# P-group generation algorithm

In mathematics, specifically group theory, finite groups of prime power order ${\displaystyle p^{n}}$, for a fixed prime number ${\displaystyle p}$ and varying integer exponents ${\displaystyle n\geq 0}$, are briefly called finite p-groups.

The p-group generation algorithm by M. F. Newman [1] and E. A. O'Brien [2] [3] is a recursive process for constructing the descendant tree of an assigned finite p-group which is taken as the root of the tree.

## Lower exponent-p central series

For a finite p-group ${\displaystyle G}$, the lower exponent-p central series (briefly lower p-central series) of ${\displaystyle G}$ is a descending series ${\displaystyle (P_{j}(G))_{j\geq 0}}$ of characteristic subgroups of ${\displaystyle G}$, defined recursively by

${\displaystyle (1)\qquad P_{0}(G):=G}$ and ${\displaystyle P_{j}(G):=\lbrack P_{j-1}(G),G\rbrack \cdot P_{j-1}(G)^{p}}$, for ${\displaystyle j\geq 1}$.

Since any non-trivial finite p-group ${\displaystyle G>1}$ is nilpotent, there exists an integer ${\displaystyle c\geq 1}$ such that ${\displaystyle P_{c-1}(G)>P_{c}(G)=1}$ and ${\displaystyle \mathrm {cl} _{p}(G):=c}$ is called the exponent-p class (briefly p-class) of ${\displaystyle G}$. Only the trivial group ${\displaystyle 1}$ has ${\displaystyle \mathrm {cl} _{p}(1)=0}$. Generally, for any finite p-group ${\displaystyle G}$, its p-class can be defined as ${\displaystyle \mathrm {cl} _{p}(G):=\min \lbrace c\geq 0\mid P_{c}(G)=1\rbrace }$.

The complete lower p-central series of ${\displaystyle G}$ is therefore given by

${\displaystyle (2)\qquad G=P_{0}(G)>\Phi (G)=P_{1}(G)>P_{2}(G)>\cdots >P_{c-1}(G)>P_{c}(G)=1}$,

since ${\displaystyle P_{1}(G)=\lbrack P_{0}(G),G\rbrack \cdot P_{0}(G)^{p}=\lbrack G,G\rbrack \cdot G^{p}=\Phi (G)}$ is the Frattini subgroup of ${\displaystyle G}$.

For the convenience of the reader and for pointing out the shifted numeration, we recall that the (usual) lower central series of ${\displaystyle G}$ is also a descending series ${\displaystyle (\gamma _{j}(G))_{j\geq 1}}$ of characteristic subgroups of ${\displaystyle G}$, defined recursively by

${\displaystyle (3)\qquad \gamma _{1}(G):=G}$ and ${\displaystyle \gamma _{j}(G):=\lbrack \gamma _{j-1}(G),G\rbrack }$, for ${\displaystyle j\geq 2}$.

As above, for any non-trivial finite p-group ${\displaystyle G>1}$, there exists an integer ${\displaystyle c\geq 1}$ such that ${\displaystyle \gamma _{c}(G)>\gamma _{c+1}(G)=1}$ and ${\displaystyle \mathrm {cl} (G):=c}$ is called the nilpotency class of ${\displaystyle G}$, whereas ${\displaystyle c+1}$ is called the index of nilpotency of ${\displaystyle G}$. Only the trivial group ${\displaystyle 1}$ has ${\displaystyle \mathrm {cl} (1)=0}$.

The complete lower central series of ${\displaystyle G}$ is given by

${\displaystyle (4)\qquad G=\gamma _{1}(G)>G^{\prime }=\gamma _{2}(G)>\gamma _{3}(G)>\cdots >\gamma _{c}(G)>\gamma _{c+1}(G)=1}$,

since ${\displaystyle \gamma _{2}(G)=\lbrack \gamma _{1}(G),G\rbrack =\lbrack G,G\rbrack =G^{\prime }}$ is the commutator subgroup or derived subgroup of ${\displaystyle G}$.

The following Rules should be remembered for the exponent-p class:

Let ${\displaystyle G}$ be a finite p-group.

