ABSTRACT Systematic laboratory measurements were performed to determine the steady drift of isolated, tabular ice-floe models in regular waves. A relatively simple expression for the steady wave-induced drift of two-dimensional floes of uniform thickness and density in deep water was developed. This empirical wave-drift formula indicates that the drift velocity depends primarily on the ratio of wave period to ice-floe roll period, wave steepness and length of the floe. The highest drift velocity is produced by waves with period somewhat greater than the ice-floe roll period, with fundamentally different drift behavior evident on either side of this maximum (long-wave and short-wave drift response). The dependence on wave height was found not to be linear but increased with wave length (from approximately H1/2 to H3/2). A simplified equation for the roll period of a prismatic body of uniform density and rectangular cross-section was developed and found to be of practical value. INTRODUCTION An object floating freely at the surface is generally exposed to the combined action of winds, waves and currents, and will normally undergo oscillatory motions and also experience a net drift with respect to the seafloor. When ocean currents and wind forcing can be neglected, the body will drift in the approximate direction of wave advance while simultaneously undergoing horizontal oscillations at the wave frequency. Drift rates are typically small so that the net drift is often only of interest over time periods much longer than the wave period. Horizontal motions diminish as the wave length decreases, so that the associated oscillatory velocities can often be neglected compared to the drift velocity, i.e., the steady drift velocity and maximum instantaneous velocity are practically identical for short waves. By time-averaging the horizontal motions of the floating body (over periods large compared to the wave period), the so-called steady drift velocity is obtained. Although the steady drift excludes motions at wave frequency, it is still of practical interest because (a) it permits time-of-travel estimates to be made (when will an abandoned barge or an ice floe reach a given location); (b) it provides information about the rate of dispersal to be expected for a group of floating objects of different size (wave-induced separation of ice floes); and (c) it can be used to estimate the maximum kinetic energy with which drifting objects can strike a fixed or floating structure (at least for the case of relatively short waves). We report here the results of laboratory measurements of the steady (mean) drift of ice-floe models in regular waves. This is the first part of an ongoing effort to define the wave-induced motions (both mean and instantaneous) of ice-floes nearing the point of impact with fixed structures or sand and gravel beaches. The present investigation was performed with regular progressive waves as the only driving force. In the absence of winds and large-scale ocean currents, the body is propelled by wave action alone, and a separation into a wave-induced surface current Vs and a wave-thrust velocity may be conceptually helpful.