This report outlines scientific issues whose resolution will help advance and define the field of multiphase flow. It presents the findings of four study groups and of a workshop sponsored by the Program on Engineering Physics of the Department of Energy. The reason why multiphase flows are much more difficult to analyze than single phase flows is that the phases assume a large number of complicated configurations. Therefore, it should not be surprising that the understanding of why the phases configure in a certain way is the principal scientific issue. Research is needed which identifies the microphysics controlling the organization of the phases, which develops physical models for the resultant multiscale interactions and which tests their validity in integrative experiments/theories that look at the behavior of a system. New experimental techniques and recently developed direct numerical simulations will play important roles in this endeavor. In gas-liquid flows a top priority is to develop an understanding of why the liquid phase in quasi-fully-developed pipe flow changes from one configuration to another. Mixing flows offer a more complicated situation in which several patterns can exist at the same time. They introduce new physical challenges. A second priority is to provide a quantitative description of the phase distribution for selected fully developed flows and for simple mixing flows (that could include heat transfer and phase change). Microphysical problems of interest are identified-including the coupling of molecular and macroscopic behavior that can be observed in many situations and the formation/destruction of interfaces in the coalescence/breakup of drops and bubbles. Solid-fluid flows offer a simpler system in that interfaces are not changing. However, a variety of patterns exist, that depend on the properties of the particles, their concentration and the Reynolds number characterizing the relative velocity. A top priority is the development of a physical understanding of inertial instabilities which give rise to structural features that have a large range of scales. Important microphysical problems are the understanding of particle/particle interactions, particle/boundary interactions that include the effect of wall roughness, and the influence of particles on fluid turbulence. These behaviors can differ depending on the characteristics of the particles, their size distribution and their concentration. For large concentrations, such as exist in granular flows, instabilities associated with particle-particle interactions often cause separations in systems which are not homogeneous. These instabilities are not well understood. Appropriate averaged equations could provide a way to use an understanding of the microphysics obtained in simple systems to describe more complicated flows. The formulation of these equations presents physical challenges since the structure of the phase distribution could affect the choice of averaging methods and closure relations. Universal computational approaches appear to be out of reach at present. (C) 2003 Elsevier Ltd. All rights reserved.