1. Rule: ${\displaystyle \mathrm {cl} (G)\leq \mathrm {cl} _{p}(G)}$, since the ${\displaystyle \gamma _{j}(G)}$ descend more quickly than the ${\displaystyle P_{j}(G)}$.
2. Rule: If ${\displaystyle \vartheta \in \mathrm {Hom} (G,{\tilde {G}})}$, for some group ${\displaystyle {\tilde {G}}}$, then ${\displaystyle \vartheta (P_{j}(G))=P_{j}(\vartheta (G))}$, for any ${\displaystyle j\geq 0}$.
3. Rule: For any ${\displaystyle c\geq 0}$, the conditions ${\displaystyle N\triangleleft G}$ and ${\displaystyle \mathrm {cl} _{p}(G/N)=c}$ imply ${\displaystyle P_{c}(G)\leq N}$.
4. Rule: Let ${\displaystyle c\geq 0}$. If ${\displaystyle \mathrm {cl} _{p}(G)=c}$, then ${\displaystyle \mathrm {cl} _{p}(G/P_{k}(G))=\min(k,c)}$, for all ${\displaystyle k\geq 0}$, in particular, ${\displaystyle \mathrm {cl} _{p}(G/P_{k}(G))=k}$, for all ${\displaystyle 0\leq k\leq c}$.

## Parents and descendant trees

The parent ${\displaystyle \pi (G)}$ of a finite non-trivial p-group ${\displaystyle G>1}$ with exponent-p class ${\displaystyle \mathrm {cl} _{p}(G)=c\geq 1}$ is defined as the quotient ${\displaystyle \pi (G):=G/P_{c-1}(G)}$ of ${\displaystyle G}$ by the last non-trivial term ${\displaystyle P_{c-1}(G)>1}$ of the lower exponent-p central series of ${\displaystyle G}$. Conversely, in this case, ${\displaystyle G}$ is called an immediate descendant of ${\displaystyle \pi (G)}$. The p-classes of parent and immediate descendant are connected by ${\displaystyle \mathrm {cl} _{p}(G)=\mathrm {cl} _{p}(\pi (G))+1}$.

A descendant tree is a hierarchical structure for visualizing parent-descendant relations between isomorphism classes of finite p-groups. The vertices of a descendant tree are isomorphism classes of finite p-groups. However, a vertex will always be labelled by selecting a representative of the corresponding isomorphism class. Whenever a vertex ${\displaystyle \pi (G)}$ is the parent of a vertex ${\displaystyle G}$ a directed edge of the descendant tree is defined by ${\displaystyle G\to \pi (G)}$ in the direction of the canonical projection ${\displaystyle \pi :G\to \pi (G)}$ onto the quotient ${\displaystyle \pi (G)=G/P_{c-1}(G)}$.

In a descendant tree, the concepts of parents and immediate descendants can be generalized. A vertex ${\displaystyle R}$ is a descendant of a vertex ${\displaystyle P}$, and ${\displaystyle P}$ is an ancestor of ${\displaystyle R}$, if either ${\displaystyle R}$ is equal to ${\displaystyle P}$ or there is a path

${\displaystyle (5)\qquad R=Q_{0}\to Q_{1}\to \cdots \to Q_{m-1}\to Q_{m}=P}$, where ${\displaystyle m\geq 1}$,

of directed edges from ${\displaystyle R}$ to ${\displaystyle P}$. The vertices forming the path necessarily coincide with the iterated parents ${\displaystyle Q_{j}=\pi ^{j}(R)}$ of ${\displaystyle R}$, with ${\displaystyle 0\leq j\leq m}$:

${\displaystyle (6)\qquad R=\pi ^{0}(R)\to \pi ^{1}(R)\to \cdots \to \pi ^{m-1}(R)\to \pi ^{m}(R)=P}$, where ${\displaystyle m\geq 1}$.

