Finite Index Rigidity of Relatively Hyperbolic Groups

Alessandro Sisto; Gon Rahamim; Nir Lazarovich

REVIEW 2 major objections 5 minor 27 references

Finite Index Rigidity of Relatively Hyperbolic Groups

T0 review · 2 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper claims that torsion-free relatively hyperbolic groups with non-relatively-hyperbolic peripheral subgroups are finite index rigid: any two isomorphic finite-index subgroups have the same index.

desk verdict A significant finite-index rigidity theorem for relatively hyperbolic groups, with the right ideas, but the written proof has a genuine gap around the boundary homeomorphism that needs patching before it is airtight. read the letter →

arxiv 2509.04323 v1 pith:EJ4JWMOR submitted 2025-09-04 math.GR math.GT

Nir Lazarovich , Gon Rahamim , Alessandro Sisto This is my paper · ORCID

classification math.GRmath.GT MSC 20F6720F6520J05

keywords finiteindexrigidityrelativelyhyperbolicgroupsrelativeclassifyingspacecomplexityBowditchboundaryweightedsingularpatternscohomologyperipheralsubgroups

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proves a rigidity theorem for a large class of groups: if a torsion-free group G is hyperbolic relative to proper subgroups that are themselves not relatively hyperbolic, then G is finite index rigid—isomorphic finite-index subgroups must have equal index. It also proves a more general statement: whenever an isomorphism between finite-index subgroups preserves the induced peripheral structure, the indices are equal, with no extra hypothesis on the peripheral subgroups. The proof introduces a complexity invariant for a group pair, namely the minimal number of cells in a simplicial relative classifying space, and shows this complexity grows linearly with the index of a finite-index subgroup. This linear growth is then converted, via a limit-inferior multiplicative invariant, into equality of indices for isomorphic subgroup pairs.

What carries the argument

The key object is the complexity C(G,P): the minimal number of cells in a simplicial relative classifying space for a group pair (G,P), where the peripheral subgroups P are represented by subcomplexes that are classifying spaces for them. The proof shows that for a torsion-free one-ended relatively hyperbolic group, the complexity of a finite-index subgroup H with its induced peripheral structure grows at least linearly with [G:H] and at most linearly via covers of a minimal complex. The lower bound is obtained by resolving a globally stable bicombing on a combinatorial cusped hyperbolic graph into a weighted singular pattern on the 2-skeleton of the classifying complex, bounding the total w

What would settle it

Find a group satisfying the hypotheses of Theorem 1.1, for instance a torsion-free group hyperbolic relative to a non-virtually-cyclic nilpotent subgroup, with two finite-index subgroups H and H' that are isomorphic but have [G:H] ≠ [G:H']. The theorem predicts no such pair exists; a more localized check is whether the boundary extension built in Section 7 can fail to be bijective, which would invalidate the application of Theorem 6.1.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: if G is a torsion-free group not isomorphic to Z, hyperbolic relative to a finite collection of type-F and non-relatively-hyperbolic proper subgroups, then G is finite index rigid. The proof proceeds by first establishing Theorem 1.5, which says that if two finite-index subgroups H and H' of a torsion-free relatively hyperbolic group G admit an isomorphism preserving their induced peripheral structures, then [G:H] = [G:H']. The paper then shows that, under the hypotheses of Theorem 1.1, every isomorphism between finite-index subgroups automatically preserves peripheral structures: peripheral subgroups are quasiconvex, undistorted, and non-relatively-hyperbol

Load-bearing premise

The proof's lower bound on complexity assumes that the continuous extension of the map Phi to Bowditch boundaries constructed in Section 7 is a homeomorphism; the text only proves continuity before invoking the theorem that needs bijectivity.

