Isomorphism of \'etale fundamental groups lifts to isomorphism of stratified fundamental group

A.J. Parameswaran; Indranil Biswas; Manish Kumar

REVIEW 2 major objections 5 minor 23 references

Isomorphism of \'etale fundamental groups lifts to isomorphism of stratified fundamental group

T0 review · 2 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read If a finite generically smooth morphism between smooth projective varieties over an algebraically closed field of characteristic $p>0$ induces an isomorphism of étale fundamental groups, then it also induces an isomorphism of stratified…

desk verdict New relative Gieseker result, but the proof has a false divisor claim in Section 6 that needs a repaired ordering argument. read the letter →

arxiv 2506.03829 v1 pith:3GVN7QCT submitted 2025-06-04 math.AG

Indranil Biswas , Manish Kumar , A.J. Parameswaran This is my paper

classification math.AG MSC 14H3014G1714H6014J60

keywords genuineramificationfundamentalgroupstratifiedbundlesétaleGiesekerconjecturepositivecharacteristicF-divisiblevectorrepresentationspaces

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

This paper proves a relative version of Gieseker's conjecture: if a finite generically smooth morphism $f: Y \to X$ between smooth projective varieties over an algebraically closed field of characteristic $p > 0$ induces an isomorphism of étale fundamental groups, then the induced map of stratified fundamental groups is also an isomorphism. The stratified fundamental group is a finer pro-algebraic invariant than the étale one, so the result says these two invariants cannot diverge under this hypothesis. The proof works by showing that the pullback functor between the Tannakian categories of stratified bundles on $X$ and $Y$ is an equivalence of categories, not merely that certain stratified bundles are trivial. A key step is a hom-vanishing statement for semistable vector bundles pulled back along genuinely ramified covers, together with a density argument for periodic points and a reduction to the field $\mathbb{F}_p$ via spreading out.

What carries the argument

The argument takes place in the Tannakian category of stratified vector bundles, presented as $F$-divisible vector bundles: sequences $\{E_i\}_{i\ge 0}$ of vector bundles with isomorphisms $E_i \cong F_X^* E_{i+1}$ for absolute Frobenius $F_X$, whose Tannaka dual is the stratified fundamental group. The load-bearing mechanism is Proposition 2.3, a hom-vanishing statement asserting that for a genuinely ramified map (one inducing a surjection on étale fundamental groups), the natural map $\mathrm{Hom}_X(V,W)\longrightarrow \mathrm{Hom}_Y(f^*V,f^*W)$ is an isomorphism for semistable bundles $V,W$ of the same slope. This is proved by constructing negative line bundles $L_j \subsetneq \mathcal{O}_Y$ inside the purely inseparable part of $f_*\mathcal{O}_Y$, using an ordering of the Galois group and the claim that intersections of components of $Y\times_X Y$ are divisors. Around this, the proof uses representation spaces of semistable bundles with trivialized fiber, the density lemma (Lemma 2.2) for periodic points of a rational dominant self-map, and a spreading-out theorem (Theorem 4.2) that reduces the general case to the field $\mathbb{F}_p$.

What would settle it

One concrete test is to construct a projective genuinely ramified Galois cover $Y\to X$ with smooth $X,Y$ where some automorphism has an isolated fixed point, then compute the intersections of the components of $Y\times_X Y$; if any such intersection is not a divisor, the paper's Remark 6.3 fails in that case, and checking whether $\pi_1^{\mathrm{str}}(f)$ is still an isomorphism would settle whether the theorem's statement needs modification.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1. Let $f\colon Y\longrightarrow X$ be a finite generically smooth morphism of smooth projective varieties over an algebraically closed field $k$ of characteristic $p>0$ such that the induced homomorphism $\pi_1^{\mathrm{et}}(f)\colon \pi_1^{\mathrm{et}}(Y,y)\longrightarrow \pi_1^{\mathrm{et}}(X,f(y))$ is an isomorphism. Then the induced homomorphism $\pi_1^{\mathrm{str}}(f)\colon \pi_1^{\mathrm{str}}(Y)\longrightarrow \pi_1^{\mathrm{str}}(X)$ of stratified fundamental groups is also an isomorphism. This generalizes Gieseker's conjecture, which is recovered by taking $X = \mathbb{P}^d$ with a normalization morphism. The paper also notes that genuine ramification alone (surjectivity on étale fundamental groups) suffices to conclude surjectivity of the stratified fundamental group homomorphism.