They can also be viewed as the successive quotients ${\displaystyle Q_{j}=R/P_{c-j}(R)}$ of p-class ${\displaystyle c-j}$ of ${\displaystyle R}$ when the p-class of ${\displaystyle R}$ is given by ${\displaystyle \mathrm {cl} _{p}(R)=c\geq m}$:

${\displaystyle (7)\qquad R\simeq R/P_{c}(R)\to R/P_{c-1}(R)\to \cdots \to R/P_{c+1-m}(R)\to R/P_{c-m}(R)\simeq P}$, where ${\displaystyle c\geq m\geq 1}$.

In particular, every non-trivial finite p-group ${\displaystyle G>1}$ defines a maximal path (consisting of ${\displaystyle c=\mathrm {cl} _{p}(G)}$ edges)

${\displaystyle (8)\qquad G\simeq G/1=G/P_{c}(G)\to \pi (G)=G/P_{c-1}(G)\to \pi ^{2}(G)=G/P_{c-2}(G)\to \cdots }$

${\displaystyle \cdots \to \pi ^{c-1}(G)=G/P_{1}(G)\to \pi ^{c}(G)=G/P_{0}(G)=G/G\simeq 1}$

ending in the trivial group ${\displaystyle \pi ^{c}(G)=1}$. The last but one quotient of the maximal path of ${\displaystyle G}$ is the elementary abelian p-group ${\displaystyle \pi ^{c-1}(G)=G/P_{1}(G)\simeq C_{p}^{d}}$ of rank ${\displaystyle d=d(G)}$, where ${\displaystyle d(G)=\dim _{\mathbb {F} _{p}}(H^{1}(G,\mathbb {F} _{p}))}$ denotes the generator rank of ${\displaystyle G}$.

Generally, the descendant tree ${\displaystyle {\mathcal {T}}(G)}$ of a vertex ${\displaystyle G}$ is the subtree of all descendants of ${\displaystyle G}$, starting at the root ${\displaystyle G}$. The maximal possible descendant tree ${\displaystyle {\mathcal {T}}(1)}$ of the trivial group ${\displaystyle 1}$ contains all finite p-groups and is exceptional, since the trivial group ${\displaystyle 1}$ has all the infinitely many elementary abelian p-groups with varying generator rank ${\displaystyle d\geq 1}$ as its immediate descendants. However, any non-trivial finite p-group (of order divisible by ${\displaystyle p}$) possesses only finitely many immediate descendants.

## p-covering group, p-multiplicator and nucleus

Let ${\displaystyle G}$ be a finite p-group with ${\displaystyle d}$ generators. Our goal is to compile a complete list of pairwise non-isomorphic immediate descendants of ${\displaystyle G}$. It turns out that all immediate descendants can be obtained as quotients of a certain extension ${\displaystyle G^{\ast }}$ of ${\displaystyle G}$ which is called the p-covering group of ${\displaystyle G}$ and can be constructed in the following manner.

We can certainly find a presentation of ${\displaystyle G}$ in the form of an exact sequence

${\displaystyle (9)\qquad 1\longrightarrow R\longrightarrow F\longrightarrow G\longrightarrow 1}$,

where ${\displaystyle F}$ denotes the free group with ${\displaystyle d}$ generators and ${\displaystyle \vartheta :\ F\longrightarrow G}$ is an epimorphism with kernel ${\displaystyle R:=\ker(\vartheta )}$. Then ${\displaystyle R\triangleleft F}$ is a normal subgroup of ${\displaystyle F}$ consisting of the defining relations for ${\displaystyle G\simeq F/R}$. For elements ${\displaystyle r\in R}$ and ${\displaystyle f\in F}$, the conjugate ${\displaystyle f^{-1}rf\in R}$ and thus also the commutator ${\displaystyle \lbrack r,f\rbrack =r^{-1}f^{-1}rf\in R}$ are contained in ${\displaystyle R}$. Consequently, ${\displaystyle R^{\ast }:=\lbrack R,F\rbrack \cdot R^{p}}$ is a characteristic subgroup of ${\displaystyle R}$, and the p-multiplicator ${\displaystyle R/R^{\ast }}$ of ${\displaystyle G}$ is an elementary abelian p-group, since

${\displaystyle (10)\qquad \lbrack R,R\rbrack \cdot R^{p}\leq \lbrack R,F\rbrack \cdot R^{p}=R^{\ast }}$.

