, II and III series of interferometers. Similar techniques have now been applied at other astronomical telescope arrays, including the Keck Interferometer and the Palomar Testbed Interferometer.

Aerial view of the ESO/NAOJ/NRAO ALMA construction site.

In the 1980s the aperture synthesis interferometric imaging technique was extended to visible light and infrared astronomy by the Cavendish Astrophysics Group, providing the first very high resolution images of nearby stars. In 1995 this technique was demonstrated on an array of separate optical telescopes for the first time, allowing a further improvement in resolution, and allowing even higher resolution imaging of stellar surfaces. Software packages such as BSMEM or MIRA are used to convert the measured visibility amplitudes and closure phases into astronomical images. The same techniques have now been applied at a number of other astronomical telescope arrays, including the Navy Precision Optical Interferometer, the Infrared Spatial Interferometer and the IOTA array. A number of other interferometers have made closure phase measurements and are expected to produce their first images soon, including the VLTI, the CHARA array and Le Coroller and Dejonghe's Hypertelescope prototype. If completed, the MRO Interferometer with up to ten movable telescopes will produce among the first higher fidelity images from a long baseline interferometer. The Navy Optical Interferometer took the first step in this direction in 1996, achieving 3-way synthesis of an image of Mizar; then a first-ever six-way synthesis of Eta Virginis in 2002; and most recently "closure phase" as a step to the first synthesized images produced by geostationary satellites.

Modern astronomical interferometry

Further information: Astronomical optical interferometry

Astronomical interferometry is principally conducted using Michelson (and sometimes other type) interferometers. The principal operational interferometric observatories which use this type of instrumentation include VLTI, NPOI, and CHARA.

The Navy Precision Optical Interferometer (NPOI), a 437 ma baselined optical/near-infrared, 6-beam Michelson Interferometer at 2163 m elevation on Anderson Mesa in Northern Arizona, USA. Four additional 1.8-meter telescopes are being installed starting from 2013.

Light collected by three ESO VLT auxiliary telescopes, and combined using the technique of interferometry.

This image shows one of a series of sophisticated optical and mechanical systems called star separators for the Very Large Telescope Interferometer (VLTI).

Current projects will use interferometers to search for extrasolar planets, either by astrometric measurements of the reciprocal motion of the star (as used by the Palomar Testbed Interferometer and the VLTI), through the use of nulling (as will be used by the Keck Interferometer and Darwin) or through direct imaging (as proposed for Labeyrie's Hypertelescope).

Engineers at the European Southern Observatory ESO designed the Very Large Telescope VLT so that it can also be used as an interferometer. Along with the four 8.2-metre (320 in) unit telescopes, four mobile 1.8-metre auxiliary telescopes (ATs) were included in the overall VLT concept to form the Very Large Telescope Interferometer (VLTI). The ATs can move between 30 different stations, and at present, the telescopes can form groups of two or three for interferometry.

When using interferometry, a complex system of mirrors brings the light from the different telescopes to the astronomical instruments where it is combined and processed. This is technically demanding as the light paths must be kept equal to within 1/1000 mm (the same order as the wavelength of light) over distances of a few hundred metres. For the Unit Telescopes, this gives an equivalent mirror diameter of up to 130 metres (430 ft), and when combining the auxiliary telescopes, equivalent mirror diameters of up to 200 metres (660 ft) can be achieved. This is up to 25 times better than the resolution of a single VLT unit telescope.

The VLTI gives astronomers the ability to study celestial objects in unprecedented detail. It is possible to see details on the surfaces of stars and even to study the environment close to a black hole. With a spatial resolution of 4 milliarcseconds, the VLTI has allowed astronomers to obtain one of the sharpest images ever of a star. This is equivalent to resolving the head of a screw at a distance of 300 km (190 mi).

Notable 1990s results included the Mark III measurement of diameters of 100 stars and many accurate stellar positions, COAST and NPOI producing many very high resolution images, and Infrared Stellar Interferometer measurements of stars in the mid-infrared for the first time. Additional results include direct measurements of the sizes of and distances to Cepheid variable stars, and young stellar objects.

Two of the Atacama Large Millimeter/submillimeter array (ALMA) 12-metre antennas gaze at the sky at the observatory's Array Operations Site (AOS), high on the Chajnantor plateau at an altitude of 5000 metres in the Chilean Andes.

