In section 5, we prove the result mentioned above, in the context of abstract root systems. It is independent from the rest of the paper.
I am indebted to my advisor Robert Kottwitz for suggesting the problem and many useful comments. I also thank Christian Kaiser for some helpful conversations.
This project was carried out during a stay at the MaxPlanckInstitut für Mathematik in Bonn, and I am grateful to the Institut for its support and hospitality.
2
$L$
chambers and
$\mathcal{P}$
chambers
Let
$G$
be a connected reductive group over
$\mathbb{R}$
, and
$T$
a maximal torus of
$G$
. Assume that
$G$
has a discrete series, or equivalently, that
$G$
has an elliptic maximal torus.
Write
${T}_{c}$
, respectively
$A$
, for the maximal compact, resp. split, subtori of
$T$
with centralizers
$L$
, resp.
$M$
, in
$G$
. Write
$R$
for the root system of
$T$
in
$G$
, and
${R}_{L}$
, resp.
${R}_{M}$
, for the set of roots of
$T$
in
$L$
, resp.
$M$
. Then
${R}_{L}$
is the subset of
$R$
consisting of real roots, and
${R}_{M}$
is the subset of imaginary roots. Write
${W}_{L}$
and
${W}_{M}$
for the respective Weyl groups. They are commuting subgroups of
$W$
with trivial intersection. Note that
${W}_{L}$
fixes each root in
${R}_{M}$
.
$A$
is contained as a split maximal torus in
${L}_{der}$
, the derived group of
$L$
, and we may identify
${R}_{L}$
with the set of roots of
$A$
in
${L}_{der}$
.
Write
${\mathfrak{a}}_{M}$
for
${X}_{*}\left(A\right){\otimes}_{\mathbb{Z}}\mathbb{R}$
. For any
$\alpha \in R\backslash {R}_{M}$
the root hyperplane
${H}_{\alpha}$
of
${X}^{*}(T{)}_{\mathbb{R}}:={X}_{*}(T)\otimes \mathbb{R}$
gives a hyperplane in
${\mathfrak{a}}_{M}$
. Let us consider two kinds of chambers in
${\mathfrak{a}}_{M}$
obtained from these. Define
$\mathcal{P}$
chambers to be those obtained by deleting from
${\mathfrak{a}}_{M}$
all the hyperplanes
${H}_{\alpha}$
, with
$\alpha \in R\backslash {R}_{M}$
. Define
$L$
chambers to be those obtained by deleting all the
${H}_{\alpha}$
with
$\alpha \in {R}_{L}$
. The latter are the Weyl chambers for
$A$
in
${L}_{der}$
; therefore
${W}_{L}$
acts simply transitively on them.
Observe that
${R}_{L}\subset (R\backslash {R}_{M})$
. Any additional hyperplanes coming from roots in
$R\backslash ({R}_{L}\cup {R}_{M})$
divide the
$L$
chambers into
$\mathcal{P}$
chambers.
Thus every
$\mathcal{P}$
chamber is contained in a unique
$L$
chamber.
Write
$\mathcal{P}\left(M\right)$
for the set of parabolic subgroups of
$G$
admitting
$M$
as a Levi component. There is a onetoone correspondence between
$\mathcal{P}\left(M\right)$
and the set of
$\mathcal{P}$
chambers in
${\mathfrak{a}}_{M}$
, obtained as follows: for
$P=MN\in \mathcal{P}\left(M\right)$
, the corresponding
$\mathcal{P}$
chamber is
$${\mathfrak{a}}_{P}^{+}=\{x\in {\mathfrak{a}}_{M}:\langle \alpha ,x\rangle >0\text{, for all}\alpha \in {R}_{N}\},$$
where
${R}_{N}$
denotes the set of roots of
$T$
in
$Lie\left(N\right)$
.
Recall that the set of
$L$
chambers is in bijection with the set of Borel subgroups of
$L$
containing
$T$
, or equivalently the set of positive root systems
${R}_{L}^{+}$
in the root system
${R}_{L}$
.
Now let
${C}_{P}$
be a
$\mathcal{P}$
chamber, and let
$P=MN$
be the corresponding element of
$\mathcal{P}\left(M\right)$
. It is easy to see that
${R}_{N}\cap {R}_{L}$
is a positive system in
${R}_{L}$
, and this corresponds to an
$L$
chamber
${C}_{L}$
. Thus we have defined a map
${C}_{P}\mapsto {C}_{L}$
from the set of
$\mathcal{P}$
chambers to the set of
$L$
chambers.
It is the obvious one which associates to
${C}_{P}$
the unique
$L$
chamber containing
${C}_{P}$
.
3 A Linear Combination of Characters
A stable virtual character is a finite
$\mathbb{Z}$
linear combination
$\Theta $
of characters
${\Theta}_{\pi}$
so that
$$\Theta \left(\gamma \right)=\Theta \left({\gamma}^{\prime}\right)$$
whenever
$\gamma $
and
${\gamma}^{\prime}$
are regular, stably conjugate elements of
$G(\mathbb{R})$
.
In Lemma 4.1 of [2], it is proved that for a stable virtual character
$\Theta $
on
$G(\mathbb{R})$
, the function
$$\gamma \mapsto \left{D}_{M}^{G}\right(\gamma ){}^{\frac{1}{2}}\Theta (\gamma )$$
on
${T}_{reg}(\mathbb{R})$
extends continuously to
$T(\mathbb{R})$
. A key ingredient of the proof is the fact that the expression at the bottom of page 497 is a linear combination of irreducible finitedimensional representations of
$M$
. In this section we will compute explicitly the coefficients and the representations involved, in the case where the element
$a$
appearing in the proof is equal to
$1$
.
We translate the setup of the proof in [2] as follows. We take
$\Gamma $
to be the identity component of
$T(\mathbb{R})$
. The root system
${R}_{\Gamma}$
is then simply
${R}_{L}$
. Fix a positive root system
${R}_{L}^{+}$
in
${R}_{L}$
, and let
$C$
be the corresponding
$L$
chamber in
${\mathfrak{a}}_{M}$
. We then choose a parabolic subgroup
$P=MN$
so that
${R}_{L}\cap {R}_{N}\subseteq {R}_{L}^{+}$
. Note that
${R}_{L}\cap {R}_{N}$
is also a system of positive roots, so this condition is equivalent to having
${R}_{L}\cap {R}_{N}={R}_{L}^{+}$
.
Thus we simply require that the
$\mathcal{P}$
chamber corresponding to
$P$
be contained in
$C$
.
Although at the end of our computations we will allow
$\gamma $
to be nonregular, we choose now
$\gamma $
to be a regular element of
$\Gamma ={T}_{c}(\mathbb{R})\cdot exp\left(\overline{C}\right)$
.
The expression is
$$\begin{array}{c}{\sum}_{B}m\left(B\right)\frac{{\Delta}_{P}\left(\gamma \right)\cdot {\lambda}_{B}\left(\gamma \right)}{{\Delta}_{B}\left(\gamma \right)}.\end{array}$$ 
(1)