Now we can define the p-covering group of ${\displaystyle G}$ by

${\displaystyle (11)\qquad G^{\ast }:=F/R^{\ast }}$,

and the exact sequence

${\displaystyle (12)\qquad 1\longrightarrow R/R^{\ast }\longrightarrow F/R^{\ast }\longrightarrow F/R\longrightarrow 1}$

shows that ${\displaystyle G^{\ast }}$ is an extension of ${\displaystyle G}$ by the elementary abelian p-multiplicator. We call

${\displaystyle (13)\qquad \mu (G):=\dim _{\mathbb {F} _{p}}(R/R^{\ast })}$

the p-multiplicator rank of ${\displaystyle G}$.

Let us assume now that the assigned finite p-group ${\displaystyle G\simeq F/R}$ is of p-class ${\displaystyle \mathrm {cl} _{p}(G)=c}$. Then the conditions ${\displaystyle R\triangleleft F}$ and ${\displaystyle \mathrm {cl} _{p}(F/R)=c}$ imply ${\displaystyle P_{c}(F)\leq R}$, according to the rule (R3), and we can define the nucleus of ${\displaystyle G}$ by

${\displaystyle (14)\qquad P_{c}(G^{\ast })=P_{c}(F)\cdot R^{\ast }/R^{\ast }\leq R/R^{\ast }}$

as a subgroup of the p-multiplicator. Consequently, the nuclear rank

${\displaystyle (15)\qquad \nu (G):=\dim _{\mathbb {F} _{p}}(P_{c}(G^{\ast }))\leq \mu (G)}$

of ${\displaystyle G}$ is bounded from above by the p-multiplicator rank.

## Allowable subgroups of the p-multiplicator

As before, let ${\displaystyle G}$ be a finite p-group with ${\displaystyle d}$ generators.

Proposition. Any p-elementary abelian central extension

${\displaystyle (16)\qquad 1\to Z\to H\to G\to 1}$

of ${\displaystyle G}$ by a p-elementary abelian subgroup ${\displaystyle Z\leq \zeta _{1}(H)}$ such that ${\displaystyle d(H)=d(G)=d}$ is a quotient of the p-covering group ${\displaystyle G^{\ast }}$ of ${\displaystyle G}$.

For the proof click show on the right hand side.

Proof

The reason is that, since ${\displaystyle d(H)=d(G)=d}$, there exists an epimorphism ${\displaystyle \psi :\ F\to H}$ such that ${\displaystyle \vartheta =\omega \circ \psi }$, where ${\displaystyle \omega :\ H\to H/Z\simeq G}$ denotes the canonical projection. Consequently, we have

${\displaystyle R=\ker(\vartheta )=\ker(\omega \circ \psi )=(\omega \circ \psi )^{-1}(1)=\psi ^{-1}(\omega ^{-1}(1))=\psi ^{-1}(Z)}$

and thus ${\displaystyle \psi (R)=\psi (\psi ^{-1}(Z))=Z}$. Further, ${\displaystyle \psi (R^{p})=Z^{p}=1}$, since ${\displaystyle Z}$ is p-elementary, and ${\displaystyle \psi (\lbrack R,F\rbrack )=\lbrack Z,H\rbrack =1}$, since ${\displaystyle Z}$ is central. Together this shows that ${\displaystyle \psi (R^{\ast })=\psi (\lbrack R,F\rbrack \cdot R^{p})=1}$ and thus ${\displaystyle \psi }$ induces the desired epimorphism ${\displaystyle \psi ^{\ast }:\ G^{\ast }\to H}$ such that ${\displaystyle H\simeq G^{\ast }/\ker(\psi ^{\ast })}$.

In particular, an immediate descendant ${\displaystyle H}$ of ${\displaystyle G}$ is a p-elementary abelian central extension

${\displaystyle (17)\qquad 1\to P_{c-1}(H)\to H\to G\to 1}$

of ${\displaystyle G}$, since

${\displaystyle 1=P_{c}(H)=\lbrack P_{c-1}(H),H\rbrack \cdot P_{c-1}(H)^{p}}$ implies ${\displaystyle P_{c-1}(H)^{p}=1}$ and ${\displaystyle P_{c-1}(H)\leq \zeta _{1}(H)}$,

where ${\displaystyle c=\mathrm {cl} _{p}(H)}$.

