The fate of shoot meristems of the long day (LD) plant Silene coeli-rosa in culture was examined (complete, reverted or arrested flowers) to establish whether these different patterns were related to a particular stage of morphogenesis and the extent to which the fate of the pattern was regulated by either added plant growth regulators (PGRs) or changing the carbohydrate source in the medium. In particular, the frequency of reversion was measured to test the stability of the determined state for each whorl. The plants were given various inductive treatments (4–7 LD, 7 LD + 1 to 3 SD) and the apices were explained onto Murashige-Skoog medium supplemented with 3% sucrose (controls) ± IAA, ± kinetin, ± GA3 or onto the basal medium containing 1 or 3% sucrose, glucose or fructose or 7% sucrose. The apices were examined 12 weeks later. When the data were pooled from all inductive treatments, IAA resulted in more reversions, GA3 caused more arrested flowers while kinetin hardly affected the pattern of meristem fate compared with the controls. However, each PGR treatment did not perturb the pattern of organ formation for those apices that formed either arrested or complete flowers. The time for determination (days) of the earlier formed whorls (determination times for the controls in brackets): sepals (2), stamens 1–5 (3) and petals (3), was shortened by about a day in all PGR treatments whereas the corresponding times for the later formed whorls: stamens 6–10 (4) and carpels (4), were either lengthened to 5 days or unaffected. The response of the apices to the various sugars was simply a reflection of concentration. Hence, more complete flowers formed at 7 or 3% and more flowers were arrested at 1 % regardless of the sugar moiety. However, the frequency of reversion was similar on each of the media. Pooling all data from all treatments enabled a statistical analysis of the pattern of reversion and the pattern of arrest. Reversion was more common from apices which exhibited the later-formed whorls (stamens 6–10 and carpels) than from the earlier whorls. Moreover, the stronger the inductive treatment the less frequent was reversion. The most common stage of arrest was at the stamen 6–10 whorl and this was particularly so for the GA3 treatment. The data indicated that reversion could occur from any whorl, which suggests that determination of each whorl is independent of the next. This conclusion is underlined by the more frequent occurrence of reversion from the carpel whorl. However, the longer the inductive treatment the less likelihood of reversion; this suggests that in Silene, the floral stimulus is required continuously to stabilise the determined state of each whorl and to ensure smooth completion of floral morphogenesis.