The sum ranges over Borels containing
$T$
, which correspond to elements of
$W$
.
Here
${\lambda}_{B}$
is the
$B$
dominant highest weight of
$E$
,
$${\Delta}_{B}={\prod}_{\alpha >0}(1{\alpha}^{1})\text{, and}{\Delta}_{P}={\prod}_{\alpha \in {R}_{N}}(1{\alpha}^{1}).$$
Fix now a Borel
$B$
of
$G$
with
$T\subseteq B\subseteq P$
, for the rest of this paper. Recall the set of Kostant representatives
${W}^{M}$
for the Weyl group
${W}_{M}$
of
$M$
, relative to B. It is the set
$\{w\in W{w}^{1}{R}_{M}^{+}\subset {R}^{+}\}$
.
If
$w\in W$
, write
$w*B$
for
$wB{w}^{1}$
.
We will use the observation that for
$\omega \in {W}^{M},(\omega *B{)}_{M}={B}_{M}$
.
Indeed, if
$\alpha \in {R}^{+}\cap {R}_{M}$
, then
${\omega}^{1}\alpha \in {R}^{+}$
, which implies that
$\alpha \in \omega {R}^{+}\cap {R}_{M}$
.
Our sum (1) breaks up as
$$\begin{array}{c}{\sum}_{\omega \in {W}^{M}}m(\omega *B)\cdot {\Delta}_{P}\left(\gamma \right)\cdot {\sum}_{{w}_{M}\in {W}_{M}}\frac{{w}_{M}\left(\omega {\lambda}_{B}\right)\left(\gamma \right)}{{\Delta}_{{w}_{M}\omega *B}\left(\gamma \right)}.\end{array}$$ 
(2)