Definition. A subgroup ${\displaystyle M/R^{\ast }\leq R/R^{\ast }}$ of the p-multiplicator of ${\displaystyle G}$ is called allowable if it is given by the kernel ${\displaystyle M/R^{\ast }=\ker(\psi ^{\ast })}$ of an epimorphism ${\displaystyle \psi ^{\ast }:\ G^{\ast }\to H}$ onto an immediate descendant ${\displaystyle H}$ of ${\displaystyle G}$.

An equivalent characterization is that ${\displaystyle 1 is a proper subgroup which supplements the nucleus

${\displaystyle (18)\qquad (M/R^{\ast })\cdot (P_{c}(F)\cdot R^{\ast }/R^{\ast })=R/R^{\ast }}$.

Therefore, the first part of our goal to compile a list of all immediate descendants of ${\displaystyle G}$ is done, when we have constructed all allowable subgroups of ${\displaystyle R/R^{\ast }}$ which supplement the nucleus ${\displaystyle P_{c}(G^{\ast })=P_{c}(F)\cdot R^{\ast }/R^{\ast }}$, where ${\displaystyle c=\mathrm {cl} _{p}(G)}$. However, in general the list

${\displaystyle (19)\qquad \lbrace F/M\quad \mid \quad M/R^{\ast }\leq R/R^{\ast }{\text{ is allowable }}\rbrace }$,

where ${\displaystyle G^{\ast }/(M/R^{\ast })=(F/R^{\ast })/(M/R^{\ast })\simeq F/M}$, will be redundant, due to isomorphisms ${\displaystyle F/M_{1}\simeq F/M_{2}}$ among the immediate descendants.

## Orbits under extended automorphisms

Two allowable subgroups ${\displaystyle M_{1}/R^{\ast }}$ and ${\displaystyle M_{2}/R^{\ast }}$ are called equivalent if the quotients ${\displaystyle F/M_{1}\simeq F/M_{2}}$, that are the corresponding immediate descendants of ${\displaystyle G}$, are isomorphic.

Such an isomorphism ${\displaystyle \varphi :\ F/M_{1}\to F/M_{2}}$ between immediate descendants of ${\displaystyle G=F/R}$ with ${\displaystyle c=\mathrm {cl} _{p}(G)}$ has the property that ${\displaystyle \varphi (R/M_{1})=\varphi (P_{c}(F/M_{1}))=P_{c}(\varphi (F/M_{1}))=P_{c}(F/M_{2})=R/M_{2}}$ and thus induces an automorphism ${\displaystyle \alpha \in \mathrm {Aut} (G)}$ of ${\displaystyle G}$ which can be extended to an automorphism ${\displaystyle \alpha ^{\ast }\in \mathrm {Aut} (G^{\ast })}$ of the p-covering group ${\displaystyle G^{\ast }=F/R^{\ast }}$of ${\displaystyle G}$. The restriction of this extended automorphism ${\displaystyle \alpha ^{\ast }}$ to the p-multiplicator ${\displaystyle R/R^{\ast }}$ of ${\displaystyle G}$ is determined uniquely by ${\displaystyle \alpha }$.

Since ${\displaystyle \alpha ^{\ast }(M/R^{\ast })\cdot P_{c}(F/R^{\ast })=\alpha ^{\ast }\lbrack M/R^{\ast }\cdot P_{c}(F/R^{\ast })\rbrack =\alpha ^{\ast }(R/R^{\ast })=R/R^{\ast }}$, each extended automorphism ${\displaystyle \alpha ^{\ast }\in \mathrm {Aut} (G^{\ast })}$ induces a permutation ${\displaystyle \alpha ^{\prime }}$ of the allowable subgroups ${\displaystyle M/R^{\ast }\leq R/R^{\ast }}$. We define ${\displaystyle P:=\langle \alpha ^{\prime }\mid \alpha \in \mathrm {Aut} (G)\rangle }$ to be the permutation group generated by all permutations induced by automorphisms of ${\displaystyle G}$. Then the map ${\displaystyle \mathrm {Aut} (G)\to P}$, ${\displaystyle \alpha \mapsto \alpha ^{\prime }}$ is an epimorphism and the equivalence classes of allowable subgroups ${\displaystyle M/R^{\ast }\leq R/R^{\ast }}$ are precisely the orbits of allowable subgroups under the action of the permutation group ${\displaystyle P}$.