High on the Chajnantor plateau in the Chilean Andes, the European Southern Observatory (ESO), together with its international partners, is building ALMA, which will gather radiation from some of the coldest objects in the Universe. ALMA will be a single telescope of a new design, composed initially of 66 high-precision antennas and operating at wavelengths of 0.3 to 9.6 mm. Its main 12-meter array will have fifty antennas, 12 metres in diameter, acting together as a single telescope – an interferometer. An additional compact array of four 12-metre and twelve 7-meter antennas will complement this. The antennas can be spread across the desert plateau over distances from 150 metres to 16 kilometres, which will give ALMA a powerful variable "zoom". It will be able to probe the Universe at millimetre and submillimetre wavelengths with unprecedented sensitivity and resolution, with a resolution up to ten times greater than the Hubble Space Telescope, and complementing images made with the VLT interferometer.

Optical interferometers are mostly seen by astronomers as very specialized instruments, capable of a very limited range of observations. It is often said that an interferometer achieves the effect of a telescope the size of the distance between the apertures; this is only true in the limited sense of angular resolution. The amount of light gathered—and hence the dimmest object that can be seen—depends on the real aperture size, so an interferometer would offer little improvement as the image is dim (the thinned-array curse). The combined effects of limited aperture area and atmospheric turbulence generally limits interferometers to observations of comparatively bright stars and active galactic nuclei. However, they have proven useful for making very high precision measurements of simple stellar parameters such as size and position (astrometry), for imaging the nearest giant stars and probing the cores of nearby active galaxies.

For details of individual instruments, see the list of astronomical interferometers at visible and infrared wavelengths.

A simple two-element optical interferometer. Light from two small telescopes (shown as lenses) is combined using beam splitters at detectors 1, 2, 3 and 4. The elements creating a 1/4-wave delay in the light allow the phase and amplitude of the interference visibility to be measured, which give information about the shape of the light source.

A single large telescope with an aperture mask over it (labelled Mask), only allowing light through two small holes. The optical paths to detectors 1, 2, 3 and 4 are the same as in the left-hand figure, so this setup will give identical results. By moving the holes in the aperture mask and taking repeated measurements, images can be created using aperture synthesis which would have the same quality as would have been given by the right-hand telescope without the aperture mask. In an analogous way, the same image quality can be achieved by moving the small telescopes around in the left-hand figure — this is the basis of aperture synthesis, using widely separated small telescopes to simulate a giant telescope.

At radio wavelengths, interferometers such as the Very Large Array and MERLIN have been in operation for many years. The distances between telescopes are typically 10–100 km (6.2–62.1 mi), although arrays with much longer baselines utilize the techniques of Very Long Baseline Interferometry. In the (sub)-millimetre, existing arrays include the Submillimeter Array and the IRAM Plateau de Bure facility. The Atacama Large Millimeter Array has been fully operational since March 2013.

Max Tegmark and Matias Zaldarriaga have proposed the Fast Fourier Transform Telescope which would rely on extensive computer power rather than standard lenses and mirrors. If Moore's law continues, such designs may become practical and cheap in a few years.

Progressing quantum computing might eventually allow more extensive use of interferometry, as newer proposals suggest.

Significant figures

From Wikipedia, the free encyclopedia

https://en.wikipedia.org/wiki/Significant_figures

Significant figures, also referred to as significant digits or sig figs, are specific digits within a number written in positional notation that carry both reliability and necessity in conveying a particular quantity. When presenting the outcome of a measurement (such as length, pressure, volume, or mass), if the number of digits exceeds what the measurement instrument can resolve, only the number of digits within the resolution's capability are dependable and therefore considered significant.

For instance, if a length measurement yields 114.8 mm, using a ruler with the smallest interval between marks at 1 mm, the first three digits (1, 1, and 4, representing 114 mm) are certain and constitute significant figures. Even digits that are uncertain yet reliable are also included in the significant figures. In this scenario, the last digit (8, contributing 0.8 mm) is likewise considered significant despite its uncertainty.

Another example involves a volume measurement of 2.98 L with an uncertainty of ± 0.05 L. The actual volume falls between 2.93 L and 3.03 L. Even if certain digits are not completely known, they are still significant if they are reliable, as they indicate the actual volume within an acceptable range of uncertainty. In this case, the actual volume might be 2.94 L or possibly 3.02 L, so all three digits are considered significant.

