Yoshida transfer theorem

Yoshida's transfer theorem characterizes when the normalizer of a Sylow subgroup controls $$p$$ -transfer. It says that the only group-theoretic obstruction is a quotient of the Sylow $$p$$ -subgroup isomorphic to the regular wreath product $C_p\wr C_p$ . The Hall–Wielandt consequence follows immediately: a Sylow $$p$$ -subgroup of nilpotency class less than $$p$$ has a normalizer that controls $$p$$ -transfer. ^{[@isaacs2008]}

p-transfer and control

Let $P\in\operatorname{Syl}_p(G)$ and define

A^p(G)=\bigcap\{N\triangleleft G\mid G/N\text{ is an abelian }p\text{-group}\}.

This is the smallest normal subgroup with abelian $$p$$ -group quotient and is also the kernel of transfer from $$G$$ to $$P/P'$$ .

Control permits the maximal abelian $$p$$ -quotient of $$G$$ to be detected inside the smaller group $$H$$ . If $$A^p(H)<H$$ , then control gives $$A^p(G)<G$$ , so except when $$|G|=p$$ , the group $$G$$ cannot be simple.

Control of fusion implies control of transfer, but the converse is false. Yoshida's theorem is powerful precisely because a Sylow normalizer may control transfer without controlling all fusion in $$P$$ .

The obstruction wreath product

The regular wreath product

W=C_p\wr C_p=B\rtimes\langle\sigma\rangle

has base group $B\cong C_p^p$ , with the top generator $\sigma$ cyclically permuting the $$p$$ coordinates. Consequently

|W|=p^{p+1}, \qquad \operatorname{cl}(W)=p,

and $$W$$ contains elements of order $$p^2$$ . It is the standard smallest model in which transfer degenerates across an index- $$p$$ layer.

Yoshida's theorem

The theorem holds for every prime. For $$p=2$$ , the obstruction is $C_2\wr C_2\cong D_8$ , so the theorem remains meaningful even though the corollary "class less than $$p$$ " then covers only abelian Sylow $$2$$ -subgroups.

Pretransfer across an index-p subgroup

If $V(M)\nsubseteq\Phi(M)$ , the norm-like product is nonzero in $M/\Phi(M)$ . The cyclic action then generates a regular $\mathbf F_pC_p$ -module and forces $C_p\wr C_p$ to be a quotient of $$P$$ . ^{[@isaacs2008]}

Proof architecture

Group-theoretic route to Yoshida's theorem

Assume that $$N=N_G(P)$$ does not control $$p$$ -transfer. The failure-of-control lemma gives

M\triangleleft N, \qquad |N:M|=p,

such that every pretransfer $U:G\to N$ satisfies $U(G)\subseteq M$ . In particular $N'\leq M$ , and $S=P\cap M$ is a normal subgroup of index $$p$$ in $$P$$ .

Choose a pretransfer $V:G\to N$ and apply the Mackey double-coset formula to $$V|_P$$ . It decomposes the restriction into local pretransfers from $$P$$ to intersections $P\cap N^x$ . Because $$N$$ normalizes all of $$P$$ , the principal double-coset term is the index- $$p$$ pretransfer from $$P$$ to $$S$$ ; the remaining terms pass through smaller intersections and are absorbed in the appropriate Frattini subgroups.

If all local terms vanished in $S/\Phi(S)$ , the principal term would force $V(P)\nsubseteq M$ , contradicting $V(G)\subseteq M$ . Hence some index- $$p$$ pretransfer is nontrivial in a Frattini quotient. The index- $$p$$ formula turns this into a cyclic module of length $$p$$ , and the preceding criterion constructs a normal subgroup $$K$$ with

P/K\cong C_p\wr C_p.

This contradicts the absence of such a quotient, so $$N$$ controls transfer.

Yoshida's original proof was character-theoretic. Isaacs gives a more group-theoretic proof based on pretransfer, the Mackey formula, and finite $$p$$ -group structure. ^{[@yoshida1978]} ^{[@isaacs2008]}

Hall–Wielandt consequence

For odd $$p$$ , this contains the result for abelian Sylow subgroups, whose class is $$1$$ . At $$p=2$$ , the abelian case is normally obtained from the earlier fusion-control theorem.

Other sufficient conditions

Yoshida's criterion is stronger than a class bound because it excludes only one particular quotient:

regular $$p$$ -groups have no quotient $C_p\wr C_p$ ;
if $$p>2$$ and $$P$$ is nonabelian metacyclic, all its subgroups and quotients are metacyclic, while $C_p\wr C_p$ is not;
powerful $$p$$ -groups and several broader classes also exclude the obstruction;
exponent and power-collection identities can rule it out without bounding nilpotency class.

Sharpness and counterexamples

Modern extension

Transfer and the hyperfocal subgroup can also be defined for saturated fusion systems. Tate- and Yoshida-type theorems show that control of transfer by a subsystem controls stronger $$p$$ -residual structure, with the same wreath-like fusion pattern appearing as an obstruction. ^[@diaz2011]

This article reconstructs the group-theoretic proof in Isaacs, Section 10A, including the control definition, obstruction group, index- $$p$$ pretransfer formula, Mackey decomposition, and principal consequences. The original character-theoretic proof and fusion-system generalization provide independent context. ^{[@isaacs2008]} ^{[@yoshida1978]} ^[@diaz2011]