In [GL] Graham and Lehrer have presented
a nice criterion for quasi-heridity of a cellular algebra
which we will now verify in our case.
This will prove Theorem
7.5.
To this aim we have to investigate the
bilinear form
on the
standard modules
. We must show that they are not zero
([GL, 3.10]). Let us first calculate
the Grammatrix of
with respect to
the basis
of
. We abbreviate its entries by
. According to the
definition in [GL, 2.3], these numbers satisfy
Such a congruence relation is valid with
being
independent of
and
by the axioms of a cellular algebra
(see [GL, 1.7]). Dualizing this congruence we obtain the following
counterpart in the cellular coalgebra
:
modulo
. According to the calculations
for the verification of axiom
in the previous section
we see using the notations from there that
The bilinear form
is different from zero if this is the case
for a single entry
. We calculate
where
is the
-tableau
with constant rows
for all
and
.
Obviously
is a standard tableau with respect to both orders on n, namely
as well as
. Furthermore the reverse symplectic condition holds
because
. Consequently we have
. Note that
does not contain any pair of associated indices
. The content
of
is given by
Consider the endomorphism
It is easy to see that commutes with
and
for all
. Consequently it is an endomorphism of the
comodul
(in fact it is the idempotent of
corresponding to the weightspace with weight
). There is an induced
action of
on
from the right defined by
For a bideterminant we have
Applying to the defining equation (30) of
we see that
if
by linear independence of
.
Since
is the only element in
having
content
it follows
if
and
we conclude
By [GL, Remark 3.10], this finishes the proof of Theorem 7.5.