Editorial extensions

If this is right

Fundamental groups of complete finite-volume manifolds of pinched negative curvature are finite index rigid.
Non-abelian limit groups are finite index rigid.
Torsion-free groups hyperbolic relative to nilpotent subgroups are finite index rigid.
Free-by-cyclic groups with an exponentially growing automorphism are finite index rigid.
For any torsion-free relatively hyperbolic group, an isomorphism between finite-index subgroups that preserves the induced peripheral structure forces equal indices.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The linear-growth mechanism may extend beyond relative hyperbolicity: any setup with a hyperbolic cusped space admitting a globally stable bicombing and a cohomological boundary-control theorem could support a similar complexity-vs-index argument.
The complexity invariant might be computable in explicit examples, potentially giving effective bounds on indices and an algorithmic way to detect when isomorphic finite-index subgroups cannot have different indices.
The peripheral-structure-preserving theorem suggests a relative analogue of Mostow-style rigidity: in many relatively hyperbolic groups, isomorphism of finite-index subgroups may automatically preserve peripheral structures whenever the peripherals are 'large enough' in a coarse-geometric sense.
The paper's approach could also yield finite index rigidity for groups that are only virtually torsion-free, by passing to a torsion-free finite-index subgroup and tracking index changes.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a real result, but the proof as written has at least one genuine gap that needs fixing before it can be trusted. The main theorem — finite index rigidity for torsion-free relatively hyperbolic groups with NRH peripherals — is a natural extension of Lazarovich’s hyperbolic rigidity, and the applications to pinched negative curvature, limit groups, and free-by-cyclic groups are substantial. The new technical core is a uniform lower bound on the complexity of relative classifying spaces over all finite-index subgroups, combining weighted patterns from [18] with Manning–Wang’s cohomological results. That is the right strategy and the result is likely true.

The serious issue is in Section 7. Theorem 6.1 requires the map Φ to extend to a homeomorphism of Bowditch boundaries. The proof only cites Lemma 6.3, which gives a continuous extension. Continuity does not imply the boundary map is bijective, and without bijectivity the cohomology pushforward in Theorem 6.1 can collapse the nontrivial class, so the uniform quasi-surjectivity conclusion does not follow. This is a real gap in the written argument.

That said, it is probably repairable. An arbitrary H-equivariant vertex map between the universal covers of finite classifying spaces for H and G is a quasi-isometry (because H has finite index in G), and the extension constructed in §7 should be a quasi-isometry of cusped spaces, which would induce the required boundary homeomorphism. The authors just don’t say that, and a referee cannot be expected to supply it silently.

There is also a smaller equivariance typo in §8.1, where Proposition 4.2 is applied to a complex with a G-action when the natural action is by H. The argument likely goes through with H substituted throughout, but as written it is inconsistent.

The paper relies on [18] and [20] for heavy machinery, which is fine. The gaps are in the new parts. I would not desk-reject this; the main idea is credible and the missing pieces are fixable. But it needs a serious referee who will either verify the boundary homeomorphism or demand the missing argument. If repaired, this is a strong paper. For a reading group, I’d say maybe — the ideas are worth discussing despite the incomplete state. I’d cite it once the gaps are patched.

Recommendation: send to peer review, with the boundary-homeomorphism issue as the main point to resolve.

Referee Report

2 major / 5 minor

Summary. The paper proves finite index rigidity for torsion-free relatively hyperbolic groups whose peripheral subgroups are non-relatively-hyperbolic (NRH), and a more general rigidity statement for isomorphisms that preserve induced peripheral structures. The engine is a complexity invariant C(G,P), defined as the minimal number of cells in a simplicial relative classifying space. The authors prove a linear lower bound alpha[G:H] <= C(H,P_H) for finite-index H (Theorem 1.4) by combining weighted singular patterns from globally stable bicombings (Sections 3-4), cohomological uniform quasi-surjectivity for cylindrical cusped spaces (Sections 5-6), and a map-construction argument (Section 7). The upper bound C(H,P_H) <= beta[G:H] is a covering-space argument. A Reznikov-style liminf then converts these bounds into index rigidity.

