The Arithmetic of Quaternion Algebra
The Arithmetic of Quaternion Algebra
The Arithmetic of Quaternion Algebra
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
30 CHAPTER 2. QUATERNION ALGEBRA OVER A LOCAL FIELD<br />
Use N(O) to denote the normalizer <strong>of</strong> an Eichler order O <strong>of</strong> M(2, K) in<br />
GL(2, K). By definition N(O) = {x ∈ GL(2, K)|xOx−1 = O}. Let O1, O2<br />
be the maximal orders containing O. <strong>The</strong> inner automorphism associating to<br />
an element <strong>of</strong> N(O) fixes the couple (O1, O2). <strong>The</strong> study <strong>of</strong> two-sided ideal<br />
<strong>of</strong> maximal order has showed that the two-sided ideal <strong>of</strong> a maximal order is<br />
generated by the nonzero elements <strong>of</strong> K. Hence it follows O = On �<br />
with<br />
�<br />
n ≥ 1.<br />
0 1<br />
. We can<br />
therefore we see that N(On) is generated by K × On × and<br />
π n 0<br />
verify without difficulty that the reduced discriminant <strong>of</strong> an Eichler order equals<br />
its level.<br />
<strong>The</strong> tree <strong>of</strong> maximal order<br />
Definition 2.8. (Serre [3], Kurihara [1]). A graph is given by<br />
– a set S(X) whose element is called a vertex <strong>of</strong> X,<br />
– a set Ar(X) whose element is called an edge <strong>of</strong> X,<br />
– a mapping: Ar(X) → S(X) × S(X) defined by y ↦→ (s, s ′ ) where s is called<br />
the origin <strong>of</strong> y and s ′ the extremity <strong>of</strong> y,<br />
– an involution <strong>of</strong> Ar(X) denoted by y ↦→ ¯y such that the origin <strong>of</strong> y to be the<br />
extremity <strong>of</strong> ¯y and such that<br />
(1) y �= ¯y<br />
.<br />
A chain <strong>of</strong> a graph X is a sequence <strong>of</strong> edges (y1, ..., yi+1...) such that the extremity<br />
<strong>of</strong> yi to be the origin <strong>of</strong> yi+1 for all i. To give a chain is equivalent to<br />
give a sequence <strong>of</strong> vertices such that two consecutive vertices to be always the<br />
origin and the extremity <strong>of</strong> an edge. A finite chain (y1, ..., yn) is called to have<br />
the length n, and we say it joins the origin <strong>of</strong> y1 and the extremity <strong>of</strong> yn. A pair<br />
(yi, ¯yi) in a chain is called a loop. A finite chain without any loop such that the<br />
origin <strong>of</strong> y1 to be the extremity <strong>of</strong> yn is called a circuit. A graph is connect if<br />
it always exists a chain joining two distinct vertices. A tree is a connect graph<br />
and without circuit.<br />
We see the set X <strong>of</strong> maximal orders <strong>of</strong> M(2, K) is provided with a structure<br />
<strong>of</strong> graph denoted still by X , such that these maximal orders are the vertices <strong>of</strong><br />
X and the pair (O, O ′ ) <strong>of</strong> maximal orders <strong>of</strong> distance 1 are the edges <strong>of</strong> X.<br />
Lemma 2.2.5. Let O be a maximal order. <strong>The</strong> maximal orders at distance n<br />
to O are the extremities <strong>of</strong> some chains which have no loops, their origins are<br />
O and the length are n.<br />
Pro<strong>of</strong>. Let O ′ be a maximal order such that d(O, O ′ ) = n. <strong>The</strong>refore, O =<br />
End(e1R + e2R) and O ′ = End(e1R + e2π n R), for an appropriate basis (e1, e2)<br />
<strong>of</strong> V . <strong>The</strong> sequence <strong>of</strong> vertices (O, O1, ..., Oi, ..., O ′ ), where Oi = End(e1R +<br />
e2π i R), 1 ≤ i ≤ n − 1, is a chain without loops, joining O and O ′ , <strong>of</strong> length n.<br />
Inversely, provided there is a chain <strong>of</strong> length n > 2 given by a sequence <strong>of</strong><br />
vertices (O0, ..., On). <strong>The</strong>re exist the R-lattices Li ⊃ Li+1 ⊃ Liπ such that<br />
Oi = End(Li) for 0 ≤ i ≤ n. the chain has no loops if Liπ �= Li+2 for<br />
0 ≤ i ≤ n − 2. WE have<br />
Li+1 ⊃ Liπ ⊃ Li+1π<br />
Li+1 ⊃ Li+2 ⊃ Li+1π