We would prefer the denominator inside the sum to be
${\Delta}_{{w}_{M}*{B}_{M}}\left(\gamma \right)$
.
(Recall that
${B}_{M}=B\cap M$
.) Note that
${\Delta}_{P}\cdot {\Delta}_{{B}_{M}}={\Delta}_{B}$
, since
${R}^{+}$
is the disjoint union of
${R}_{M}^{+}$
and
${R}_{N}$
.
So we consider the quantity
$$\begin{array}{c}\frac{{\Delta}_{P}\cdot {\Delta}_{{w}_{M}*{B}_{M}}}{{\Delta}_{{w}_{M}\omega *B}}=\frac{{\Delta}_{B}\cdot {\Delta}_{{w}_{M}*{B}_{M}}}{{\Delta}_{{B}_{M}}\cdot {\Delta}_{{w}_{M}\omega *B}}.\end{array}$$ 
(3)

Observe that if
$\mathcal{\mathcal{B}}$
is a Borel,
${\Delta}_{\mathcal{\mathcal{B}}}={\delta}_{\mathcal{\mathcal{B}}}\cdot {\rho}_{\mathcal{\mathcal{B}}}^{1}$
, where
${\delta}_{\mathcal{\mathcal{B}}}={\prod}_{\alpha >0}({\alpha}^{\frac{1}{2}}{\alpha}^{\frac{1}{2}})$
and
${\rho}_{\mathcal{\mathcal{B}}}$
is the usual half sum of positive roots. Since
${\delta}_{w*\mathcal{\mathcal{B}}}=\varepsilon \left(w\right){\delta}_{\mathcal{\mathcal{B}}}$
, we compute that
$$\frac{{\Delta}_{w*\mathcal{\mathcal{B}}}}{{\Delta}_{\mathcal{\mathcal{B}}}}=\varepsilon \left(w\right)\cdot ({\rho}_{\mathcal{\mathcal{B}}}w{\rho}_{\mathcal{\mathcal{B}}}).$$
Thus 3 becomes
$$\varepsilon \left(\omega \right)\left({w}_{M}\right(\omega {\rho}_{B}{\rho}_{{B}_{M}}){\rho}_{B}+{\rho}_{{B}_{M}}).$$
Next observe that for
${w}_{M}\in {W}_{M}$
,
$${w}_{M}({\rho}_{B}{\rho}_{{B}_{M}})={\rho}_{B}{\rho}_{{B}_{M}}.$$
Indeed, the roots of
${R}^{+}$
not in
${R}_{M}^{+}$
are in
${R}_{N}$
, and are thus normalized by
${W}_{M}$
. So the above expression simplifies to
$$\varepsilon \left(\omega \right)\cdot {w}_{M}(\omega {\rho}_{B}{\rho}_{B}).$$
We can therefore rewrite 2 as
$$\begin{array}{c}{\sum}_{\omega \in {W}^{M}}m(\omega *B)\cdot \varepsilon \left(\omega \right)\cdot {\sum}_{{w}_{M}\in {W}_{M}}\frac{{w}_{M}\left(\omega \right({\lambda}_{B}+{\rho}_{B}){\rho}_{B})\left(\gamma \right)}{{\Delta}_{{w}_{M}*{B}_{M}}\left(\gamma \right)}.\end{array}$$ 
(4)

Since
$\omega $
is a Kostant representative, the weight
$\omega ({\lambda}_{B}+{\rho}_{B}){\rho}_{B}$
is positive for
${B}_{M}$
, and we may use the Weyl character formula to rewrite this as
$$\begin{array}{c}{\sum}_{\omega \in {W}^{M}}m(\omega *B)\cdot \varepsilon \left(\omega \right)\cdot tr(\gamma ;{V}_{\omega ({\lambda}_{B}+{\rho}_{B}){\rho}_{B}}^{M}).\end{array}$$ 
(5)

