On the K-theory of algebraic tori

Branko Juran; Florian Riedel; Qingyuan Bai; Shachar Carmeli

REVIEW 2 major objections 4 minor 13 references

On the K-theory of algebraic tori

T0 review · 2 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The algebraic K-theory of any torus is the equivariant homology of a topological torus built from its character lattice.

desk verdict A substantial, likely correct computation of K-theory for all algebraic tori via a motivic Fourier transform; the proof is coherent, with a few spots that need tightening rather than any visible fatal flaw. read the letter →

arxiv 2507.12954 v1 pith:UOPQ4K3B submitted 2025-07-17 math.KT math.AGmath.AT

Qingyuan Bai , Shachar Carmeli , Branko Juran , Florian Riedel This is my paper · ORCID

classification math.KTmath.AGmath.AT MSC 14F4219D4519L4714L10

keywords algebraicK-theorytorimotivichomotopytheoryFouriertransformequivarianttopologicalmirrorArtin-TatemotivesEilenberg-Mooreformula

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 that the algebraic K-theory of any algebraic torus $T$ over a field $F$ is naturally isomorphic to the equivariant homology of a topological torus $\mathcal{T}(T)$, the quotient of the character lattice's real span by the lattice itself, with coefficients in the $G$-equivariant K-theory spectrum of $F$. This turns the computation of K-groups of tori, including non-split ones, into a topological homology calculation and generalizes the fundamental theorem for the multiplicative group, $K_*(G_m) \simeq K_*(F) \oplus K_{*-1}(F)$, as well as the previously known presentation of $K_0(T)$. The proof works in motivic homotopy theory, where a naturally constructed motivic Fourier transform is shown to be an equivalence of commutative $K_F$-algebras, and the result is then transported to genuine equivariant spectra by an assembly argument.

What carries the argument

The load-bearing object is the motivic Fourier transform, a natural map of commutative $K_F$-algebras from the K-homology of the étale delooping of the character lattice to the $K_F$-motive of the torus. It is assembled from a parametrized Fourier transform formalism and a categorical Fourier--Mukai transform for categories of perfect complexes, and it is shown to be an equivalence by means of a motivic Eilenberg--Moore formula. The geometric heart of that formula is the Artin--Tate property of the zero section of the classifying stack of a Weil-restricted split torus, proved by induction over an equivariant cell decomposition of affine space.

What would settle it

Compute $K_1$ of a non-split torus, for example the norm-one torus of a non-Galois cubic extension of $\mathbb{Q}$, using the Atiyah--Hirzebruch spectral sequence of Corollary 5.11 and compare with the group obtained from a toric-variety presentation; any disagreement between the two would show the claimed equivalence is false.

Watch

Extended reading notes

Core claim

The central discovery is that a torus's character lattice completely determines its K-theory through a Fourier-type duality. The paper constructs, for any torus $T$ with character lattice $\Lambda$, a natural map $\mathfrak{F}^{\mathrm{mot}}_{\Sigma\Lambda(T)}: K_F[\Sigma\Lambda(T)] \to K^T_F$ between the $K_F$-homology of the étale delooping of the character lattice and the $K_F$-motive of $T$, and proves it is an equivalence of commutative $K_F$-algebras in the stable motivic category (Theorem 4.7). Applying the motivic-to-equivariant comparison functor, this yields a natural equivalence $K_G(F)[\mathcal{T}(T)] \cong K_G(T)$ of commutative $K_G(F)$-algebras (Theorem 5.8). In particular, the algebraic K-theory of $T$ identifies with the $G$-equivariant homology of the topological mirror $\mathcal{T}(T)$ with coefficients in $K_G(F)$, as graded rings with convolution product.

Load-bearing premise

The main equivalence depends on the Artin--Tate property of the zero section of the classifying stack of a Weil-restricted split torus, a strong cellularity condition verified by an equivariant cell decomposition of affine space; if that cellularity (or the normal-bundle condition feeding it) failed for some field, the Eilenberg--Moore formula would not hold and the theorem would not follow.

Editorial extensions

If this is right

K-groups of a torus are governed by an Atiyah--Hirzebruch spectral sequence $E^2_{p,q} = H^G_p(\mathcal{T}(T); \pi_q K_G(F))$ converging to $K_{p+q}(T)$.
The equivalence is one of $E_\infty$-algebras, so all multiplicative structure on $K_*(T)$ is encoded by the topological mirror and its convolution product.
The $K_0$ statement recovers the known presentation of $K_0$ of tori in terms of Picard groups of intermediate extensions, now as part of a uniform higher-K-theory computation.
The known computation of the K-theory of rank-one tori (norm-one quadrics) appears as the special case of the equivariant cell structure of the topological circle.

Reading between the lines

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

The same Fourier-strategy might compute K-theory of other group schemes, such as abelian varieties, once an Artin--Tate or cellularity statement for their classifying stacks is available; the paper's proof relies on the Weil resolution, which is special to tori.
Read constructively, Theorem 5.8 gives an explicit machine for higher K-groups: if the equivariant cohomology of the character lattice can be expressed through Galois cohomology of the lattice (as in the $K_0$ case), then $K_n(T)$ for all $n$ admits a concrete presentation in terms of fixed fields.
The topological-mirror formula suggests a Pontryagin-style duality for K-theory, so that the K-cohomology of a torus is the K-homology of its dual; testing this on tori over more general bases could provide a template for duality theorems in motivic homotopy theory.