Load-bearing premise

The proof of Proposition 2.3 rests on Remark 6.3, which asserts that for a genuinely ramified Galois cover with smooth $X$ and $Y$, the intersection $Y_\gamma \cap Y_\delta$ of two components of the fiber product $Y\times_X Y$ is a divisor on $Y$, justified by purity of the branch locus; this can fail when an automorphism's fixed locus has codimension at least two.

Editorial extensions

If this is right

Gieseker's conjecture follows as a direct consequence: if $\pi_1^{\mathrm{et}}(X)$ is trivial, a normalization morphism $X\to \mathbb{P}^d$ has trivial étale fundamental group on both sides, so Theorem 1.1 forces $\pi_1^{\mathrm{str}}(X)$ to be trivial.
When the étale fundamental group homomorphism is only assumed surjective (the genuinely ramified case), the same proof shows that the stratified fundamental group homomorphism is surjective as well (Remark 5.1).
The pullback functor between the categories of stratified bundles on $X$ and $Y$ is fully faithful for every genuinely ramified map, giving a standalone tool for transferring homomorphisms of semistable bundles along ramified covers.
For a morphism spread out over a connected base, Theorem 4.2 shows that being genuinely ramified, or inducing an étale fundamental group isomorphism, is an open condition on the base; this specialization result is what makes the reduction to $\mathbb{F}_p$ legitimate.

Reading between the lines

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

The divisor assertion in Remark 6.3 is the most delicate point: purity of the branch locus guarantees that the total branch locus of a finite map with smooth target is a divisor, but it does not automatically make the fixed locus of every individual automorphism a divisor. A projective action whose quotient is smooth while a non-identity element has an isolated fixed point would be a natural test
One could try to extend the theorem to finite generically smooth morphisms with controlled wild ramification, where the étale fundamental group is replaced by a tame fundamental group; the stratification machinery might adapt if the appropriate purity statement holds.
The reduction to $\mathbb{F}_p$ suggests a template for other Tannakian invariants built from Frobenius-divisible structures: any invariant whose objects are semistable bundles with Frobenius structure and whose moduli spaces carry a dominant Verschiebung could be controlled by the same density argument.
It is worth testing whether the theorem remains true without the smoothness of $Y$ or $X$; the proof uses smoothness in the divisor purity step, so a singular compactification of the $S_3$ example might separate the theorem from its current proof.

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

The paper proves a genuine new step on Gieseker's conjecture: if a finite generically smooth map between smooth projective varieties in characteristic p induces an isomorphism on étale fundamental groups, then it does so on stratified fundamental groups. That is a real improvement over the triviality results of Esnault-Mehta, Esnault-Srinivas, and Sun, and the proof strategy—representation spaces, Verschiebung, spreading out to F_p—is sensible.

The main result is likely true, but the proof as written has a load-bearing gap in Section 6. Lemma 6.4 constructs line bundles from the intersections Y_{γ_j} ∩ Y_{γ_{η(j)}} in the fiber product. Remark 6.3 asserts these intersections are divisors, citing purity of the branch locus. That is false in general. For example, with Y = (P^1)^3 and Γ = S_3 acting by permuting factors, the quotient is P^3 and the map is genuinely ramified, but a 3-cycle has fixed locus the diagonal, a curve of codimension 2. So Y_e ∩ Y_{(123)} is not a divisor. The proof needs an ordering argument: because the quotient is smooth, Γ is generated by pseudo-reflections, so one can order the elements so each ratio γ_j^{-1}γ_{η(j)} is a pseudo-reflection and the relevant intersections are divisors. No such argument is given. Since Lemma 6.5 and Proposition 2.3 rest on this, the gap is central, though probably repairable.

There are also smaller issues. Several lemmas are stated to be 'exactly identical' to proofs in [BP] for curves without spelling out the higher-dimensional verification. Lemma 6.1 as written has a type mismatch—the pullback of a torsion sheaf cannot embed into O_X^{⊕(d-1)}. These are fixable but need attention.