Here
${V}_{\omega ({\lambda}_{B}+{\rho}_{B}){\rho}_{B}}^{M}$
denotes the irreducible finitedimensional representation of
$M$
with highest weight
$\omega ({\lambda}_{B}+{\rho}_{B}){\rho}_{B}$
.
4 A Formula for
${\Phi}_{M}(\gamma ,{\Theta}^{E})$
To identify 5 with
${\Phi}_{M}(\gamma ,{\Theta}^{E})$
, we replace
$m(\omega *B)$
with
$n(\gamma ,\omega *B)$
as on page 500 of [2], and multiply it by the factor
${\delta}_{P}^{\frac{1}{2}}\left(\gamma \right)$
:
$$\begin{array}{c}{\delta}_{P}^{\frac{1}{2}}\left(\gamma \right)\cdot {\sum}_{\omega \in {W}^{M}}n(\gamma ,\omega *B)\cdot \varepsilon \left(\omega \right)\cdot tr(\gamma ;{V}_{\omega ({\lambda}_{B}+{\rho}_{B}){\rho}_{B}}^{M}).\end{array}$$ 
(6)

Here
${\delta}_{P}$
is the modulus character of
$P$
. (We are still only considering regular
$\gamma $
.) Write
${A}_{G}$
for the split component of the center of
$G$
. Let
${\lambda}_{0}\in {X}^{*}\left({A}_{G}\right)$
denote the character by which
${A}_{G}$
acts on
$E$
. It extends to
${X}^{*}(T{)}_{\mathbb{R}}$
in the usual way, and is
$W$
invariant.
Let
${T}_{e}$
denote the subtorus of
$T$
generated by
${T}_{c}$
and
${A}_{G}$
. It is the maximal subtorus of
$T$
which is elliptic in
$G$
.
Write
${p}_{M}$
for the projection from
${X}^{*}(T{)}_{\mathbb{R}}$
to
${X}^{*}(A{)}_{\mathbb{R}}$
, and note that it is
${W}_{L}$
invariant. The group
${W}_{L}$
fixes each root of
$M$
, thus it acts on
${W}^{M}$
. For every orbit of this action, there is a unique member
$\omega $
so that
${p}_{M}\left(\omega \right({\lambda}_{B}+{\rho}_{B}{\lambda}_{0}\left)\right)$
is dominant with respect to
$C$
. We denote the set of these elements by
${W}^{LM}$
, one element for each orbit of
${W}_{L}$
on
${W}_{M}$
.
If
$\lambda \in {X}^{*}\left(T\right)$
and
${w}_{L}\in {W}_{L}$
, then plainly
${w}_{L}\lambda \lambda \in {\mathfrak{a}}_{M}^{*}$
. Write
$({\chi}_{{w}_{L},\omega ,B},{\mathbb{C}}_{{w}_{L},\omega ,B})$
for the onedimensional representation of
$M$
, acting through
$A$
, with weight
${w}_{L}\omega ({\lambda}_{B}+{\rho}_{B})\omega ({\lambda}_{B}+{\rho}_{B})$
. Note that
${T}_{c}$
and
${A}_{G}$
act trivially on
${\mathbb{C}}_{{w}_{L},\omega ,B}$
, thus so does
${T}_{e}$
.
Thus we have
$${V}_{{w}_{L}\omega ({\lambda}_{B}+{\rho}_{B}){\rho}_{B}}^{M}\sim ={V}_{\omega ({\lambda}_{B}+{\rho}_{B}){\rho}_{B}}^{M}\otimes {\mathbb{C}}_{{w}_{L},\omega ,B}.$$
Our formula 6 is now (replacing
$\omega \in {W}^{M}$
with
${w}_{L}\omega $
, where
$\omega $
is now in
${W}^{LM}$
):
$$\begin{array}{c}{\delta}_{P}^{\frac{1}{2}}\left(\gamma \right)\cdot {\sum}_{\omega \in {W}^{LM}}\varepsilon \left(\omega \right)\cdot tr(\gamma ;{V}_{\omega ({\lambda}_{B}+{\rho}_{B}){\rho}_{B}}^{M})\cdot {\sum}_{{w}_{L}\in {W}_{L}}\varepsilon \left({w}_{L}\right)\cdot {\chi}_{{w}_{L},\omega ,B}\left(\gamma \right)\cdot n(\gamma ,{w}_{L}\omega *B),\end{array}$$ 
(7)

