Frobenius result on simple groups of order (p^3-p)/2

The complete list of pairs of non-isomorphic finite simple groups having the same order is well-known. In particular for p>3, PSL_2(Z/p) is the “only” simple group of order (p^3-p)/2. It’s less well-known that Frobenius proved this uniqueness result in 1902. This note presents a version of Frobenius’ argument that might be used in an undergraduate honors algebra course. It also includes a short modern proof, aimed at the same audience, of the much earlier result that PSL_2(Z/p) is simple for p>3; a result stated by Galois in 1832.

💡 Research Summary

The paper revisits a little‑known result of Frobenius from 1902: for any prime (p>3) the only finite simple group of order ((p^{3}-p)/2) is (\mathrm{PSL}{2}(\mathbb{Z}/p)). The author presents Frobenius’s original argument in a form suitable for an undergraduate honors algebra course, and then supplies a modern, streamlined proof of the same classification together with a short proof that (\mathrm{PSL}{2}(\mathbb{Z}/p)) is itself simple for (p>3) (a fact originally asserted by Galois in 1832). The paper also sketches the generalisation that (\mathrm{PSL}_{2}(F)) is simple for any field (F) with more than three elements.

The exposition begins by recalling the standard description of (\mathrm{PSL}{2}(\mathbb{Z}/p)) as the group of projective linear fractional transformations on the set (\mathbb{Z}/p\cup{\infty}). The two basic maps (z\mapsto z+1) (translation) and (z\mapsto -1/z) generate the whole group, and the kernel of the action of (SL{2}(\mathbb{Z}/p)) is ({\pm I}).

The Classification Theorem is then stated: let (G) be a transitive permutation group on (\mathbb{Z}/p\cup{\infty}) containing all translations and of order ((p^{3}-p)/2). Either (a) the involution (z\mapsto -1/z) lies in (G) – in which case (G) is permutation‑isomorphic to (\mathrm{PSL}_{2}(\mathbb{Z}/p)) – or (b) (p=7) and (G) contains a specific involution with cycle structure ((0,\infty)(1,3)(2,6)(4,5)) or ((0,\infty)(1,5)(2,3)(4,6)); in this exceptional case (G) has a normal subgroup of order 8.

The proof proceeds by Sylow theory and 2‑transitivity. Because the translation subgroup is the unique Sylow‑(p) subgroup, it is normal in the stabiliser of (\infty); consequently the whole group is doubly transitive. The stabiliser of ({0,\infty}) splits into two cosets: (K), fixing both points, and (\overline K), swapping them. One shows that (K) consists exactly of the scalar maps (z\mapsto a z) with (a) a quadratic residue, so (|K|=(p-1)/2). The behaviour of (\overline K) depends on the parity of (p) modulo 4.

When (p\equiv1\pmod4), (-1) is a square, and (\overline K) stabilises the sets of squares (R) and non‑squares (N). A counting argument (Lemma 3.2) shows that every element of (\overline K) has order 2, and a careful analysis (Lemmas 3.3–3.6) forces the existence of an element (\lambda) acting as (z\mapsto z-1) on (R) and (z\mapsto cz-1) on (N). By a further calculation one proves (c=1), which yields the involution (z\mapsto -1/z) and places us in case (a).

When (p\equiv3\pmod4), (-1) is a non‑square, so (\overline K) interchanges (R) and (N). One introduces a parameter (c\in N) and an integer (n) describing how (\overline K) conjugates the scalar maps. Two sub‑cases arise. In the “main case” (c=-1); Lemma 4.4 shows that the only solutions of (x^{n}=x) in ((\mathbb{Z}/p)^{\times}) are (1) and (-1), which forces (n\equiv-1\pmod{(p-1)/2}) and again yields the involution (z\mapsto -1/z). In the “special case” (c\neq-1) one derives algebraic relations (c^{3}=-1) and either (c^{4}+3=0) or (3c^{4}+1=0). These equations force (p=7) and give the explicit involutions listed in the theorem. The group generated by translations, the map (z\mapsto2z) and the involution is shown to have order 168 with a normal subgroup of order 8, matching the structure described in case (b).

The corollary follows: for any prime (p>3) the only simple group of order ((p^{3}-p)/2) is (\mathrm{PSL}{2}(\mathbb{Z}/p)). The case (p=3) is trivial ((|G|=12) and the only simple group of that order is (A{4})).

The final section gives a concise modern proof that (\mathrm{PSL}{2}(F)) is simple whenever (|F|>3). By exhibiting a non‑scalar matrix in any non‑trivial normal subgroup (N) of (SL{2}(F)) and using the elementary subgroups (P={\begin{pmatrix}1&0\* &1\end{pmatrix}}) and (P’={\begin{pmatrix}1&*\0&1\end{pmatrix}}), one shows that (P\subseteq N) and hence (P’\subseteq N); consequently (N=SL_{2}(F)). Quotienting by the centre yields the simplicity of (\mathrm{PSL}_{2}(F)).

Overall, the paper succeeds in translating Frobenius’s early 20th‑century argument into a clear, pedagogical format, while also providing a modern, self‑contained proof of the simplicity of (\mathrm{PSL}_{2}) groups. The treatment of the exceptional (p=7) case illustrates how delicate the interplay between Sylow theory, quadratic residues, and permutation transitivity can be, and it offers a concrete example that can be explored in an undergraduate honors course.