A long-standing question in the theory of rational points of algebraic surfaces is whether a K3 surface X over a number field K acquires a Zariski-dense set of L-rational points over some finite extension L/K. In this case, we say X has potential density of rational points. In case X C has Picard rank greater than 1, Bogomolov and Tschinkel [2] have shown in many cases that X has potential density of rational points, using the existence of elliptic fibrations on X or large automorphism groups of X. By contrast, we do not know a single example of a K3 surface X/K with geometric Picard number 1 which can be shown to have potential density of rational points; nor is there an example which we can show not to have potential density of rational points. In fact, the situation is even worse; the moduli space of polarized K3 surfaces of a given degree contains a countable union of subvarieties, each parametrizing a family of K3 surfaces with geometric Picard number greater than 1. SinceQ is countable, it is not a priori obvious that these subvarieties don't cover theQ-points of the moduli space. In other words, it is a non-trivial fact that there exists a K3 surface over any number field with geometric Picard number 1! In this note, we correct this slightly embarrassing situation by proving the following theorem:Theorem 1. Let d be an even positive integer. Then there exists a number field K and a polarized K3 surface X/K, of degree d, such that rank Pic(X C ) = 1.The main idea is to use an argument of Serre on -adic groups to reduce the problem to proving the existence of K3 surfaces whose associated mod-n Galois representations have large image for some finite n; we then use Hilbert's irreducibility theorem and global Torelli for K3's to complete the proof.Acknowledgment: This note is the result of a conversation between the author, Brendan Hassett, and A.J. de Jong, which took place at the American Institute of Mathematics during the workshop, "Rational and integral points on higher-dimensional varieties." It should also be pointed out that the main idea, in case d = 4, is implicit in the final remark of [3].