Net: the theorem is worth refereeing. A competent referee could work through the details and likely fill the gap. As submitted, I would not cite it, but I would ask for revision.

Referee Report

2 major / 5 minor

Summary. The paper proves Theorem 1.1: if f: Y → X is a finite generically smooth morphism of smooth projective varieties over an algebraically closed field of characteristic p > 0 and the induced homomorphism of étale fundamental groups is an isomorphism, then the induced homomorphism of stratified fundamental groups is an isomorphism. The proof works by showing that the pullback functor on stratified bundles is an equivalence. The main ingredients are a Hom-vanishing result (Proposition 2.3) for genuinely ramified maps, proved through Galois covers and negative line bundles; a reduction to the case k = F_p using representation spaces and Hrushovski's density theorem; and a specialization theorem for étale fundamental groups. The authors also observe that their theorem implies the Gieseker conjecture on triviality of stratified fundamental groups.

Significance. If correct, the result is a genuine relative generalization of the Esnault–Mehta theorem and the Esnault–Srinivas relative result, and it strengthens the conclusion from triviality of stratified bundles to an equivalence of Tannakian categories. The strategy of proving essential surjectivity of the stratified pullback functor, rather than only conservation of triviality, is valuable. The paper also contains a useful specialization theorem (Theorem 4.2) for genuine ramification. However, the central Hom-vanishing proposition rests on a false geometric assertion in Section 6, so the main theorem is not established as written. The gap appears local and likely repairable, but it is load-bearing.

major comments (2)

[Section 6, Remark 6.3 and Lemma 6.4] The assertion in Remark 6.3 that every nonempty intersection Y_{γ_j} ∩ Y_{γ_{η(j)}} is a divisor on Y is false in general. Let Y = (P^1)^3 and let Γ = S_3 act by permuting the factors, so X = Y/Γ ≅ P^3 is smooth and f is finite generically smooth; since the étale fundamental groups of both (P^1)^3 and P^3 are trivial, f is genuinely ramified. For the 3-cycle σ = (123), the fixed locus is the diagonal {x = y = z} ≅ P^1, of codimension 2 in Y, so Y_e ∩ Y_σ is not a divisor. Purity of the branch locus constrains the branch locus of the quotient map, not the fixed loci of individual group elements. This matters because Lemma 6.4 defines L_i = O_Y(-D_i^0) with D_i^0 required to be an effective Cartier divisor; if D_i^0 has codimension at least two, the line bundle is not defined. Since Lemma 6.5 and Proposition 2.3 depend on the construction of these negative line bundles, the proof of Theorem 1.1 has a serious gap at this point.
[Section 6, Lemma 6.2 and Lemma 6.4] The proof of Lemma 6.2 is said to be identical to [BP, Lemma 3.4] and to use only the connectedness of Y ×_X Y. In dimension one, any nonempty intersection of two components is zero-dimensional and hence gives a divisor on the curve, but in higher dimension connectedness gives no information on the codimension of the intersection. The paper needs a stronger ordering statement: for each j, the element (γ_j)^{-1}γ_{η(j)} should be chosen to have divisorial fixed locus, or the proof must otherwise show that the particular pairs supplied by the ordering avoid fixed loci of codimension at least two. No such argument is given, and the smoothness of X and Y does not by itself imply it, as the example in the previous comment shows.

minor comments (5)

[Section 2] In the paragraph after the fiber functor, the induced homomorphism is written as π_et^1(f) : π_str^1(Y,y) → π_str^1(X,x); this should be π_str^1(f).
[Section 6, Lemma 6.4] The text says 'the lie bundle on Y' where it should say 'the line bundle on Y'.
[Section 6, proof of Proposition 2.3] The sequence displayed in (6.22) is written as a short exact sequence, but Hom(V,–) is only left exact; the later deduction only needs left exactness, so the display should be corrected to avoid claiming surjectivity at the final term.
[Section 3, proof of Theorem 3.1] The notation π_et(f) : Vect_et(X) → Vect_et(Y) conflates the pullback functor with fundamental-group notation; a notation such as f^* would be clearer.
[Section 6, Lemma 6.5] The proof of Lemma 6.5 is omitted on the grounds that it is identical to [BP, Lemma 4.1]; in view of the higher-dimensional issues discussed above, a short indication of the slope computation would help the reader verify the claim.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: Theorem 1.1 is proved from prior published lemmas, not from its own conclusion; score 2 reflects heavy self-citation rather than a circular derivation.