Significance. If the proof is completed, the paper substantially extends the first author's finite-index rigidity theorem for hyperbolic groups to a broad relative setting, covering fundamental groups of finite-volume negatively curved manifolds, limit groups, and exponentially growing free-by-cyclic groups. The complexity invariant is natural and is defined independently of the target theorem; the linear bounds are derived rather than fitted. The proof strategy is credible and the paper is well organized. The main reservation is a missing verification of a boundary-homeomorphism hypothesis in Section 7; I regard the gap as likely repairable rather than fatal.

major comments (2)

[§7, Step 3; Theorem 6.1] Theorem 6.1 is applied to conclude X subseteq N_{R0}(Phi(Ccyl(K))). Its hypothesis requires Phi to extend continuously to a homeomorphism dPhi: dCcyl(K,B) -> dCcyl(X,A). In Step 3 the authors write: 'By Lemma 6.3, Phi extends continuously to a map dPhi. It now follows from Theorem 6.1...' But Lemma 6.3 proves only continuity of dPhi; it does not establish injectivity or surjectivity. The proof of Theorem 6.1 uses (dPhi)^* being an isomorphism on Cech cohomology to pull back a nonzero class and contradict Phi(C(K)) cap B_{R0} = empty. If dPhi is merely continuous, (dPhi)^* need not be injective, so the contradiction fails. Consequently the uniform quasi-surjectivity that underpins the lower bound in Section 8.1 is not established as written. This is likely fixable, e.g. by proving the constructed Phi is a quasi-isometry and using the standard boundary homeomorphism, or by stating a weaker
[§8.1, Proposition 4.2 application] The proof of Theorem 1.4 says: 'Let K be an aspherical simplicial complex, such that Vol(K/H)=C(H,P_H). By Proposition 4.2, there exist a G-equivariant map Phi_0: L_0 -> X_0...' This is inconsistent with the setup: K is a classifying space for H, so K carries an H-action, not necessarily a G-action. Proposition 4.2 as stated requires a free cocompact G-action on K. The resolution and bound (8.1) can only be applied with H in place of G; the map should be H-equivariant. Since Theorem 7.1 is then invoked for an H-equivariant extension, I suspect this is a typo, but as written the application of Proposition 4.2 is unjustified and should be corrected by stating and using the H-version of Proposition 4.2.

minor comments (5)

[§4, Claim 4.4] In the displayed inequality after (4.3), there are typos: d(\tilde{\Psi}((e_i)_-), \tilde{\Phi}((e_i)_+)) should be d(\tilde{\Psi}((e_i)_-), \tilde{\Psi}((e_i)_+)); and several parentheses are misplaced, e.g. d(\Psi(e_i)_-), \Psi((e_i)_+)) should be d(\Psi((e_i)_-), \Psi((e_i)_+)). These do not affect the argument but should be corrected.
[§7, Equation (7.1)] The displayed implication contains a stray LaTeX control sequence `/Leftr⫯g⊸tl⫯ne⇒`; it should be a simple implication arrow.
[§6, Theorem 6.1 proof] The arrow in the diagram `i /leftr⫯g⊸tl⫯ne →H^k_c(C(X))` is malformed and should be typeset as a standard arrow.
[§3, Lemma 3.4(T3)] The proof says 'the action G↷C(X) is free' and concludes track stabilizers are trivial. Proposition 4.2 is stated without a torsion-freeness assumption. If the proposition is intended in that generality, either add torsion-free as a hypothesis or replace 'trivial' by 'finite' (which suffices for the argument).
[§4, Claim 4.3] In the proof, 'local compactness of C(x)' should be 'local compactness of C(X)'.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the complexity invariant and weighted-pattern bounds are independent inputs; main risk is a §7 boundary-homeomorphism gap, not circularity. Score 2 reflects non-load-bearing self-citation of [18].

