The stem fructan content (FCstem) increased in W1
andW0 until 3000 and 3500 GDD, respectively (Fig. 7). Afterward a strong fructans reduction was detected in
both treatments until halving in early November. Except
for the two last harvests,W0 showed a much higher
FCstem compared to W1. However in the second part of
the cycleW1 showed a higher fructan accumulation rate
until equating W0 at about 3500GDD (Fig. 7). Therefore,
at the final harvest the stem fructan yields (FYstem)
of W0 and W1 did not vary significantly despite the
higher stem biomass of W1. Nonetheless, in both years
FYstem of the last but one harvest was three-fold compared
to that of the final harvest. The FCstem appeared
not significantly correlated with SDWbut it was highly
correlated with stem fructan length index (DPstem)
(r = 0.84**). The maximum DPstem were reached at
about 3400GDD (early September) in both treatments.
The late harvest negatively influenced DPstem only in
W0. DPstem of W0 were in fact 9.7 and 5.8 in the two
last harvests.
As for stems, tubers of W0 showed a significant
higher fructan content (FCtub) only in the first part
of the growing season. Thereafter, the increasing of
FCtub in W1 was higher than in W0 and no statistically
differences between treatments were found at the
final harvest although in 1999 the tuber fructan yield
(FYtub)was significant higher inW1. The tubers fructan
length index (DPtub) appeared significantly correlated
with DPstem (r = 0.72**), but slightly higher than this
(Fig. 7). According to other experiments (Incoll and
Neales, 1970) the highest DPtub was reached before the
maximum TDW (about 1800 GDD). Water restoration
did not increase DPtub but it delayed (about 700 GDD)
the reaching of the maximum values.