Eventually, our goal to compile a list ${\displaystyle \lbrace F/M_{i}\mid 1\leq i\leq N\rbrace }$ of all immediate descendants of ${\displaystyle G}$ will be done, when we select a representative ${\displaystyle M_{i}/R^{\ast }}$ for each of the ${\displaystyle N}$ orbits of allowable subgroups of ${\displaystyle R/R^{\ast }}$ under the action of ${\displaystyle P}$. This is precisely what the p-group generation algorithm does in a single step of the recursive procedure for constructing the descendant tree of an assigned root.

## Capable p-groups and step sizes

A finite p-group ${\displaystyle G}$ is called capable (or extendable) if it possesses at least one immediate descendant, otherwise it is terminal (or a leaf). The nuclear rank ${\displaystyle \nu (G)}$ of ${\displaystyle G}$ admits a decision about the capability of ${\displaystyle G}$:

• ${\displaystyle G}$ is terminal if and only if ${\displaystyle \nu (G)=0}$.
• ${\displaystyle G}$ is capable if and only if ${\displaystyle \nu (G)\geq 1}$.

In the case of capability, ${\displaystyle G=F/R}$ has immediate descendants of ${\displaystyle \nu =\nu (G)}$ different step sizes ${\displaystyle 1\leq s\leq \nu }$, in dependence on the index ${\displaystyle (R/R^{\ast }:M/R^{\ast })=p^{s}}$ of the corresponding allowable subgroup ${\displaystyle M/R^{\ast }}$ in the p-multiplicator ${\displaystyle R/R^{\ast }}$. When ${\displaystyle G}$ is of order ${\displaystyle \vert G\vert =p^{n}}$, then an immediate descendant of step size ${\displaystyle s}$ is of order ${\displaystyle \#(F/M)=(F/R^{\ast }:M/R^{\ast })=(F/R^{\ast }:R/R^{\ast })\cdot (R/R^{\ast }:M/R^{\ast })}$ ${\displaystyle =\#(F/R)\cdot p^{s}=\vert G\vert \cdot p^{s}=p^{n}\cdot p^{s}=p^{n+s}}$.

For the related phenomenon of multifurcation of a descendant tree at a vertex ${\displaystyle G}$ with nuclear rank ${\displaystyle \nu (G)\geq 2}$ see the article on descendant trees.

The p-group generation algorithm provides the flexibility to restrict the construction of immediate descendants to those of a single fixed step size ${\displaystyle 1\leq s\leq \nu }$, which is very convenient in the case of huge descendant numbers (see the next section).

## Numbers of immediate descendants

We denote the number of all immediate descendants, resp. immediate descendants of step size ${\displaystyle s}$, of ${\displaystyle G}$ by ${\displaystyle N}$, resp. ${\displaystyle N_{s}}$. Then we have ${\displaystyle N=\sum _{s=1}^{\nu }\,N_{s}}$. As concrete examples, we present some interesting finite metabelian p-groups with extensive sets of immediate descendants, using the SmallGroups identifiers and additionally pointing out the numbers ${\displaystyle 0\leq C_{s}\leq N_{s}}$ of capable immediate descendants in the usual format ${\displaystyle (N_{1}/C_{1};\ldots ;N_{\nu }/C_{\nu })}$ as given by actual implementations of the p-group generation algorithm in the computer algebra systems GAP and MAGMA.

First, let ${\displaystyle p=3}$.

We begin with groups having abelianization of type ${\displaystyle (3,3)}$. See Figure 4 in the article on descendant trees.

