Quantitative arithmetic of projective varieties (Q1030571)

Let \(X\) be an absolutely irreducible algebraic variety defined over \(\mathbb{Q}\). This book is concerned with the behaviour, as \(B\to\infty\), of the counting function \[ N_X(B):=\#\{P\in X(\mathbb{Q}): H(P)\leq B\}, \] where \(H(P)\) is the standard multiplicative height. There are three principal motivations for this. Firstly, it is a study of interest in its own right. One finds that the behaviour of \(N_X(B)\) is linked to the intrinsic geometry of the variety \(X\), but that ``geometry over \(\mathbb{Q}\)'' also has features of its own, just as algebraic geometry over \(\mathbb{R}\) does. Secondly, there are many instances in which a study of \(N_X(B)\) shows that \(N_X(B)\to\infty\) as \(B\) grows, thereby proving that \(X(\mathbb{Q})\) is non-empty. Often this is most natural way to show that the corresponding Diophantine system has rational solutions. Thirdly there are many diverse applications of estimates for \(N_X(B)\) in seemingly unrelated questions in analytic number theory. The use of Vinogradov's mean-value theorem to prove the sharpest known form of the prime number theorem is a good example. Chapter 1, ``Introduction'' describes the basic heuristic for predicting the coarse behaviour of \(N_X(B)\), and explains phenomena that can cause departures from this expectation. The relevance to such classical questions as Euler's conjecture and Waring's problem is also outlined. The second chapter ``The Manin conjectures'' concerns Manin's more precise formulation of the previous heuristic. A certain amount of familiarity with algebraic geometry is necessary at this point, as the author runs through the Picard group, the cone of effective divisors, Fano varieties and del Pezzo surfaces. He outlines a number of results that have been obtained for the latter, some in which \(X\) is singular, and others in which \(X\) is smooth. The chapter ends with a brief look at universal torsors and their use in connection with Manin's conjecture. Chapter 3 is ``The dimension growth conjecture'', which asserts that if \(X\) is absolutely irreducible, is not linear, and has dimension \(d\), then \[ N_X(B)\ll_{\varepsilon,X} B^{d+\varepsilon} \] for any \(\varepsilon> 0\). A number of results towards this conjecture are described, and an application of exponential sums to the problem is presented. The next chapter, ``Uniform bounds for curves and surfaces'', is devoted principally to applications of the ``determinant method'', which shows for example that \[ N_X(B)\ll_{\varepsilon, d,n} B^{2/d+\varepsilon} \] for a curve \(X\in\mathbb{P}^n\) of degree \(d\). There is also an account of consequences of the reviewer's ``Rank Hypothesis'', as well as a discussion of results in which \(X\) is a surface rather than a curve. Up to this point the book has consisted largely of survey material, but Chapters 5 and 6 give detailed treatments respectively of the asymptotic formula for \(N_X(B)\) in the case of a certain \(A_1\) del Pezzo surface of degree 6, and of an upper bound (of the predicted order of magnitude) when \(X\) is a certain \(D_4\) del Pezzo surface of degree 3. In both cases the pasage to the universal torsor is described in full, along with the necessary analytic techniques. Chapter 7 is concerned with ``slicing'' arguments, using Siegel's lemma. These enable one to reduce the estimation on \(N_X(B)\) for a variety \(X\) of higher dimension, to a problem in which one has only curves. The final chapter introduces the circle method and gives a detailed proof of the asymptotic formula for \(N_X(B)\) in the case in which \(X\) is a smooth quartic hypersurface of dimension at least 47. (This is a particular case of a result of Birch.) There is then a discussion of how the circle method might apply to diagonal cubic surfaces. The chapter concludes with a discussion of the convergence of the singular series, relating this issue with Manin's conjecture and with Peyre's conjecture for the leading constant. Naturally it is impossible to give much by way of proofs in a book with this extensive scope. However the author has managed to give the main ideas without involving too many technicalities. There is much that readers will need to search out from other sources, but they will have been given an excellent starting point. The most important feature of the book is the way it presents the geometric and analytic aspects of the theory on a unified equal footing. The interface between these two fields has been a very productive subject in recent years, and this book is likely to be of considerable value to anyone, graduate student and up, interested in this area.