Of course we now wish to simplify the inner sum. Recall from page 500 of [2] that
$$n(\gamma ,{w}_{L}\omega *B)=\overline{c}(x,{p}_{M}({w}_{L}\omega {\lambda}_{B}+{w}_{L}\omega {\rho}_{B}{\lambda}_{0}\left)\right),$$
where
$x$
is in the interior of
$C$
. Here
$\overline{c}(x,\lambda )$
is the integervalued “stable discrete series constant” on
$$\left({X}_{*}\right(A/{A}_{G}{)}_{\mathbb{R}}{)}_{reg}\times \left({X}^{*}\right(A/{A}_{G}{)}_{\mathbb{R}}{)}_{reg},$$
as defined, for instance, on page 493 of [2]. Recall that
${\lambda}_{0}\in {X}^{*}(T{)}_{\mathbb{R}}$
is obtained from the character
${\lambda}_{0}\in {X}^{*}\left({A}_{G}\right)$
by which
${A}_{G}$
acts on
$E$
, and is thus
$W$
invariant.
As
${p}_{M}$
commutes with
${w}_{L}$
, the inner sum of 7 is now
$$\begin{array}{c}{\sum}_{{w}_{L}\in {W}_{L}}\varepsilon \left({w}_{L}\right)\cdot \overline{c}(x,{w}_{L}\Lambda )\cdot {\chi}_{{w}_{L},\omega ,B}\left(\gamma \right),\end{array}$$ 
(8)

where
$\Lambda ={p}_{M}(\omega {\lambda}_{B}+\omega {\rho}_{B}{\lambda}_{0})$
.
We would like to consider the limit of
8 as
$x$
approaches
$0$
. Recall we can write
$\gamma ={\gamma}_{c}\cdot exp\left(x\right)$
, with
${\gamma}_{c}\in {T}_{c}(\mathbb{R})$
and
$x$
in
$\overline{C}$
. Also recall that
$\gamma $
is still regular (not for long!). Consider the above formula with
${\gamma}_{c}$
fixed and
$x$
going to
$0$
along regular elements of
$\overline{C}$
. Fix some element
${x}_{0}$
in the interior of
$C$
. The value
$$\overline{c}(x,{w}_{L}\Lambda )=\overline{c}({x}_{0},{w}_{L}\Lambda )$$
is unchanged, but
${\chi}_{{w}_{L},\omega ,B}\left(\gamma \right)$
approaches
${\chi}_{{w}_{L},\omega ,B}\left({\gamma}_{c}\right)=1$
.
Thus
8 converges to
$${\sum}_{{w}_{L}\in {W}_{L}}\varepsilon \left({w}_{L}\right)\cdot \overline{c}({x}_{0},{w}_{L}\Lambda )$$
for some
${x}_{0}\in C$
.
But this is simply
$(1{)}^{q\left(L\right)}{W}_{L}$
, by Proposition 1(ii) in Section 5 below. Here we use that
$\omega \in {W}^{LM}$
. Note that
$1$
is in the Weyl group of the root system by the argument on page 499 of [2].
It is easy to modify this argument to get the same limit as
$x$
approaches an element of
${X}_{*}({A}_{G}{)}_{\mathbb{R}}$
.
Finally note that
${\delta}_{P}$
is a positive character and therefore trivial on the compact group
${T}_{c}(\mathbb{R})$
. It is thus trivial on
${T}_{e}(\mathbb{R})$
.
Now consider irregular
$\gamma $
. We take the limit in 7 and obtain our theorem:
Theorem 1.
If
$\gamma \in {T}_{e}(\mathbb{R})$
, then
$$\begin{array}{c}{\Phi}_{M}(\gamma ,{\Theta}^{E})=(1{)}^{q\left(L\right)}\cdot {W}_{L}\cdot {\sum}_{w\in {W}^{LM}}\varepsilon (w)\cdot tr(\gamma ;{V}_{w({\lambda}_{B}+{\rho}_{B}){\rho}_{B}}^{M}).\end{array}$$ 
(9)