• The group ${\displaystyle \langle 27,3\rangle }$ of coclass ${\displaystyle 1}$ has ranks ${\displaystyle \nu =2}$, ${\displaystyle \mu =4}$ and descendant numbers ${\displaystyle (4/1;7/5)}$, ${\displaystyle N=11}$.
• The group ${\displaystyle \langle 243,3\rangle =\langle 27,3\rangle -\#2;1}$ of coclass ${\displaystyle 2}$ has ranks ${\displaystyle \nu =2}$, ${\displaystyle \mu =4}$ and descendant numbers ${\displaystyle (10/6;15/15)}$, ${\displaystyle N=25}$.
• One of its immediate descendants, the group ${\displaystyle \langle 729,40\rangle =\langle 243,3\rangle -\#1;7}$, has ranks ${\displaystyle \nu =2}$, ${\displaystyle \mu =5}$ and descendant numbers ${\displaystyle (16/2;27/4)}$, ${\displaystyle N=43}$.

In contrast, groups with abelianization of type ${\displaystyle (3,3,3)}$ are partially located beyond the limit of computability.

• The group ${\displaystyle \langle 81,12\rangle }$ of coclass ${\displaystyle 2}$ has ranks ${\displaystyle \nu =2}$, ${\displaystyle \mu =7}$ and descendant numbers ${\displaystyle (10/2;100/50)}$, ${\displaystyle N=110}$.
• The group ${\displaystyle \langle 243,37\rangle }$ of coclass ${\displaystyle 3}$ has ranks ${\displaystyle \nu =5}$, ${\displaystyle \mu =9}$ and descendant numbers ${\displaystyle (35/3;2783/186;81711/10202;350652/202266;\ldots )}$, ${\displaystyle N>4\cdot 10^{5}}$ unknown.
• The group ${\displaystyle \langle 729,122\rangle }$ of coclass ${\displaystyle 4}$ has ranks ${\displaystyle \nu =8}$, ${\displaystyle \mu =11}$ and descendant numbers ${\displaystyle (45/3;117919/1377;\ldots )}$, ${\displaystyle N>10^{5}}$ unknown.

Next, let ${\displaystyle p=5}$.

Corresponding groups with abelianization of type ${\displaystyle (5,5)}$ have bigger descendant numbers than for ${\displaystyle p=3}$.

• The group ${\displaystyle \langle 125,3\rangle }$ of coclass ${\displaystyle 1}$ has ranks ${\displaystyle \nu =2}$, ${\displaystyle \mu =4}$ and descendant numbers ${\displaystyle (4/1;12/6)}$, ${\displaystyle N=16}$.
• The group ${\displaystyle \langle 3125,3\rangle =\langle 125,3\rangle -\#2;1}$ of coclass ${\displaystyle 2}$ has ranks ${\displaystyle \nu =3}$, ${\displaystyle \mu =5}$ and descendant numbers ${\displaystyle (8/3;61/61;47/47)}$, ${\displaystyle N=116}$.

## Schur multiplier

Via the isomorphism ${\displaystyle \mathbb {Q} /\mathbb {Z} \to \mu _{\infty }}$, ${\displaystyle {\frac {n}{d}}\mapsto \exp({\frac {n}{d}}\cdot 2\pi i)}$ the quotient group ${\displaystyle \mathbb {Q} /\mathbb {Z} =\lbrace {\frac {n}{d}}\cdot \mathbb {Z} \mid d\geq 1,\ 0\leq n\leq d-1\rbrace }$ can be viewed as the additive analogue of the multiplicative group ${\displaystyle \mu _{\infty }=\lbrace z\in \mathbb {C} \mid z^{d}=1{\text{ for some integer }}d\geq 1\rbrace }$ of all roots of unity.

Let ${\displaystyle p}$ be a prime number and ${\displaystyle G}$ be a finite p-group with presentation ${\displaystyle G=F/R}$ as in the previous section. Then the second cohomology group ${\displaystyle M(G):=H^{2}(G,\mathbb {Q} /\mathbb {Z} )}$ of the ${\displaystyle G}$-module ${\displaystyle \mathbb {Q} /\mathbb {Z} }$ is called the Schur multiplier of ${\displaystyle G}$. It can also be interpreted as the quotient group ${\displaystyle M(G)=(R\cap \lbrack F,F\rbrack )/\lbrack F,R\rbrack }$.

