Winter hardiness of genotypes pre-selected from Cerapadus × Prunus avium L. crosses was studied for 3 years (2000)(2001)(2002) in comparison with clonal cherry rootstocks, presently grown in the Czech Republic using artificial freezing of the budwood applied just after the phase of deep dormancy. With a temperature drop to freezing, both the mean and the maximum rate of general frost injury was markedly increased. The greatest damage of the tested material (roughly at the level of LD 50 on the average) occurred after the application of combined low temperatures -25°C for 4 hours + -20°C for 66 hours. The results of laboratory tests were compared with the damage of natural frost that occurred during the first half of January, 2002. The single observed years did not differ from each other in the extent of injury, on the average. From standard cherry rootstocks, P-HL-B was generally the most sensitive to freeze injury. Its weak winter hardiness was approximately the same, or even somewhat worse, than that of rootstock Colt. The rootstock P-HL-C was classified as medium sensitive to winter frosts, while rootstock P-HL-A was scored as winter hardy. The average frost injury score of all 48 selected Cerapadus × Prunus avium L. genotypes included in the study was 6.7, whereas that of all control cherry rootstocks was only 5.9. According to the results of this study, the tested genotypes were classified into 5 groups with different classes of frost resistance or susceptibility. The most frequent was the class of medium resistance. The following five genotypes were the most winter hardy: CPH VODÁRNA, CPH 43, CPH 17, CPH 22 and CPH 49. On the basis of the obtained results, suggestions for improving testing procedures are also given. cut from mother trees during the first half of January. This term was chosen using the experience of MITTEL- STÄDT and WOLFRAM (1996), who had tested similar clonal material in similar climatic conditions. The term following the deep dormancy phase was convenient, because the treated material could easily be used for subsequent growing tests.After storing for a short time in a cold store, the tested shoots were treated by a freeze temperature in a laboratory freeze chamber, with a possibility of temperature control and measuring accuracy of 0.1°C. The treated temperature was reached within two hours. At first, several different freezing temperatures were applied for different durations of time and were verified in 2000, from which three of the most suitable treatments were chosen and used for this three year study: -15°C for 22 hours (D), -20°C for 22 hours (E) and -25°C for 4 hours plus -20 o C for 66 hours (F). In January, 2002 a natural frost at Holovousy occurred about one week before the budwood was cut for testing. The deepest natural freezing outdoors took place from January 4 th till January 6 th with temperature fluctuating between -15°C to -18°C. After a freeze treatment, the tested shoots were, together with the untreated controls, placed into bulbs with water and kept at r...