full rationale

Walking the derivation chain: the lower bound α[G:H] ≤ C(H,PH) is obtained by (i) defining complexity C as the minimal number of cells in a simplicial relative classifying space (Definition 1.3); (ii) bounding the total weight of a singular pattern on a minimal 2-complex above by complexity (Proposition 4.2); (iii) constructing a quasi-surjective map Φ via a Rips complex and applying cohomological uniform quasi-surjectivity (Theorem 6.1); (iv) converting disjoint R0-balls in X/H into tracks of weight ≥ 1/λ. Each input is a theorem about classifying spaces, bicombings, or cohomology, not the target rigidity statement. The heavy use of [18] supplies the weighted-pattern machinery and several elementary lemmas whose proofs are either repeated in the text or said to 'readily carry over' because they use only the bicombing axioms; this is self-citation, but the relative-rigidity theorem does not reduce by construction to [18]'s finite-index-rigidity theorem. Likewise, the Reznikov liminf invariant in §8 is multiplicative via the elementary covering inequality C(L∩H) ≤ [L:L∩H]C(L), not by assuming the desired index equality. No equation in the paper fits a parameter to data and then renames it a prediction, and no target quantity appears inside its own hypothesis. One genuine gap exists in §7, Step 3: the text says 'By Lemma 6.3, Φ extends continuously to a map ∂Φ. It now follows from Theorem 6.1...' but Theorem 6.1 requires ∂Φ to be a homeomorphism, while Lemma 6.3 only establishes continuity. This is a correctness gap (possibly fixable by strengthening the boundary argument), not a circularity, so I have not counted it as a circular step; it is noted here per the reviewing rule to flag missing support explicitly.

Assumptions & free parameters 0 free parameters · 6 assumptions · 1 invented entities

The central claim rests on several deep external theorems in relatively hyperbolic group theory, but none of them are free parameters fitted to the result. The complexity invariant is the only newly introduced mathematical object, and it has independent grounding through finiteness and linear-growth results proven in the paper.

assumptions (6)

domain assumption Existence of globally-stable bicombing on combinatorial cusped graphs for relatively hyperbolic pairs (Mineyev; Groves-Manning).
Invoked in Theorem 2.11 and used throughout Sections 3-4 to define resolutions and weight patterns; this is an external deep theorem.
domain assumption Dahmani's theorem: torsion-free relatively hyperbolic groups with type-F peripherals admit finite relative classifying spaces.
Used to ensure C(G,P) is finite and to obtain the upper bound via covers; referenced in Section 2.4 and used in Section 8.1.
domain assumption Manning-Wang: relative cohomology of (G,P) is isomorphic to reduced cohomology of the Bowditch boundary, and cylindrical cusped spaces are quasi-isometric to combinatorial cusped spaces.
Used in Section 5 and Theorem 6.1; imported from [20].
standard math Rips complexes of hyperbolic spaces are contractible for suitable parameters (Bridson-Haefliger).
Used in Section 7 to extend maps by filling disks.
domain assumption Osin's results: peripheral subgroups are quasiconvex and distinct conjugates of infinite peripheral subgroups cannot be strictly nested.
Used in the proof of Theorem 1.1 to show isomorphisms preserve peripheral structure.
domain assumption NRH is a quasi-isometry invariant (Drutu), and thick groups are NRH (Behrstock-Drutu-Mosher).
Used in Theorem 1.1 and Corollary 1.2 to verify that peripheral subgroups are NRH.

invented entities (1)

Relative classifying space complexity C(G,P) independent evidence
purpose: Attaches a number to a relatively hyperbolic pair so that it multiplies by the index under finite covers; used to prove index rigidity.
Well-defined as the minimum number of cells in a simplicial relative classifying space; finiteness follows from Dahmani's theorem, and the paper proves linear bounds, so it is not an ad hoc construct pulled from a hat.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Finite Index Rigidity of Relatively Hyperbolic Groups." pith.science (2026). https://pith.science/paper/EJ4JWMOR

@misc{pith2026250904323,
  author       = {Pith},
  title        = {Pith review of: Finite Index Rigidity of Relatively Hyperbolic Groups},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EJ4JWMOR}},
  note         = {Machine review of arXiv:2509.04323}
}

read the original abstract

We prove that, given a torsion-free relatively hyperbolic group G with non-relatively-hyperbolic peripherals, isomorphic finite index subgroups of G have the same index. This applies for instance to fundamental groups of finite-volume negatively curved manifolds, to limit groups, and to free-by-cyclic groups. More generally, we show that if two finite index subgroups of a relatively hyperbolic group are isomorphic via a map that respects their peripheral structures, then their indices in the ambient group are equal. The proof relies on demonstrating that the number of simplices in a simplicial classifying space of a finite index subgroup in a relatively hyperbolic group grows linearly with its index. These results generalize earlier work of the first author in the context of hyperbolic groups.