The following types of digits are not considered significant:

• Leading zeros. For instance, 013 kg has two significant figures—1 and 3—while the leading zero is insignificant since it does not impact the mass indication; 013 kg is equivalent to 13 kg, rendering the zero unnecessary. Similarly, in the case of 0.056 m, there are two insignificant leading zeros since 0.056 m is the same as 56 mm, thus the leading zeros do not contribute to the length indication.
• Trailing zeros when they serve as placeholders. In the measurement 1500 m, the trailing zeros are insignificant if they simply stand for the tens and ones places (assuming the measurement resolution is 100 m). In this instance, 1500 m indicates the length is approximately 1500 m rather than an exact value of 1500 m.
• Spurious digits that arise from calculations resulting in a higher precision than the original data or a measurement reported with greater precision than the instrument's resolution.

Among a number's significant figures, the most significant is the digit with the greatest exponent value (the leftmost significant figure), while the least significant is the digit with the lowest exponent value (the rightmost significant figure). For example, in the number "123," the "1" is the most significant figure, representing hundreds (10²), while the "3" is the least significant figure, representing ones (10⁰).

To avoid conveying a misleading level of precision, numbers are often rounded. For instance, it would create false precision to present a measurement as 12.34525 kg when the measuring instrument only provides accuracy to the nearest gram (0.001 kg). In this case, the significant figures are the first five digits (1, 2, 3, 4, and 5) from the leftmost digit, and the number should be rounded to these significant figures, resulting in 12.345 kg as the accurate value. The rounding error (in this example, 0.00025 kg = 0.25 g) approximates the numerical resolution or precision. Numbers can also be rounded for simplicity, not necessarily to indicate measurement precision, such as for the sake of expediency in news broadcasts.

Significance arithmetic encompasses a set of approximate rules for preserving significance through calculations. More advanced scientific rules are known as the propagation of uncertainty.

Radix 10 (base-10, decimal numbers) is assumed in the following. (See unit in the last place for extending these concepts to other bases.)

Identifying significant figures

Rules to identify significant figures in a number

Digits in light blue are significant figures; those in black are not.

Note that identifying the significant figures in a number requires knowing which digits are reliable (e.g., by knowing the measurement or reporting resolution with which the number is obtained or processed) since only reliable digits can be significant; e.g., 3 and 4 in 0.00234 g are not significant if the measurable smallest weight is 0.001 g.

• Non-zero digits within the given measurement or reporting resolution are significant.
  • 91 has two significant figures (9 and 1) if they are measurement-allowed digits.
  • 123.45 has five significant digits (1, 2, 3, 4 and 5) if they are within the measurement resolution. If the resolution is 0.1, then the last digit 5 is not significant.
• Zeros between two significant non-zero digits are significant (significant trapped zeros).
  • 101.12003 consists of eight significant figures if the resolution is to 0.00001.
  • 125.340006 has seven significant figures if the resolution is to 0.0001: 1, 2, 5, 3, 4, 0, and 0.
• Zeros to the left of the first non-zero digit (leading zeros) are not significant.
  • If a length measurement gives 0.052 km, then 0.052 km = 52 m so 5 and 2 are only significant; the leading zeros appear or disappear, depending on which unit is used, so they are not necessary to indicate the measurement scale.
  • 0.00034 has 2 significant figures (3 and 4) if the resolution is 0.00001.
• Zeros to the right of the last non-zero digit (trailing zeros) in a number with the decimal point are significant if they are within the measurement or reporting resolution.
  • 1.200 has four significant figures (1, 2, 0, and 0) if they are allowed by the measurement resolution.
  • 0.0980 has three significant digits (9, 8, and the last zero) if they are within the measurement resolution.
  • 120.000 consists of six significant figures (1, 2, and the four subsequent zeroes) if, as before, they are within the measurement resolution.
• Trailing zeros in an integer may or may not be significant, depending on the measurement or reporting resolution.
  • 45,600 has 3, 4 or 5 significant figures depending on how the last zeros are used. For example, if the length of a road is reported as 45600 m without information about the reporting or measurement resolution, then it is not clear if the road length is precisely measured as 45600 m or if it is a rough estimate. If it is the rough estimation, then only the first three non-zero digits are significant since the trailing zeros are neither reliable nor necessary; 45600 m can be expressed as 45.6 km or as 4.56 × 10⁴ m in scientific notation, and neither expression requires the trailing zeros.
• An exact number has an infinite number of significant figures.
  • If the number of apples in a bag is 4 (exact number), then this number is 4.0000... (with infinite trailing zeros to the right of the decimal point). As a result, 4 does not impact the number of significant figures or digits in the result of calculations with it.
• A mathematical or physical constant has significant figures to its known digits.
  • π is a specific real number with several equivalent definitions. All of the digits in its exact decimal expansion 3.14159265358979323... are significant. Although many properties of these digits are known — for example, they do not repeat, because π is irrational — not all of the digits are known. As of 19 August 2021, more than 62 trillion digits have been calculated. A 62 trillion-digit approximation has 62 trillion significant digits. In practical applications, far fewer digits are used. The everyday approximation 3.14 has three significant decimal digits and 7 correct binary digits. The approximation 22/7 has the same three correct decimal digits but has 10 correct binary digits. Most calculators and computer programs can handle the 16-digit expansion 3.141592653589793, which is sufficient for interplanetary navigation calculations.
  • The Planck constant is ${\displaystyle h=6.62607015\times {10}^{-34}\mathrm{J}\cdot \mathrm{s}}$ and is defined as an exact value so that it is more properly defined as ${\displaystyle h=6.62607015(0)\times {10}^{-34}\mathrm{J}\cdot \mathrm{s}}$.

