Plants were drip irrigated. Irrigation was triggered when two of
three tensiometers gave readings below −8 and −26 kPa, when the
plants grew in the absence of and with WR, respectively. During the
WR period, the volume of solution applied at each irrigation was
limited to 20 ml per plant, to prevent the complete rehydration of
the substrate and the return of the plant to conditions in which
water was not limiting. This system made it possible to apply moderate
or severe WR to the plants for at least 95% of the WR period
(Table 2). From bud break to the FBV stage, i.e. for a mean of 15 days,
the young rose plants were grown without WR. At the FBV stage,
plants were randomly assigned to five treatment groups: (i) a control
group with no WR (WR0), (ii) four groups in which temporary
WR was applied for 7 days (WR7), 14 days (WR14), 21 days (WR21)
or 35 days (WR35). After the WR period, the plants were subirrigated
for 4 h and then grown withoutWR, as in theWR0 group, until
the secondary axes had reached their final length, when they had
stopped growing but no flower bud had appeared (blind shoots) or
when the flower had reached at least the petal color visible (PCV)
stage, i.e. a mean of 59 days after FBV of primary axis (Fig. 1A)