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

Good paper. The headline result is real: they compute the full algebraic K-theory of every algebraic torus as G-equivariant homology of its topological mirror, generalizing Quillen's split case and recovering Merkurjev-Panin for K0. The motivic Fourier transform is genuinely new, and the proof strategy is coherent. I think the theorem is right.

What is new: they do not stop at K0. They produce a natural equivalence of commutative K_G(F)-algebras K_G(F)[T(T)] ≅ K_G(T), and identify the source with the motivic homology of the étale delooping. The reduction through the Weil resolution and the motivic Eilenberg-Moore formula is elegant. Proposition 3.13 and its motivic consequence Proposition 3.20 handle the character decomposition; Proposition 4.20 is the load-bearing geometric input, and the Artin-Tate dévissage by induction over an equivariant cell decomposition of affine space is plausible and in line with the Merkurjev-Panin argument.

Soft spots, in proportion. Proposition 3.13 is lightly sketched: after reducing to Gm, it asserts the infinite coproduct-to-product comparison works and then cites Moulinos for the equivalence Dqc(A)[Z] ≅ Dqc(BGm,A). Likely correct, but a referee should ask for a fuller justification of that comparison. Corollary 5.13 leaves the verification of the coend relations to the reader; since recovering Merkurjev-Panin is one of the advertised payoffs, those relations should be written out. Both are minor.

The stress-test concern about Proposition 4.20 condition (2) is the right place to look. The identification of the normal bundle of Z_{J_t} inside U_{k-1} with the pullback of the representation F^{I\J_t} is stated in one sentence. It is credible because the embedding is linear and equivariant, but the compatibility with the Galois action after Weil restriction, and specifically that the class lies in the image of alpha^*, is not spelled out. I do not see an actual failure; it reads as terseness, not a broken induction. Still, that single condition is worth a careful referee pass, since the rest of Theorem 4.7 is formal once it holds.

On circularity: none. The paper builds on the chromatic Fourier transform of Barthel-Carmeli-Schlank-Yanovski and on Martini-Wolf parametrized category theory; those are external frameworks, not the target result. The cited prior work covers K0 and the split/rank-one cases, and the full statement is new.

Bottom line: substantial, likely correct, clearly worth serious referee time. The referee should ask for details in 3.13 and 4.20 and for the missing relations in 5.13; those are revisions, not reasons to reject. I would cite it and bring it to our reading group.

Referee Report

2 major / 4 minor

Summary. The paper proves that for an algebraic torus T over a field F, the algebraic K-theory spectrum K(T) is naturally equivalent to the G-equivariant homology of the topological mirror T(T) with coefficients in the G-equivariant K-theory spectrum K_G(F), where G is the absolute Galois group of F. The proof proceeds motivically: the authors construct a motivic Fourier transform K_F[M] → K_F^{M^∨} and show that for M = ΣΛ(T) it is an equivalence, using a motivic Eilenberg–Moore formula for a Weil resolution of T. The main geometric input is an Artin–Tate dévissage for the section Spec(F)→BG_m^I. The paper also derives a spectral sequence computing K_*(T), recovers Merkurjev–Panin's K_0 formula, and reproduces Swan's computation for rank-one tori.

Significance. If the proof is correct, this is a substantial result: it gives a highly structured, ring-level computation of the algebraic K-theory of arbitrary algebraic tori as equivariant homology of a compact topological torus, generalizing both Quillen's fundamental theorem and Merkurjev–Panin's K_0 computation. The construction of the motivic Fourier transform and the Artin–Tate Eilenberg–Moore formalism are interesting in their own right and likely to be reused. The exposition is careful about the main dependencies, and the paper is explicit about the crucial geometric input. The main theorem is falsifiable and the framework is coherent; the remaining concerns are specific proof gaps rather than signs of a false central claim.

major comments (2)

[§3.2, Proposition 3.13] The proof of Proposition 3.13 contains the following assertion: 'Since in Catpf the map from the infinite coproduct to the infinite product is fully faithful and hence a monomorphism, the colimit over the discrete space Z in the sheaf category Γ_A(Mod_Onc_S) is computed sectionwise.' This is not justified and is load-bearing: it identifies the sections of Onc_S[Z] with the coproduct ∐_Z Dperf(A), which is then used to identify the Fourier transform with the category of Z-graded objects and hence to prove Proposition 3.20. Colimits in categories of sheaves of categories are not generally computed sectionwise; fully faithfulness of the coproduct-to-product map only shows the sectionwise coproduct is separated. The authors should either prove that the presheaf colimit is already a sheaf in this case or supply a reference for the exact statement being used.
[§4.2–4.3, Corollary 4.21 and Proposition 4.6] Corollary 4.21 states that the section s:S→BW is Artin over K_BW. By Definition 4.8, this means that s_♯s^*K_BW (equivalently s_♯K_S) is Artin in Mod_{K_BW}(SH(BW)). However, in the proof of Proposition 4.6 the object whose Artin property is needed in Corollary 4.19 is A = K^S_{BW} = s_*K_S. The text passes from s_♯K_S to s_*K_S without comment. These two objects are not formally identical; the identification would follow, for example, from a Bott-periodic version of Atiyah duality that kills the cotangent twist. Since the entire Eilenberg–Moore step depends on this identification, it should be stated and proved explicitly.