I. R. Shafarevich [4] has proved that the difference between the relation rank ${\displaystyle r(G)=\dim _{\mathbb {F} _{p}}(H^{2}(G,\mathbb {F} _{p}))}$ of ${\displaystyle G}$ and the generator rank ${\displaystyle d(G)=\dim _{\mathbb {F} _{p}}(H^{1}(G,\mathbb {F} _{p}))}$ of ${\displaystyle G}$ is given by the minimal number of generators of the Schur multiplier of ${\displaystyle G}$, that is ${\displaystyle r(G)-d(G)=d(M(G))}$.

N. Boston and H. Nover [5] have shown that ${\displaystyle \mu (G_{j})-\nu (G_{j})\leq r(G)}$, for all quotients ${\displaystyle G_{j}:=G/P_{j}(G)}$ of p-class ${\displaystyle \mathrm {cl} _{p}(G_{j})=j}$, ${\displaystyle j\geq 0}$, of a pro-p group ${\displaystyle G}$ with finite abelianization ${\displaystyle G/G^{\prime }}$.

Furthermore, J. Blackhurst (in the appendix On the nucleus of certain p-groups of a paper by N. Boston, M. R. Bush and F. Hajir [6]) has proved that a non-cyclic finite p-group ${\displaystyle G}$ with trivial Schur multiplier ${\displaystyle M(G)}$ is a terminal vertex in the descendant tree ${\displaystyle {\mathcal {T}}(1)}$ of the trivial group ${\displaystyle 1}$, that is, ${\displaystyle M(G)=1}$ ${\displaystyle \Rightarrow }$ ${\displaystyle \nu (G)=0}$.

### Examples

• A finite p-group ${\displaystyle G}$ has a balanced presentation ${\displaystyle r(G)=d(G)}$ if and only if ${\displaystyle r(G)-d(G)=0=d(M(G))}$, that is, if and only if its Schur multiplier ${\displaystyle M(G)=1}$ is trivial. Such a group is called a Schur group and it must be a leaf in the descendant tree ${\displaystyle {\mathcal {T}}(1)}$.
• A finite p-group ${\displaystyle G}$ satisfies ${\displaystyle r(G)=d(G)+1}$ if and only if ${\displaystyle r(G)-d(G)=1=d(M(G))}$, that is, if and only if it has a non-trivial cyclic Schur multiplier ${\displaystyle M(G)}$. Such a group is called a Schur+1 group.

## References

1. ^ Newman, M. F. (1977). Determination of groups of prime-power order. pp. 73-84, in: Group Theory, Canberra, 1975, Lecture Notes in Math., Vol. 573, Springer, Berlin.
2. ^ O'Brien, E. A. (1990). "The p-group generation algorithm". J. Symbolic Comput. 9: 677–698. doi:10.1016/s0747-7171(08)80082-x.
3. ^ Holt, D. F., Eick, B., O'Brien, E. A. (2005). Handbook of computational group theory. Discrete mathematics and its applications, Chapman and Hall/CRC Press.CS1 maint: Multiple names: authors list (link)
4. ^ Shafarevich, I. R. (1964). "Extensions with given points of ramification (Russian)". Inst. Hautes \'Etudes Sci., Publ. Math. (English transl. in Amer. Math. Soc. Transl. (2) 59 (1966), 128-149). 18: 71–95.
5. ^ Boston, N., Nover, H. (2006). Computing pro-p Galois groups. Proceedings of the 7th Algorithmic Number Theory Symposium 2006, Lecture Notes in Computer Science 4076, 1-10, Springer, Berlin.CS1 maint: Multiple names: authors list (link)
6. ^ Boston, N., Bush, M. R., Hajir, F. (2013). "Heuristics for p-class towers of imaginary quadratic fields". Math. Ann.: (preprint: arXiv:1111.4679v1 [math.NT], 2011).CS1 maint: Multiple names: authors list (link)