Ways to denote significant figures in an integer with trailing zeros

The significance of trailing zeros in a number not containing a decimal point can be ambiguous. For example, it may not always be clear if the number 1300 is precise to the nearest unit (just happens coincidentally to be an exact multiple of a hundred) or if it is only shown to the nearest hundreds due to rounding or uncertainty. Many conventions exist to address this issue. However, these are not universally used and would only be effective if the reader is familiar with the convention:

• An overline, sometimes also called an overbar, or less accurately, a vinculum, may be placed over the last significant figure; any trailing zeros following this are insignificant. For example, 1300 has three significant figures (and hence indicates that the number is precise to the nearest ten).

• Less often, using a closely related convention, the last significant figure of a number may be underlined; for example, "1300" has two significant figures.

• A decimal point may be placed after the number; for example "1300." indicates specifically that trailing zeros are meant to be significant.

As the conventions above are not in general use, the following more widely recognized options are available for indicating the significance of number with trailing zeros:

• Eliminate ambiguous or non-significant zeros by changing the unit prefix in a number with a unit of measurement. For example, the precision of measurement specified as 1300 g is ambiguous, while if stated as 1.30 kg it is not. Likewise 0.0123 L can be rewritten as 12.3 mL

• Eliminate ambiguous or non-significant zeros by using Scientific Notation: For example, 1300 with three significant figures becomes 1.30×10³. Likewise 0.0123 can be rewritten as 1.23×10⁻². The part of the representation that contains the significant figures (1.30 or 1.23) is known as the significand or mantissa. The digits in the base and exponent (10³ or 10⁻²) are considered exact numbers so for these digits, significant figures are irrelevant.

• Explicitly state the number of significant figures (the abbreviation s.f. is sometimes used): For example "20 000 to 2 s.f." or "20 000 (2 sf)".

• State the expected variability (precision) explicitly with a plus–minus sign, as in 20 000 ± 1%. This also allows specifying a range of precision in-between powers of ten.

Rounding to significant figures

Rounding to significant figures is a more general-purpose technique than rounding to n digits, since it handles numbers of different scales in a uniform way. For example, the population of a city might only be known to the nearest thousand and be stated as 52,000, while the population of a country might only be known to the nearest million and be stated as 52,000,000. The former might be in error by hundreds, and the latter might be in error by hundreds of thousands, but both have two significant figures (5 and 2). This reflects the fact that the significance of the error is the same in both cases, relative to the size of the quantity being measured.

To round a number to n significant figures:

1. If the n + 1 digit is greater than 5 or is 5 followed by other non-zero digits, add 1 to the n digit. For example, if we want to round 1.2459 to 3 significant figures, then this step results in 1.25.
2. If the n + 1 digit is 5 not followed by other digits or followed by only zeros, then rounding requires a tie-breaking rule. For example, to round 1.25 to 2 significant figures:
   • Round half away from zero (also known as "5/4") rounds up to 1.3. This is the default rounding method implied in many disciplines if the required rounding method is not specified.
   • Round half to even, which rounds to the nearest even number. With this method, 1.25 is rounded down to 1.2. If this method applies to 1.35, then it is rounded up to 1.4. This is the method preferred by many scientific disciplines, because, for example, it avoids skewing the average value of a long list of values upwards.
3. For an integer in rounding, replace the digits after the n digit with zeros. For example, if 1254 is rounded to 2 significant figures, then 5 and 4 are replaced to 0 so that it will be 1300. For a number with the decimal point in rounding, remove the digits after the n digit. For example, if 14.895 is rounded to 3 significant figures, then the digits after 8 are removed so that it will be 14.9.