Figures

Figures reproduced from arXiv: 2509.04323 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 26 canonical work pages

[18]

Finite index rigidity of hyperbolic groups

Nir Lazarovich. Finite index rigidity of hyperbolic groups. arXiv preprint arXiv:2302.04484, 2023

work page Pith review arXiv 2023
[20]

Cohomology and the Bowditch boundary

Jason F Manning and Oliver H Wang. Cohomology and the Bowditch boundary. Michigan Mathematical Journal , 69(3):633–669, 2020

work page 2020
[1]

Thick metric spaces, relative hyperbolicity, and quasi-isometric rigidity.Math

Jason Behrstock, Cornelia Drut ¸u, and Lee Mosher. Thick metric spaces, relative hyperbolicity, and quasi-isometric rigidity.Math. Ann., 344(3):543– 595, 2009

work page 2009
[2]

The boundary of negatively curved groups

Mladen Bestvina and Geoffrey Mess. The boundary of negatively curved groups. Journal of the American Mathematical Society, 4(3):469–481, 1991

work page 1991
[3]

Relative homology and Poincar´ e duality for group pairs

Robert Bieri and Beno Eckmann. Relative homology and Poincar´ e duality for group pairs. Journal of Pure and Applied Algebra , 13(3):277–319, 1978

work page 1978
[4]

Relatively hyperbolic groups

Brian H Bowditch. Relatively hyperbolic groups. International Journal of Algebra and Computation , 22(03):1250016, 2012

2012
[5]

Metric spaces of non-positive cur- vature, volume 319

Martin R Bridson and Andr´ e Haefliger. Metric spaces of non-positive cur- vature, volume 319. Springer Science & Business Media, 2013

work page 2013
[6]

Cohomology of groups , volume 87

Kenneth S Brown. Cohomology of groups , volume 87. Springer Science & Business Media, 2012

work page 2012

Show all 27 references

[7]

Classifying spaces and boundaries for relatively hyper- bolic groups

Fran¸ cois Dahmani. Classifying spaces and boundaries for relatively hyper- bolic groups. Proceedings of the London Mathematical Society , 86(3):666– 684, 2003. 24

2003
[8]

Combination of convergence groups.Geometry & Topol- ogy, 7(2):933–963, 2003

Fran¸ cois Dahmani. Combination of convergence groups.Geometry & Topol- ogy, 7(2):933–963, 2003

2003
[9]

Relative hyperbolicity for automor- phisms of free products and free groups

Fran¸ cois Dahmani and Ruoyu Li. Relative hyperbolicity for automor- phisms of free products and free groups. Journal of Topology and Analysis, 14(01):55–92, 2022

2022
[10]

Relatively hyperbolic groups: geometry and quasi- isometric invariance

Cornelia Drut ¸u. Relatively hyperbolic groups: geometry and quasi- isometric invariance. Comment. Math. Helv. , 84(3):503–546, 2009

2009
[11]

Tree-graded spaces and asymptotic cones of groups

Cornelia Drut ¸u and Mark Sapir. Tree-graded spaces and asymptotic cones of groups. Topology, 44(5):959–1058, 2005

2005
[12]

Relatively hyperbolic groups

Benson Farb. Relatively hyperbolic groups. Geometric and functional anal- ysis, 8(5):810–840, 1998

1998
[13]

Relative hyperbolicity of free-by-cyclic extensions

Pritam Ghosh. Relative hyperbolicity of free-by-cyclic extensions. Compo- sitio Mathematica, 159(1):153–183, 2023

2023
[14]

Hyperbolic groups

Mikhael Gromov. Hyperbolic groups. In Essays in group theory , pages 75–263. Springer, 1987

1987
[15]

Dehn filling in relatively hyperbolic groups

Daniel Groves and Jason Fox Manning. Dehn filling in relatively hyperbolic groups. Israel Journal of Mathematics , 168:317–429, 2008

2008
[16]

A remark on thickness of free-by-cyclic groups

Mark Hagen. A remark on thickness of free-by-cyclic groups. Illinois J. Math., 63(4):633–643, 2019

2019
[17]