full rationale

The central claim, Theorem 1.1, is not derived by assuming the stratified fundamental group isomorphism. The proof reduces the statement to the F_p case via Theorem 4.2, then uses Theorem 3.1 and Corollary 3.2. The engine is Proposition 2.3, proved in Section 6. Although the proof of Proposition 2.3 imports several lemmas verbatim from the authors' earlier papers [BP] and [BDP] (Lemmas 6.1, 6.2, 6.4, 6.5 and the connectedness input), these are published results with independent derivations and with assumptions that do not include Theorem 1.1. Under the stated rules, such self-citations count as independent support, not circularity. The main unproved new structural assertion is Remark 6.3, which claims that Y_{γ_j} ∩ Y_{γ_{η(j)}} is a divisor by purity of the branch locus; the reviewer's counterexample (e.g. fixed loci of codimension 2) indicates a possible rigor gap in Lemma 6.4 and therefore in Proposition 2.3. This is a correctness risk, not a circular reduction: the paper does not define the divisor claim in terms of the desired isomorphism, nor does it fit any parameter to the conclusion. No fitted input is renamed as a prediction, and no uniqueness theorem is imported to forbid alternatives. The result is self-contained in the sense that the conclusion is not equivalent to the hypothesis by construction; the score is set to 2 only to acknowledge the unusually heavy reliance on the authors' own prior work, not because the derivation is circular.

Assumptions & free parameters 0 free parameters · 10 assumptions · 0 invented entities

The paper introduces no fitted parameters and no new entities. The central claim rests on a collection of standard results in Tannakian categories, moduli of vector bundles, and positive-characteristic algebraic geometry, plus several cited results from the authors' own prior work. The most fragile axioms are the structural assertions about intersections in the fiber product (Remark 6.3) and the unproved non-Galois reduction in Lemma 6.5.

assumptions (10)

standard math Equivalence between stratified vector bundles and F-divisible vector bundles
Used throughout; due to Gieseker [Gi] and dos Santos [Sa], cited in Section 2.
standard math Neutral Tannakian duality for stratified bundles yields the stratified fundamental group scheme
Section 2, from [Sa], [Gi], [DM].
standard math Existence of fine moduli spaces R(X, xi, P) of semistable bundles with trivialization
Section 2, [Sun, Theorem 2.3], a positive-characteristic generalization of Simpson.
domain assumption A Verschiebung rational map exists on representation spaces and is dominant on the closure of the Frobenius orbit
Used in the proof of Theorem 3.1; cited from [Sun, Proposition 2.5].
domain assumption Hrushovski's theorem: periodic points of a dominant rational self-map over F_p are dense
Lemma 2.2, cited from [Hr, Corollary 1.2]; used in Theorem 3.1 to find periodic points.
domain assumption Lange-Stuhler theorem: if (F^j)^*E is isomorphic to E for some j >= 1, then E is etale trivial
Used in Theorem 3.1 proof; cited from [LS].
domain assumption Connectedness of Y x_X Y for genuinely ramified maps in arbitrary dimension
Used in Lemma 6.2; cited from [BDP, Theorem 2.4(3)], a paper by two of the current authors.
ad hoc to paper For a genuinely ramified Galois map with smooth X and Y, the intersections Y_gamma and Y_delta are divisors
Remark 6.3 asserts this from purity of branch locus; it is a key geometric input for Lemma 6.4 and is not fully proven in the paper.
ad hoc to paper Lemma 6.5 in the non-Galois case follows exactly as in [BP, Lemma 4.1]
The proof of Lemma 6.5 refers to [BP] for the reduction from Galois to general genuinely ramified maps; the reduction is not detailed here.
standard math Specialization of etale fundamental groups and purity of branch locus (SGA1)
Used in Proposition 4.1 and in the proof of Theorem 4.2.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Isomorphism of \'etale fundamental groups lifts to isomorphism of stratified fundamental group." pith.science (2026). https://pith.science/paper/3GVN7QCT