For the reader's convenience, we review the definition of
${W}^{LM}$
.
The definition depends on the choice of a parabolic
$P=MN$
and a Borel subgroup
$B$
with
$T\subseteq B\subseteq P$
. The choice of
$B$
gives a set of positive roots
${R}^{+}$
for
$R$
and a set of positive roots
${R}_{M}^{+}$
for
${R}_{M}$
. It also gives
$B$
dominant elements
${\lambda}_{B}$
and
${\rho}_{B}$
of
${X}^{*}(T{)}_{\mathbb{R}}$
. The choice of
$P$
determines an
$L$
chamber
$C$
as in Section 2. Recall the character
${\lambda}_{0}$
determined by
${A}_{G}$
on
$E$
and the projection
${p}_{M}$
from
${X}^{*}(T{)}_{\mathbb{R}}$
to
${X}^{*}(A{)}_{\mathbb{R}}$
. Then
$${W}^{LM}=\{w\in W{w}^{1}{R}_{M}^{+}\subseteq {R}^{+}\text{and}{p}_{M}\left(w\right({\lambda}_{B}+{\rho}_{B}{\lambda}_{0}\left)\right)\text{is dominant w.r.t.}C\}.$$
We now evaluate 9 for
${\Phi}_{M}$
on the extreme cases for
$T$
. If
$T=A$
is split, then
$M=A$
,
$L=G$
,
${W}^{LM}$
is trivial, but so is
${T}_{c}$
. We conclude that for
$z\in {A}_{G}(\mathbb{R})$
,
$${\Phi}_{A}(z,{\Theta}^{E})=(1{)}^{q\left(G\right)}\cdot W\cdot {\lambda}_{0}(z).$$
If
$T$
is elliptic, then
$M=G$
,
$L=T$
,
${W}^{LM}$
is again trivial, and so for
$\gamma \in T$
,
$${\Phi}_{G}(\gamma ,{\Theta}^{E})=tr(\gamma ;E).$$
Note that this agrees with the results of Theorems 5.1 and 5.2 of [2], since
$$tr({\gamma}^{1};{E}^{*})=tr(\gamma ;E).$$
5 The Sum of the Stable Discrete Series Constants
Let
$(X,{X}^{*},R,\stackrel{\u02c7}{R})$
be a root system. Write
$W$
for the Weyl group of the root system, and
$\varepsilon $
for its sign character. Assume that
$R$
generates the real vector space
$X$
and that
$1\in W$
. Write
$q\left(R\right)$
for
$\left(\right{R}^{+}+dim(X\left)\right)/2$
, as in [2]. Let
${x}_{0}$
be a regular element of
$X$
, and
$\lambda $
a regular element of
${X}^{*}$
. Write
${C}_{0}$
for the chamber of
$X$
containing
${x}_{0}$
, and
${C}_{0}^{\vee}$
for its dual chamber in
${X}^{*}$
. Recall the stable discrete series constants
${\overline{c}}_{R}({x}_{0},\lambda )$
from section 3 of [2].
Proposition 1.
We have the following formulas for sums of discrete series constants:

(i)
For all such
$\lambda ,{\sum}_{w\in W}{\overline{c}}_{R}(w{x}_{0},\lambda )=\leftW\right.$

(ii)
For
$\lambda ={\lambda}_{0}\in {C}_{0}^{\vee}$
, we have
${\sum}_{w\in W}\varepsilon \left(w\right)\cdot {\overline{c}}_{R}(w{x}_{0},{\lambda}_{0})=(1{)}^{q\left(R\right)}W.$
The same formulas hold if we sum over the
$W$
orbit of
$\lambda $
rather than that of
${x}_{0}$
.
We make a few comments before beginning the proof. The proof begins by using the “inductive” property (4) of the discrete series constants from page 493 of [2], to change the sum over chambers into a sum over certain facets of
$X$
. In fact we consider those facets which separate the chambers of
$X$
, i.e., those which span the root hyperplanes
$Y$
of
$X$
.
In the course of the proof, we (mis)use the term “facet” only in reference to these particular facets, of codimension
$1$
. So a facet in this sense will be the common face of two adjacent chambers.
The hyperplanes
$Y$
have their own chambers, and we examine the relationship between the facets and these smaller chambers. Not every facet is equal to such a chamber, as in the case of
${B}_{3}$
when
$Y$
is the root hyperplane of a long root. The facets in
$Y$
give a
${B}_{2}$
system, but the chambers of
${R}_{Y}$
give an
${A}_{1}\times {A}_{1}$
system.
Finally induction on the rank of the root system gives the calculation.