Homological dimension and critical exponent of Kleinian groups

Michael Kapovich. Homological dimension and critical exponent of Kleinian groups. Geometric and Functional Analysis , 18(6):2017–2054, 2009

2017
[19]

Detour functions and quasi-isometries

Nataˇ sa Macura. Detour functions and quasi-isometries. The Quarterly Journal of Mathematics , 53(2):207–239, 2002

2002
[21]

Straightening and bounded cohomology of hyperbolic groups

Igor Mineyev. Straightening and bounded cohomology of hyperbolic groups. Geometric & Functional Analysis GAF A, 11(4):807–839, 2001

2001
[22]

Quasi-conformal mappings in n-space and the rigidity of hyperbolic space forms

George D Mostow. Quasi-conformal mappings in n-space and the rigidity of hyperbolic space forms. Publications Math´ ematiques de l’IH´ES, 34:53–104, 1968

1968
[23]

Strong rigidity of locally symmetric spaces

Grigory A Mostow. Strong rigidity of locally symmetric spaces. Number 78. Princeton University Press, 1973. 25

1973
[24]

Relatively Hyperbolic Groups: Intrinsic Geometry, Alge- braic Properties, and Algorithmic Problems: Intrinsic Geometry, Algebraic Properties, and Algorithmic Problems , volume 843

Denis V Osin. Relatively Hyperbolic Groups: Intrinsic Geometry, Alge- braic Properties, and Algorithmic Problems: Intrinsic Geometry, Algebraic Properties, and Algorithmic Problems , volume 843. American Mathemati- cal Soc., 2006

2006
[25]

Strong rigidity of Q-rank 1 lattices

Gopal Prasad. Strong rigidity of Q-rank 1 lattices. Inventiones mathemat- icae, 21(4):255–286, 1973

1973
[26]

Volumes of discrete groups and topological complexity of homology spheres

Alexander Reznikov. Volumes of discrete groups and topological complexity of homology spheres. arXiv preprint dg-ga/9506010 , 1995

1995 arXiv
[27]

Complexity volumes of splittable groups

Mihalis Sykiotis. Complexity volumes of splittable groups. Journal of Algebra, 503:409–432, 2018. Department of Mathematics, Technion – Israel Institute of Technology, Haifa, Israel E-mail address: lazarovich@technion.ac.il Department of Mathematics, Technion – Israel Institut...

2018

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [18]

Finite index rigidity of hyperbolic groups

Nir Lazarovich. Finite index rigidity of hyperbolic groups. arXiv preprint arXiv:2302.04484, 2023

work page Pith review arXiv 2023

[2] [20]

Cohomology and the Bowditch boundary

Jason F Manning and Oliver H Wang. Cohomology and the Bowditch boundary. Michigan Mathematical Journal , 69(3):633–669, 2020

work page 2020

[3] [1]

Thick metric spaces, relative hyperbolicity, and quasi-isometric rigidity.Math

Jason Behrstock, Cornelia Drut ¸u, and Lee Mosher. Thick metric spaces, relative hyperbolicity, and quasi-isometric rigidity.Math. Ann., 344(3):543– 595, 2009

work page 2009

[4] [2]

The boundary of negatively curved groups

Mladen Bestvina and Geoffrey Mess. The boundary of negatively curved groups. Journal of the American Mathematical Society, 4(3):469–481, 1991

work page 1991

[5] [3]

Relative homology and Poincar´ e duality for group pairs

Robert Bieri and Beno Eckmann. Relative homology and Poincar´ e duality for group pairs. Journal of Pure and Applied Algebra , 13(3):277–319, 1978

work page 1978

[6] [4]

Relatively hyperbolic groups

Brian H Bowditch. Relatively hyperbolic groups. International Journal of Algebra and Computation , 22(03):1250016, 2012

2012

[7] [5]

Metric spaces of non-positive cur- vature, volume 319

Martin R Bridson and Andr´ e Haefliger. Metric spaces of non-positive cur- vature, volume 319. Springer Science & Business Media, 2013

work page 2013

[8] [6]

Cohomology of groups , volume 87

Kenneth S Brown. Cohomology of groups , volume 87. Springer Science & Business Media, 2012

work page 2012