“It is evident that the primes are randomly distributed but, unfortunately, we don’t know what ‘random ’ means. ” — R. C. Vaughan (February 1990). After the first world war, Cramér began studying the distribution of prime numbers, guided by Riesz and Mittag-Leffler. His works then, and later in the midthirties, have had a profound influence on the way mathematicians think about the distribution of prime numbers. In this article, we shall focus on how Cramér’s ideas have directed and motivated research ever since. One can only fully appreciate the significance of Cramér’s contributions by viewing his work in the appropriate historical context. We shall begin our discussion with the ideas of the ancient Greeks, Euclid and Eratosthenes. Then we leap in time to the nineteenth century, to the computations and heuristics of Legendre and Gauss, the extraordinarily analytic insights of Dirichlet and Riemann, and the crowning glory of these ideas, the proof the “Prime Number Theorem ” by Hadamard and de la Vallée Poussin in 1896. We pick up again in the 1920’s with the questions asked by Hardy and Littlewood,

Abstract. We consider the problem of proving that a user has selected and correctly employed a truly random seed in the generation of her RSA key pair. This task is related to the problem of key validation, the process whereby a user proves to another party that her key pair has been generated securely. The aim of key validation is to pursuade the verifying party that the user has not intentionally weakened or reused her key or unintentionally made use of bad software. Previous approaches to this problem have been ad hoc, aiming to prove that a private key is secure against specific types of attacks, e.g., that an RSA modulus is resistant to elliptic-curve-based factoring attacks. This approach results in a rather unsatisfying laundry list of security tests for keys. We propose a new approach that we refer to as key generation with verifiable randomness (KEGVER). Our aim is to show in zero knowledge that a private key has been generated at random according to a prescribed process, and is therefore likely to benefit from the full strength of the underlying cryptosystem. Our proposal may be viewed as a kind of distributed key generation protocol involving the user and verifying party. Because the resulting private key is held solely by the user, however, we are able to propose a protocol much more practical than conventional distributed key generation. We focus here on a KEGVER protocol for RSA key generation. Key words: certificate authority, key generation, non-repudiation, publickey infrastructure, verifiable randomness, zero knowledge 1

Dedicated to Freeman Dyson, with best wishes on the occasion of his eightieth birthday. Abstract. Contrary to what would be predicted on the basis of Cramér’s model concerning the distribution of prime numbers, we develop evidence that the distribution of ψ(x + H) − ψ(x), for 0 ≤ x ≤ N, is approximately normal with mean ∼ H and variance ∼ H log N/H, when N δ ≤ H ≤ N 1−δ. Cramér [4] modeled the distribution of prime numbers by independent random variables Xn (for n ≥ 3) that take the value 1 (n is “prime”) with probability 1 / logn and take the value 0 (n is “composite”) with probability 1 − 1 / log n. If pn denotes the n th prime

Abstract. Analytic number theorists usually seek to show that sequences which appear naturally in arithmetic are “well-distributed ” in some appropriate sense. In various discrepancy problems, combinatorics researchers have analyzed limitations to equi-distribution, as have Fourier analysts when working with the “uncertainty principle”. In this article we find that these ideas have a natural setting in the analysis of distributions of sequences in analytic number theory, formulating a general principle, and giving several examples. 1.

this paper we are primarily interested in further developing the theory of quadratic polynomials for which many of the small values are prime (rather than "all" as in Rabinowitsch's result). It is well-known (see [5]) that if the class number of some imaginary quadratic field with large discriminant is one then we will have an egregious counterexample to the Generalized Riemann Hypothesis (that is, a zero of the associated Dirichlet L-function which is very close to 1). Thus Rabinowitsch's result can be informally stated as "n

