Abstract Petal abscission was studied in twelve hybrid tea rose (Rosa hybrida L.) cultivars. At about 20 °C the time to petal abscission in uncut stems in greenhouses was the same as in cut stems placed in water in the greenhouse or in a climate-controlled room. The time between petal unfolding and abscission depended on the cultivar, and varied between 12 and 35 d. The time to petal abscission of the cultivars was inversely correlated with their flower diameter at full bloom (linear regression, r2 = 0·82). In the cultivars with a relatively large flower diameter (10-18 cm) the petals fell without visible desiccation symptoms, whereas in the group with a small diameter the petals were partially or fully desiccated when shed. Fertilization occurred in some flowers of a few cultivars studied. In cultivars with a relatively large flower diameter (Papa Meilland, Cocktail, Dr. Verhage, Tineke) it had no effect on the time to abscission in Motrea, Europa, and Carolien roses, which bear small flowers, the petals fell after fertilization, whereas in unfertilized flowers of the latter group of cultivars an abscission zone just above the uppermost node became active and all parts above this node (pedicel and flower) turned brown and desiccated, though remained attached for more than a month. It is concluded that in the cultivars investigated: (a) the time to petal abscission was inversely related to their flower diameter, (b) abscised petals were more desiccated in cultivars in which the time to abscission was longer, (c) fertilization had little effect on the time to abscission in most cultivars, whereas the absence of fertilization prevented petal abscission in a number of the small-diameter cultivars where it was replaced by flower abscission, and (d) cutting and placement in water at 20 °C did not affect the time to abscission.