(Received 20 October 1975) §1. THE PLr~OSE of this paper is to present some information about the following Question: I[ M is a compact manifold [oliated by compact submani[olds (everything smooth), is there an upper bound on the volume of the leaves?t In particular, if M is a compact manifold supporting a nonsingular flow in which each orbit is periodic, is there an upper bound on the lengths of the orbits? In his thesis, Reeb ([12], for an analytic example see [3, p. 68]) describes a smooth flow on a non-compact manifold such that all orbits are periodic and such that the lengths of the orbits are not locally bounded. After our research was completed, the third author found a smooth flow on a closed 5-manifold[15, 16], which showed that the answer, in.general, was no. The former example shows that the question is global and cannot be answered by simply considering the structure of the foliation in a neighborhood of individual compact leaves. The latter example shows that some additional hypothesis on M is required. This example is worth keeping in mind while reading this paper due to its close connection with our main result. The existence of an upper bound on the volume of the leaves has rather important consequences which provide a description of the local, as well as global, structure of the foliation. The boundedness of volume near any given leaf is equivalent to the finiteness of the holonomy group of that leaf, and also to the hausdroff separation property for the topology of the leaf space near the leaf ([4], [8], see also §4). Hence, in the presence of a bound on the volume, a structure theorem due to Ehresmann [4, Theorem 4.3] provides a nice picture of the local behavior of the foliation. In the absence of such a bound the geometrical possibilities are somewhat formidable. Concerning the known cases of the Question, if the leaves have condimension 1 then the bound exists by a relatively elementary argument [12]. For periodic flows on compact 3-manifolds Epstein [3] has demonstrated the existence of a bound by a surprisingly delicate argument. Our two principal results were obtained in trying to understand Epstein's argument. Our first result is that in the presence of a certain homological assumption, the answer to the Question is yes. THEOREM 1. Suppose M is a compact smooth manifold which is smoothly foliated by compact leaves of dimension d. Suppose that the leaves are oriented in a continuous manner, and that the images of the fundamental classes of the leaves all lie in some open hal[space of the d-dimensional real homology of M. Then there is an upper bound on the volumes of the leaves of M. Consequently, all the holonomy groups of the foliation are finite. Another way of stating the homologicai condition is to say there exists a closed d-form to on M, such that to has a positive integral along each leaf. Here M may have boundary, in which case we assume that the boundary is a union of leaves. The with-boundary version of the Theorem follows from the without-boundary version by doubling M along its boundary. Two particular situations to which the theorem applies are when (i) each leaf of the foliation