A freezing process and the resulting injury or survival of biological cells is commonly characterized in terms of the cooling rate, B. Under certain circumstances, the cooling rate can be expressed as B = G · v, where G denotes the thermal gradient at the ice-liquid interface and v its velocity, respectively. To determine the influence of G and v on the morphology of the ice-liquid interface and on cell survival, a gradient freezing stage was designed. Flat capillaries could be pushed with constant velocity from a warm to a cold heat reservoir. With this setup both parameters, G and v, are independently adjustable and the resulting process of directional solidification can be observed dynamically in a light microscope. Human lymphocytes in phosphate-buffered saline with 10 vol% of dimethyl sulfoxide were used as biological test material. Viability was assessed by a membrane integrity test with fluorescein diacetate and ethidium bromide. All cells were cooled down to a final temperature of -196 °C and then rapidly thawed. The results obtained with this technique show that the viability determined after freezing and thawing with a certain cooling rate, B = G · v, may vary considerably depending on the imposed values of the thermal gradient, G, and the ice front velocity, v. In addition, the data seem to suggest that, first, the maximum viability which can be reached is governed by the cooling rate, and, second, this maximum for a given cooling rate could be achieved by establishing small temperature gradients and high interface velocities (about 30 ° K/cm and 500 (μm/sec, respectively, for the range of values of G and v tested).