@misc{pith2026250603829,
  author       = {Pith},
  title        = {Pith review of: Isomorphism of \'etale fundamental groups lifts to isomorphism of stratified fundamental group},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3GVN7QCT}},
  note         = {Machine review of arXiv:2506.03829}
}

abstract

It is shown that if a finite generically smooth morphism $f\,:\,Y\,\longrightarrow\, X$ of smooth projective varieties induces an isomorphism of the \'etale fundamental groups, then the induced map of the stratified fundamental groups $\pi_1^{str}(f)\, :\, \pi_1^{str}(Y,\, y)\,\longrightarrow\, \pi_1^{str}(X,\, f(y))$ is also an isomorphism.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

23 extracted references · 22 canonical work pages

[1]

Berthelot and A

P. Berthelot and A. Ogus, Notes on crystalline cohomology , Princeton University Press, Princeton NJ, University of Tokyo Press, Tokyo, 1978

work page 1978
[2]

Biswas and A

I. Biswas and A. J. Parameswaran, Ramified covering maps and stability of pulled back bundles, Int. Math. Res. Not. , (2022), no. 17, 12821--12851

work page 2022
[3]

Biswas, S

I. Biswas, S. Das, and A. J. Parameswaran, Genuinely ramified maps and stable vector bundles, Internat. J. Math. 33 (2022), Paper No. 2250039, 14

work page 2022
[4]

Deligne and J

P. Deligne and J. S. Milne, Tannakian Categories, Hodge cycles, motives, and Shimura varieties , by P. Deligne, J. S. Milne, A. Ogus and K.-Y. Shih, pp. 101--228, Lecture Notes in Mathematics, 900, Springer-Verlag, Berlin-Heidelberg-New York, 1982

work page 1982
[5]

Esnault and V

H. Esnault and V. Mehta, Simply connected projective manifolds in characteristic p>0 have no nontrivial stratified bundles, Invent. Math. 181 (2010), 449--465

work page 2010
[6]

Esnault and V

H. Esnault and V. Srinivas, A relative version of Gieseker's problem on stratifications in characteristic p>0 , Inter. Math. Res. Not. (2019), 5635--5648

work page 2019
[7]

Gieseker, Flat vector bundles and the fundamental group in non-zero characteristics, Ann

D. Gieseker, Flat vector bundles and the fundamental group in non-zero characteristics, Ann. Scuola Norm. Sup. Pisa Cl. Sci. 2 (1975), 1--31

work page 1975
[8]

Grothendieck, \'El\'ements de g\'eom\'etrie alg\'ebrique

A. Grothendieck, \'El\'ements de g\'eom\'etrie alg\'ebrique. IV. \'Etude locale des sch\'emas et des morphismes de sch\'emas IV, Inst. Hautes \'Etudes Sci. Publ. Math. 32 (1967), 5--361

work page 1967

Show all 23 references

[9]

Hartshorne, Algebraic geometry , Graduate Texts in Mathematics, No

R. Hartshorne, Algebraic geometry , Graduate Texts in Mathematics, No. 52. Springer-Verlag, New York-Heidelberg, 1977

1977
[10]

Hrushovski, The elementary theory of the Frobenius automorphism, preprint, available at http://www.ma.huji.ac.il/ ehud/FROB.pdf

E. Hrushovski, The elementary theory of the Frobenius automorphism, preprint, available at http://www.ma.huji.ac.il/ ehud/FROB.pdf
[11]

Lange and U

H. Lange and U. Stuhler, Vektorb\"undel auf Kurven und Darstellungen der algebraischen Fundamentalgruppe, Math. Zeit. 156 (1977), 73--83

1977
[12]

J. P. Murre, Lectures on an introduction to G rothendieck's theory of the fundamental group, Notes by S. Anantharaman, Tata Institute of Fundamental Research Lectures on Mathmatics, No 40, 1967

1967
[13]

