Abstract. We give a very general and flexible way of producing measure extensions. We obtain as corollaries many well-known and important measure extension and integral representation theorems as well as the main theorems of several recent papers.Introduction. The question of when one can extend a measure (finitely, countably, or arbitrarily additive) is a very old one. Many different procedures over the years have been applied to obtain different types of significant extension theorems, mostly in topological spaces. Many of these have found numerous important applications in diverse branches of mathematics. Some of the more recent extension theorems have found applications in solving topological questions and are giving a great deal of insight into the interplay between measure and topology (see, in particular, [4], [5], [6]). In this paper we give a single, unifying procedure, which allows one to get simultaneously many of the well-known measure extension theorems. The method presented here is particularly useful for a variety of reasons: (1) It lends itself immediately to applications in the finite or totally finite cases. (2) It gives a great deal of flexibility in the actual construction of the measure extension. (3) In most of the important cases where the extension of the measure is not unique, this procedure gives all possible ways of extending the given measure. (4) The method ties in very naturally with many well-known representation theorems, and one obtains many of these as corollaries. (5) The actual construction of the measure extension is simple.In the very simplest cases we obtain many nontrivial extension theorems for locally compact T2 spaces. When our results are applied appropriately in many of the other cases, we obtain the main theorems of Definitions. By a paving (X, E) we mean a set X together with a lattice £ of subsets of X. We will suppress X and just say that £ is a paving. If 0 E £