In financial calculations, a number is often rounded to a given number of places. For example, to two places after the decimal separator for many world currencies. This is done because greater precision is immaterial, and usually it is not possible to settle a debt of less than the smallest currency unit.

In UK personal tax returns, income is rounded down to the nearest pound, whilst tax paid is calculated to the nearest penny.

As an illustration, the decimal quantity 12.345 can be expressed with various numbers of significant figures or decimal places. If insufficient precision is available then the number is rounded in some manner to fit the available precision. The following table shows the results for various total precision at two rounding ways (N/A stands for Not Applicable).

Precision	Rounded to significant figures	Rounded to decimal places
6	12.3450	12.345000
5	12.345	12.34500
4	12.34 or 12.35	12.3450
3	12.3	12.345
2	12	12.34 or 12.35
1	10	12.3
0	—	12

Another example for 0.012345. (Remember that the leading zeros are not significant.)

Precision	Rounded to significant figures	Rounded to decimal places
7	0.01234500	0.0123450
6	0.0123450	0.012345
5	0.012345	0.01234 or 0.01235
4	0.01234 or 0.01235	0.0123
3	0.0123	0.012
2	0.012	0.01
1	0.01	0.0
0	—	0

The representation of a non-zero number x to a precision of p significant digits has a numerical value that is given by the formula:

${\displaystyle {10}^n\cdot \mathrm{round}\left(\frac{x}{{10}^n}\right)}$

where

${\displaystyle n=\lfloor {\log }_{10}(|x|)\rfloor +1-p}$

which may need to be written with a specific marking as detailed above to specify the number of significant trailing zeros.

Writing uncertainty and implied uncertainty

Significant figures in writing uncertainty

It is recommended for a measurement result to include the measurement uncertainty such as ${\displaystyle x_{best}\pm {\sigma }_x}$, where x_best and σ_x are the best estimate and uncertainty in the measurement respectively. x_best can be the average of measured values and σ_x can be the standard deviation or a multiple of the measurement deviation. The rules to write ${\displaystyle x_{best}\pm {\sigma }_x}$are:

• σ_x has only one or two significant figures as more precise uncertainty has no meaning.
  • 1.79 ± 0.06 (correct), 1.79 ± 0.96 (correct), 1.79 ± 1.96 (incorrect).
• The digit positions of the last significant figures in x_best and σ_x are the same, otherwise the consistency is lost. For example, in 1.79 ± 0.067 (incorrect), it does not make sense to have more accurate uncertainty than the best estimate. 1.79 ± 0.9 (incorrect) also does not make sense since the rounding guideline for addition and subtraction below tells that the edges of the true value range are 2.7 and 0.9, that are less accurate than the best estimate.
  • 1.79 ± 0.06 (correct), 1.79 ± 0.96 (correct), 1.79 ± 0.067 (incorrect), 1.79 ± 0.9 (incorrect).

Implied uncertainty

In chemistry (and may also be for other scientific branches), uncertainty may be implied by the last significant figure if it is not explicitly expressed. The implied uncertainty is ± the half of the minimum scale at the last significant figure position. For example, if the volume of water in a bottle is reported as 3.78 L without mentioning uncertainty, then ± 0.005 L measurement uncertainty may be implied. If 2.97 ± 0.07 kg, so the actual weight is somewhere in 2.90 to 3.04 kg, is measured and it is desired to report it with a single number, then 3.0 kg is the best number to report since its implied uncertainty ± 0.05 kg tells the weight range of 2.95 to 3.05 kg that is close to the measurement range. If 2.97 ± 0.09 kg, then 3.0 kg is still the best since, if 3 kg is reported then its implied uncertainty ± 0.5 tells the range of 2.5 to 3.5 kg that is too wide in comparison with the measurement range.