Nitsure, Schematic Harder-Narasimhan stratification, Internat

N. Nitsure, Schematic Harder-Narasimhan stratification, Internat. J. Math. 22 (2011), 1365--1373

2011
[14]

M. V. Nori, On the representations of the fundamental group, Compositio Math. 33 (1976), 29--41

1976
[15]

M. V. Nori, The fundamental group scheme, Proc. Indian Acad. Sci. Math. Sci. 91 (1982), 73--122

1982
[16]

Grothendieck, Rev\^etements \'etales et groupe fondamental , SGA 1

A. Grothendieck, Rev\^etements \'etales et groupe fondamental , SGA 1
[17]

J. P. P. dos Santos, Fundamental group schemes for stratified sheaves, J. Algebra 317 (2007), 691--713

2007
[18]

Saavedra Rivano, Cat\'egories Tannakiennes , Lecture Notes in Mathematics, Vol

N. Saavedra Rivano, Cat\'egories Tannakiennes , Lecture Notes in Mathematics, Vol. 265, Springer-Verlag, Berlin, Heidelberg, 1972

1972
[19]

Shatz, The decomposition and specialization of algebraic families of vector bundles, Compositio Math

S. Shatz, The decomposition and specialization of algebraic families of vector bundles, Compositio Math. 35 (1977), 163--187

1977
[20]

Simpson, Moduli of representations of the fundamental groups of a smooth projective variety I, Inst

C.T. Simpson, Moduli of representations of the fundamental groups of a smooth projective variety I, Inst. Hautes \'Etudes Sci. Publ. Math. 79 (1994), 47--129

1994
[21]

Stacks Project , http://stacks.math.columbia.edu/
[22]

Sun, Stratified bundles and representation spaces , Adv

X. Sun, Stratified bundles and representation spaces , Adv. Math. 345 (2019), 767--783

2019
[23]

Varshavsky, Intersection of a correspondence with a graph of Frobenius, J

Y. Varshavsky, Intersection of a correspondence with a graph of Frobenius, J. Algebraic Geom. 27 (2018), 1--20

2018

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Berthelot and A

P. Berthelot and A. Ogus, Notes on crystalline cohomology , Princeton University Press, Princeton NJ, University of Tokyo Press, Tokyo, 1978

work page 1978

[2] [2]

Biswas and A

I. Biswas and A. J. Parameswaran, Ramified covering maps and stability of pulled back bundles, Int. Math. Res. Not. , (2022), no. 17, 12821--12851

work page 2022

[3] [3]

Biswas, S

I. Biswas, S. Das, and A. J. Parameswaran, Genuinely ramified maps and stable vector bundles, Internat. J. Math. 33 (2022), Paper No. 2250039, 14

work page 2022

[4] [4]

Deligne and J

P. Deligne and J. S. Milne, Tannakian Categories, Hodge cycles, motives, and Shimura varieties , by P. Deligne, J. S. Milne, A. Ogus and K.-Y. Shih, pp. 101--228, Lecture Notes in Mathematics, 900, Springer-Verlag, Berlin-Heidelberg-New York, 1982

work page 1982

[5] [5]

Esnault and V

H. Esnault and V. Mehta, Simply connected projective manifolds in characteristic p>0 have no nontrivial stratified bundles, Invent. Math. 181 (2010), 449--465

work page 2010

[6] [6]

Esnault and V

H. Esnault and V. Srinivas, A relative version of Gieseker's problem on stratifications in characteristic p>0 , Inter. Math. Res. Not. (2019), 5635--5648

work page 2019

[7] [7]

Gieseker, Flat vector bundles and the fundamental group in non-zero characteristics, Ann

D. Gieseker, Flat vector bundles and the fundamental group in non-zero characteristics, Ann. Scuola Norm. Sup. Pisa Cl. Sci. 2 (1975), 1--31

work page 1975

[8] [8]

Grothendieck, \'El\'ements de g\'eom\'etrie alg\'ebrique

A. Grothendieck, \'El\'ements de g\'eom\'etrie alg\'ebrique. IV. \'Etude locale des sch\'emas et des morphismes de sch\'emas IV, Inst. Hautes \'Etudes Sci. Publ. Math. 32 (1967), 5--361

work page 1967