. Let d 0 (n) = pn , the n-th prime, for n 1, and let d k+1 (n) = jd k (n) \Gamma d k (n + 1)j for k 0, n 1. A well known conjecture, usually ascribed to Gilbreath but actually due to Proth in the 19-th century, says that d k (1) = 1 for all k 1. This paper reports on a computation that verified this conjecture for k ß(10 13 ) ß 3 \Theta 10 11 . It also discusses the evidence and the heuristics about this conjecture. It is very likely that similar conjectures are also valid for many other integer sequences. 1. Introduction Let p 1 = 2, p 2 = 3; : : : be the primes in their natural ordering, and set d 0 (n) = p n ; n 1 d k+1 (n) = jd k (n) \Gamma d k (n + 1)j ; k 0; n 1 : (1.1) Table 1 shows d k (n) for 0 k 20, 1 n 20. Note that d k (1) = 1 for 1 k 20. As was pointed out by H. C. Williams, Proth [15] claimed to prove that d k (1) = 1 for all k 1, but his proof was faulty. More recently, Gilbreath (unpublished) independently conjectured that d k (1) = 1 for all k ...

The set of cycle lengths of almost all permutations in Sn are “Poisson distributed”: we show that this remains true even when we restrict the number of cycles in the permutation. The formulas we develop allow us to also show that almost all permutations with a given number of cycles have a certain “normal order” (in the spirit of the Erdős-Turán theorem). Our results were inspired by analogous questions about the size of the prime divisors of “typical ” integers. 1

What follows is an expanded version of my lectures at the NATO School on Equidistribution. I have tried to keep the informal style of the lectures. In particular, I have sometimes oversimplified matters in order to convey the spirit of an argument. Lecture 1: The Cramér model and gaps between consecutive primes The prime number theorem tells us that π(x), the number of primes below x, is ∼ x / logx. Equivalently, if pn denotes the n-th smallest prime number then pn ∼ n log n. What is the distribution of the gaps between consecutive primes, pn+1 − pn? We have just seen that pn+1 − pn is approximately log n “on average”. How often do we get a gap of size 2 logn, say; or of size 1 log n? One way to make this question precise 2 is to fix an interval [α, β] (with 0 ≤ α < β) and ask for

In 1972, H.L. Montgomery and F. Dyson uncovered a surprising connection between the Theory of the Riemann Zeta function and Random Matrix Theory. For the next few decades, the major developments in the area were the numerical calculations of Odlyzko and conjectures for the moments of the Riemann Zeta function (and other L-functions) found by Conrey, Ghosh, Gonek, Heath-Brown, Hejhal and Sarnak. Recently, there have been two important advances. First, Keating and Snaith, in a 2000 paper, conjectured connections between the moments of the characteristic polynomials of random matrices and the moments of the Riemann Zeta function. Second, Katz and Sarnak proposed connections between certain families of L-functions and other matrix groups. Our goal in this paper is twofold. First, we discuss links between Random Matrix Theory and the Zeta function. Then, we describe our numerical calculation of the moments of the Zeta function and compare initial results with Random Matrix Theory predictions. 1

: In an earlier paper [FG] we showed that the expected asymptotic formula ß(x; q; a) ¸ ß(x)=OE(q) does not hold uniformly in the range q ! x= log N x, for any fixed N ? 0. There are several reasons to suspect that the expected asymptotic formula might hold, for large values of q, when a is kept fixed. However, by a new construction, we show herein that this fails in the same ranges, for a fixed and, indeed, for almost all a satisfying 0 ! jaj ! x= log N x. 1. Introduction. For any positive integer q and integer a coprime to q, we have the asymptotic formula (1:1) ß(x; q; a) ¸ ß(x) OE(q) as x ! 1, for the number ß(x; q; a) of primes p x with p j a (mod q), where ß(x) is the number of primes x, and OE is Euler's function. In fact (1.1) is known to hold uniformly for (1:2) q ! log N x and all (a; q) = 1, for every fixed N ? 0 (the Siegel--Walfisz Theorem), for almost all q ! x 1=2 = log 2+" x and all (a; q) = 1 (the Bombieri--Vinogradov Theorem) and for almost all q !...