If there is a need to write the implied uncertainty of a number, then it can be written as ${\displaystyle x\pm {\sigma }_x}$with stating it as the implied uncertainty (to prevent readers from recognizing it as the measurement uncertainty), where x and σ_x are the number with an extra zero digit (to follow the rules to write uncertainty above) and the implied uncertainty of it respectively. For example, 6 kg with the implied uncertainty ± 0.5 kg can be stated as 6.0 ± 0.5 kg.

Arithmetic

Main article: Significance arithmetic

As there are rules to determine the significant figures in directly measured quantities, there are also guidelines (not rules) to determine the significant figures in quantities calculated from these measured quantities.

Significant figures in measured quantities are most important in the determination of significant figures in calculated quantities with them. A mathematical or physical constant (e.g., π in the formula for the area of a circle with radius r as πr²) has no effect on the determination of the significant figures in the result of a calculation with it if its known digits are equal to or more than the significant figures in the measured quantities used in the calculation. An exact number such as ½ in the formula for the kinetic energy of a mass m with velocity v as ½mv² has no bearing on the significant figures in the calculated kinetic energy since its number of significant figures is infinite (0.500000...).

The guidelines described below are intended to avoid a calculation result more precise than the measured quantities, but it does not ensure the resulted implied uncertainty close enough to the measured uncertainties. This problem can be seen in unit conversion. If the guidelines give the implied uncertainty too far from the measured ones, then it may be needed to decide significant digits that give comparable uncertainty.

Multiplication and division

For quantities created from measured quantities via multiplication and division, the calculated result should have as many significant figures as the least number of significant figures among the measured quantities used in the calculation. For example,

• 1.234 × 2 = 2.468 ≈ 2
• 1.234 × 2.0 = 2.468 ≈ 2.5
• 0.01234 × 2 = 0.02468 ≈ 0.02

with one, two, and one significant figures respectively. (2 here is assumed not an exact number.) For the first example, the first multiplication factor has four significant figures and the second has one significant figure. The factor with the fewest or least significant figures is the second one with only one, so the final calculated result should also have one significant figure.

Exception

For unit conversion, the implied uncertainty of the result can be unsatisfactorily higher than that in the previous unit if this rounding guideline is followed; For example, 8 inch has the implied uncertainty of ± 0.5 inch = ± 1.27 cm. If it is converted to the centimeter scale and the rounding guideline for multiplication and division is followed, then 20.32 cm ≈ 20 cm with the implied uncertainty of ± 5 cm. If this implied uncertainty is considered as too overestimated, then more proper significant digits in the unit conversion result may be 20.32 cm ≈ 20. cm with the implied uncertainty of ± 0.5 cm.

Another exception of applying the above rounding guideline is to multiply a number by an integer, such as 1.234 × 9. If the above guideline is followed, then the result is rounded as 1.234 × 9.000.... = 11.106 ≈ 11.11. However, this multiplication is essentially adding 1.234 to itself 9 times such as 1.234 + 1.234 + … + 1.234 so the rounding guideline for addition and subtraction described below is more proper rounding approach. As a result, the final answer is 1.234 + 1.234 + … + 1.234 = 11.106 = 11.106 (one significant digit increase).

Addition and subtraction

For quantities created from measured quantities via addition and subtraction, the last significant figure position (e.g., hundreds, tens, ones, tenths, hundredths, and so forth) in the calculated result should be the same as the leftmost or largest digit position among the last significant figures of the measured quantities in the calculation. For example,

• 1.234 + 2 = 3.234 ≈ 3
• 1.234 + 2.0 = 3.234 ≈ 3.2
• 0.01234 + 2 = 2.01234 ≈ 2

with the last significant figures in the ones place, tenths place, and ones place respectively. (2 here is assumed not an exact number.) For the first example, the first term has its last significant figure in the thousandths place and the second term has its last significant figure in the ones place. The leftmost or largest digit position among the last significant figures of these terms is the ones place, so the calculated result should also have its last significant figure in the ones place.

The rule to calculate significant figures for multiplication and division are not the same as the rule for addition and subtraction. For multiplication and division, only the total number of significant figures in each of the factors in the calculation matters; the digit position of the last significant figure in each factor is irrelevant. For addition and subtraction, only the digit position of the last significant figure in each of the terms in the calculation matters; the total number of significant figures in each term is irrelevant.^{[citation needed]} However, greater accuracy will often be obtained if some non-significant digits are maintained in intermediate results which are used in subsequent calculations.

Logarithm and antilogarithm

The base-10 logarithm of a normalized number (i.e., a × 10^b with 1 ≤ a < 10 and b as an integer), is rounded such that its decimal part (called mantissa) has as many significant figures as the significant figures in the normalized number.

