Let γ be a linear combination of roots with integer coefficients, and α an indivisible root. If γ is proportional to α, then γ∈Zα.
By Prop. 15 of no. 5, C can be chosen so that α∈B(C). By Th. 3,
γ=β∈B(C)∑nββwithnβ∈Z. Thus, if
γ is proportional to
α, then
γ=nαα, which proves the corollary.
Now let S be the set of reflections sα for α∈B(C) and let T be the union of the conjugates of S under W. For α∈B(C) and w∈W, the element t=wsαw−1 of T is the orthogonal reflection sβ associated to the root β=w(α); conversely, for any indivisible root β, there exists an element w∈W such that α=w−1(β)∈B(C) (Prop. 15) and sβ=wsαw−1∈T. It follows that a bijection ψ from the set of indivisible roots to {±1}×T is obtained by associating to an indivisible root β the pair (ε,sβ), where ε=+1 if β is positive and ε=−1 if β is negative.
On the other hand, (W,S) is a Coxeter system (Th. 2) and the results of Chap. IV, § 1, no. 4 can be applied. We have seen that, if w is an element of W of length (with respect to S) equal to q, there exists a subset Tw of T, with q elements, such that, if w=s1…sq with si∈S and if
ti=s1…si−1sisi−1…s1 (for
1≤i≤q), then
Tw={t1,…,tq}. Recall that we have also defined in no. 4 of § 1 a number
η(w,t) (for
w∈W and
t∈T) equal to
+1 if
t∈/Tw and to
−1 if
t∈Tw. Finally, recall that, if we define a map
Uw from the set
{±1}×T to itself by the formula
Uw(ε,t)=(εη(w−1,t),wtw−1), the map
w↦Uw is a
homomorphism from
W to the group of permutations of the set
{±1}×T (Chap. IV, § 1, no. 4, Lemma 1).
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.