Given a set of n disjoint balls b_1, ..., b_n in R^d, we provide a data structure, of near linear size, that can answer (1 +- epsilon)-approximate k-th-nearest neighbor queries in O(log(n) + 1/epsilon^d) time, where k and epsilon are provided at query time. If k and epsilon are provided in advance, we provide a data structure to answer such queries, that requires (roughly) O(n/k) space; that is, the data structure has sublinear space requirement if k is sufficiently large.