We wish to describe an apparatus for preparative electrophoresis on polyacrylamide, starch, or other gels, that we have found useful and that has advantages over other available apparatus for the same purpose. The very high resolving power of polyacrylamide gels as introduced by Ornstein and Davis (1) has quite naturally led to design of a number of preparative methods using the same gel. The principle of carrying the materials through the gel and into a collecting electrolyte seems more appealing than extracting the gel. Designs employing this principle usually have a constant flow of collecting electrolyte and the materials of interest are prevented from leaving it by a semipermeable membrane (2-7). One device uses the gel for its separating ability but collects the material in a density gradient that is fractionally collected after electrophoresis (8). Continuous flow of collecting electrolyte through narrow slits requires mechanical strength provided by a sintered glass or porous plastic disc or careful hydrostatic balancing of the deformable gel surface. The discs are electrically resistive and tend to heat. Hydrostatic balancing requires meticulous care in setting up. Most devices make provision for cooling the surfaces of the gel tube by circulating water. Since the gel can only transfer heat by conduction no amount of external cooling will prevent the production of a temperature gradient from center to surface (9). This is about 9-10" for a gel 1.4 cm in diameter under the usual running conditions. Continuous flow of collecting electrolyte results in high dilution between applied sample and product. A considerable part of this difficulty is connected with "holdup" at the gel surface. This factor is discussed below.