Proof.
The second formula follows from the first by applying Theorem 3.2(2) on page 494 of [2].
We induce on
$r=dimX$
. The proposition is clear when
$r=0$
.
We associate these discrete series constants with the various chambers and facets of
$X$
, and introduce some appropriate notation.
Write
$c(\mathcal{C})$
for
${\overline{c}}_{R}(x,\lambda )$
, when
$x$
is in the interior of a chamber
$\mathcal{C}$
. Suppose
$F$
is a facet in
$X$
,
$y$
is in the interior of
$F$
, and
$\overline{F}:=span\left(F\right)=Y$
. Then write
$c\left(F\right)={\overline{c}}_{{R}_{Y}}(y,{\lambda}_{Y})$
, notation as on page 493 of [2].
Thus if
$F$
is the common face of distinct chambers
$\mathcal{C}$
and
${\mathcal{C}}^{\prime}$
, then
$$2c\left(F\right)=c(\mathcal{C})+c\left({\mathcal{C}}^{\prime}\right).$$
Each chamber has
$r$
faces, and it follows that
$$\begin{array}{c}r\cdot {\sum}_{\mathcal{C}}c(\mathcal{C})=2{\sum}_{F}c\left(F\right),\end{array}$$ 
(10)

where we are summing over all chambers and then all facets.
We show the right hand side of
10 is equal to
$r\cdot \leftW\right$
to prove the proposition.
Now every facet is on some root hyperplane
${X}_{\alpha}={X}_{\alpha}$
, so we have
$$2{\sum}_{F}c\left(F\right)={\sum}_{\alpha \in R}{\sum}_{\overline{F}={X}_{\alpha}}c\left(F\right).$$
We now work with the inner sum. There is a root system on
${X}_{\alpha}$
whose set of coroots is
$\stackrel{\u02c7}{R}\cap {X}_{\alpha}$
, which defines chambers
${\mathcal{C}}_{\alpha}$
in
${X}_{\alpha}$
and constants
${c}_{\alpha}\left({\mathcal{C}}_{\alpha}\right)$
. Write
${W}_{\alpha}$
for the Weyl group of
${X}_{\alpha}$
. We have
$${\sum}_{\overline{F}={X}_{\alpha}}c\left(F\right)={\sum}_{{\mathcal{C}}_{\alpha}}{\sum}_{F\subset {\mathcal{C}}_{\alpha}}{c}_{\alpha}\left({\mathcal{C}}_{\alpha}\right)={\sum}_{{\mathcal{C}}_{\alpha}}{\sum}_{{W}_{\alpha}\backslash \{F\subset {X}_{\alpha}\}}{c}_{\alpha}\left({\mathcal{C}}_{\alpha}\right)={\sum}_{{W}_{\alpha}\backslash \{F\subset {X}_{\alpha}\}}{\sum}_{{\mathcal{C}}_{\alpha}}{c}_{\alpha}\left({\mathcal{C}}_{\alpha}\right).$$
For the first equality, note that every facet
$F$
with
$\overline{F}={X}_{\alpha}$
is contained in a some chamber
${\mathcal{C}}_{\alpha}$
.
The second equality follows because
${W}_{\alpha}$
acts transitively on the chambers
${C}_{\alpha}$
.
Write
$n\left(\alpha \right)$
for the order of
${W}_{\alpha}\backslash \{F\subset {X}_{\alpha}\}$
. It is equal to the number of facets in a given chamber
${\mathcal{C}}_{\alpha}$
. Then by induction the above is merely
$$n\left(\alpha \right)\cdot \left{W}_{\alpha}\right,$$
which is exactly the number of facets in
${X}_{\alpha}$
. It follows that 10 is simply equal to twice the total number of facets in
$X$
.
Since
$W$
has
$r$
orbits on the set of facets in
$X$
, and the stabilizer in
$W$
of any facet has order
$2$
, we conclude that the total number of facets is half of
$r\cdot \leftW\right$
, as desired. □
References

J. Arthur, The
${L}^{2}$
Lefschetz numbers of Hecke operators, Invent. Math. 97 (1989), 257290.

M. Goresky, R. Kottwitz, and R. Macpherson, Discrete Series Characters and the Lefschetz Formula for Hecke operators, Duke Math. Journal, Vol. 89, No. 3 (1997),477554.
MaxPlanckInstitut fur Mathematik, Bonn, Germany Email address : spallone@mpigbonn.mpg.de