Sylow Notes
Sylow Notes
Proposition 1 Let q = pr ; let M Tn (Fq ) be the group of upper triangular matrices with 1s
along the diagonal. Then M Tn (Fq ) is a Sylow p-subgroup of GLn (Fq ).
Proof: Its clear that M Tn (Fq ) is a subgroup of GLn (Fq ), so it suffices to show that M Tn (Fq )
has order q n(n 1)/2 . This follows simply; each of the n(n 1)/2 entries strictly above the
diagonal can be any element of Fq , for a total of q n(n 1)/2 elements.
Im not going to prove this proposition, but you can find a proof in Artins Algebra on
pages 115 and 116, or in many other places, Im sure. This proposition tells us that A and
(A) are conjugate (or in the language of linear algebra, similar ), for any change-of-basis
isomorphism . Thus, the Sylow p-subgroups M Tn (Fq ) and (M Tn (Fq )) are conjugate as
well. As it turns out, all of the Sylow p-subgroups of a group G are conjugate; this is Sylows
second theorem.
Theorem 2 (Second Sylow Theorem) The Sylow p-subgroups of a group G are conju-
gate.
Theorem 3 (Third Sylow Theorem) Let s be the number of Sylow p-subgroups of G; let
|G| = pa m where p does not divide m. Then s divides m and s is congruent to 1 (mod p).
Weve seen already that we get one Sylow p-subgroup of GLn (Fq ) by taking the group of
upper triangular matrices with 1s along the diagonal. In similar fashion, the group of lower
triangular matrices with 1s along the diagonal is a Sylow p-subgroup. Since for n 2 these
two groups arent the same (and when n = 1, p doesnt divide the order of GLn (Fq )), the
number of Sylow p-subgroups of GLn (Fq ) is greater than 1, so it is at least p + 1. We notice
that p + 1 divides q 2 1, so it is an allowable number of Sylow p-subgroups. Unfortunately,
even for small q and n, GLn (Fq ) is large and there are a lot of choices of s that fulfill the
requirements of the third Sylow theorem. In principle, it shouldnt be difficult to find s; we
know one easily described Sylow p-subgroup, and we just need to conjugate it to find all
the others. This is tedious, but I dont know of a better way to count the Sylow p-subgroups.