We prove that given any small but fixed η > 0, a positive proportion of all gaps between consecutive primes are smaller than η times the average gap. We show some unconditional and conditional quantitative results in this vein. In the results the dependence on η is given explicitly, providing a new quantitative way, in addition to that of the first paper in this series, of measuring the effect of the knowledge on the level of distribution of primes.

We study so-called real zeros of holomorphic Hecke cusp forms, that is, zeros on three geodesic segments on which the cusp form (or a multiple of it) takes real values. Ghosh and Sarnak, who were the first to study this problem, showed the existence of many such zeros if many short intervals contain numbers whose prime factors all belong to a certain subset of the primes.We prove new results concerning this sieving problem which leads to improved lower bounds for the number of real zeros.

The Selberg sieve provides majorants for certain arithmetic sequences, such as the primes and the twin primes. We prove an $L^2$–$L^p$ restriction theorem for majorants of this type. An immediate application is to the estimation of exponential sums over prime $k$-tuples. Let $a_1,\cdots ,a_k$ and $b_1,\cdots ,b_k$ be positive integers. Write $h\left(\theta \right):={\sum }_{n\in X}e\left(n\theta \right)$, where $X$ is the set of all $n\le N$ such that the numbers $a_{1}n+b_1,\cdots ,a_{k}n+b_k$ are all prime. We obtain upper bounds for ${\parallel h\parallel }_{L^p\left(𝕋\right)}$, $p\>2$, which are (conditionally on the Hardy-Littlewood prime tuple conjecture) of the correct order...

Soit $q\in \mathbb{N}$, $q\ge 2$. Pour $n\in \mathbb{N}$, on note $s_q\left(n\right)$ la somme des chiffres de $n$ en base $q$. Nous donnons des majorations de sommes d’exponentielles de la forme$$G(x,y,\theta ;\alpha ,𝐡)=\sum _{x\<n\le x+y}exp\left(2i\pi ({\alpha }_1{s}_q\left(h_{1}n\right)+\cdots +{\alpha }_r{s}_q\left(h_{r}n\right)+\theta n)\right),$$pour $r\in {\mathbb{N}}^{*}$, $𝐡\in {\mathbb{N}}^{*r}$ et $\theta \in r$. De telles sommes ont déjà été étudiées dans le cas $r=1$ par Gelfond, et pour $r\ge 2$ entre autre par Coquet et Solinas. Nos résultats étendent le domaine de validité en $𝐡$ de ces précédents travaux pour $r\ge 2$, sont plus précis et ont l’avantage d’être uniformes en $x$ et $r$ et effectifs en $𝐡$. Ce contrôle soigneux des paramètres nous permet d’obtenir divers types d’applications....