• log₁₀(3.000 × 10⁴) = log₁₀(10⁴) + log₁₀(3.000) = 4.000000... (exact number so infinite significant digits) + 0.4771212547... = 4.4771212547 ≈ 4.4771.

When taking the antilogarithm of a normalized number, the result is rounded to have as many significant figures as the significant figures in the decimal part of the number to be antiloged.

• 10^4.4771 = 29998.5318119... = 30000 = 3.000 × 10⁴.

Transcendental functions

If a transcendental function ${\displaystyle f(x)}$ (e.g., the exponential function, the logarithm, and the trigonometric functions) is differentiable at its domain element x, then its number of significant figures (denoted as "significant figures of ${\displaystyle f(x)}$") is approximately related with the number of significant figures in x (denoted as "significant figures of x") by the formula

${\displaystyle (\mathrm{s}\mathrm{i}\mathrm{g}\mathrm{n}\mathrm{i}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{t}\mathrm{f}\mathrm{i}\mathrm{g}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{o}\mathrm{f}\mathrm{f}(\mathrm{x}))\approx (\mathrm{s}\mathrm{i}\mathrm{g}\mathrm{n}\mathrm{i}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{t}\mathrm{f}\mathrm{i}\mathrm{g}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{o}\mathrm{f}\mathrm{x})-{\log }_{10}\left(\left|\frac{df(x)}{dx}\frac{x}{f(x)}\right|\right)}$,

where ${\displaystyle \left|\frac{df(x)}{dx}\frac{x}{f(x)}\right|}$ is the condition number. See the significance arithmetic article to find its derivation.

Round only on the final calculation result

When performing multiple stage calculations, do not round intermediate stage calculation results; keep as many digits as is practical (at least one more digit than the rounding rule allows per stage) until the end of all the calculations to avoid cumulative rounding errors while tracking or recording the significant figures in each intermediate result. Then, round the final result, for example, to the fewest number of significant figures (for multiplication or division) or leftmost last significant digit position (for addition or subtraction) among the inputs in the final calculation.

• (2.3494 + 1.345) × 1.2 = 3.6944 × 1.2 = 4.43328 ≈ 4.4.
• (2.3494 × 1.345) + 1.2 = 3.159943 + 1.2 = 4.359943 ≈ 4.4.

Estimating an extra digit

When using a ruler, initially use the smallest mark as the first estimated digit. For example, if a ruler's smallest mark is 0.1 cm, and 4.5 cm is read, then it is 4.5 (±0.1 cm) or 4.4 cm to 4.6 cm as to the smallest mark interval. However, in practice a measurement can usually be estimated by eye to closer than the interval between the ruler's smallest mark, e.g. in the above case it might be estimated as between 4.51 cm and 4.53 cm.

It is also possible that the overall length of a ruler may not be accurate to the degree of the smallest mark, and the marks may be imperfectly spaced within each unit. However assuming a normal good quality ruler, it should be possible to estimate tenths between the nearest two marks to achieve an extra decimal place of accuracy. Failing to do this adds the error in reading the ruler to any error in the calibration of the ruler.

Estimation in statistic

Main article: Estimation

When estimating the proportion of individuals carrying some particular characteristic in a population, from a random sample of that population, the number of significant figures should not exceed the maximum precision allowed by that sample size.

Relationship to accuracy and precision in measurement

Main article: Accuracy and precision

Traditionally, in various technical fields, "accuracy" refers to the closeness of a given measurement to its true value; "precision" refers to the stability of that measurement when repeated many times. Thus, it is possible to be "precisely wrong". Hoping to reflect the way in which the term "accuracy" is actually used in the scientific community, there is a recent standard, ISO 5725, which keeps the same definition of precision but defines the term "trueness" as the closeness of a given measurement to its true value and uses the term "accuracy" as the combination of trueness and precision. (See the accuracy and precision article for a full discussion.) In either case, the number of significant figures roughly corresponds to precision, not to accuracy or the newer concept of trueness.

In computing

Main article: Floating-point arithmetic

Computer representations of floating-point numbers use a form of rounding to significant figures (while usually not keeping track of how many), in general with binary numbers. The number of correct significant figures is closely related to the notion of relative error (which has the advantage of being a more accurate measure of precision, and is independent of the radix, also known as the base, of the number system used).