minor comments (4)

[§4.3, Proposition 4.20] In the verification of condition (2), the normal bundle is identified with the pullback of the representation F^{I\J_t} along Z_{J_t}//G^I_{F_t}→BG^I_{F_t}→BG^{I\J_t}_{F_t}. To match the hypothesis of Lemma 4.15, the text should explicitly say that this representation is inflated to a representation of G^I_m,F_t with trivial action on the J_t-factor, so its class lies in the image of α^*:K0(BG^I_{F_t})→K0(Z_{J_t}//G^I_{F_t}). The step is correct, but as written it is easy to misread as using a class only on the projected stack.
[§5.2, Proposition 5.7] The map (C_ns∘U_ns)(ΣΛ(T))→ΣΛ(T) is the counit of the adjunction C_ns⊣U_ns, not the unit; the terminology 'unit map' should be corrected.
[§3.3, Lemma 3.17] The proof of Lemma 3.17 says 'the right hand map is an equivalence by Atiyah duality Theorem 2.7 and the right hand map is an equivalence by Lemma 3.6'; the duplicated phrase should be corrected so that the two maps being identified are clear.
[§5.3, Corollary 5.13] The proof leaves to the reader the verification that the coend relations coincide with the generator/relation presentation of Merkurjev–Panin. Since the corollary is stated as a recovery of a known theorem, a short indication of how the coend relations specialize would make the argument self-contained.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central equivalence is assembled from independent motivic inputs; the only author-overlap citation, [Bar+24], is a framework analog and is not load-bearing for the torus theorem.

full rationale

The paper's main theorem is not obtained by assuming the result it proves. The motivic Fourier transform of Construction 3.18 is built from the categorical Fourier transform, whose split-torus case is delegated to the independent external theorem [Mou21, Thm. 4.1], and from an assembly argument using finite étale descent and Lemma 3.17. The Eilenberg–Moore step of Proposition 4.6 rests on Proposition 4.20, an Artin–Tate property of the zero section of BG_m^I that is proven by a self-contained double induction over an equivariant cell decomposition of affine space, with the normal-bundle condition checked explicitly. Corollary 4.21, Corollary 4.19, and Proposition 4.16 are formal consequences proved in the paper. The recovered results of Quillen, Merkurjev–Panin, and Swan are stated as consequences and are not used as hypotheses. The one citation with overlapping authorship, [Bar+24], is used only as a categorical template for the parametrized Fourier transform, and Proposition 3.2 is proved directly rather than imported as evidence for the torus computation. No step reduces by construction to its own input, and no fitted parameter is renamed as a prediction. The derivation chain is therefore self-contained with respect to circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the established machinery of motivic homotopy theory for stacks (Khan-Ravi), parametrized higher category theory (Martini-Wolf), the motivic-equivariant comparison (Bachmann-Hoyois), and the character decomposition of perfect complexes on BG_m (Moulinos). No free parameters or invented entities are introduced. The paper's own geometric inputs, such as the Artin-Tate property of Weil-restricted tori, are proved rather than assumed.

assumptions (5)

standard math Stable motivic homotopy theory of nicely scalloped stacks, including six-functor formalism and motivic Thom spectra, as developed by Khan and Ravi [KR24].
The paper builds SH(X) for stacks and uses projection formulas, Atiyah duality, Bott periodicity, and descent from [KR24] throughout Sections 2-4.
standard math Parametrized higher category theory of Martini and Wolf [MW24, MW25], including parametrized Kan extensions, presentability, and the group algebra construction.
Section 2.4 and Section 3.1 rely on this formalism for the Fourier transform.
standard math The comparison between motivic and equivariant homotopy theory, in particular the adjunction C: Sp_G ⇄ SH(F) of Bachmann-Hoyois [BH21, Prop. 10.6].
Proposition 2.14 and Definition 2.15 use this to define K_G(F) and to pass from Theorem B to Theorem A.
domain assumption The character decomposition of Dperf(BG_{m,A}) as Z-graded complexes, cited from Moulinos [Mou21, Theorem 4.1].
Used in the proof of Proposition 3.13 to verify the categorical Fourier transform for the multiplicative group.
standard math The norm map framework of Cnossen-Lenz-Linskens [CLL25] for finite étale maps.
Invoked in Lemma 3.6 and Lemma 3.17 to identify assembly maps for finite étale morphisms.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the K-theory of algebraic tori." pith.science (2026). https://pith.science/paper/UOPQ4K3B

@misc{pith2026250712954,
  author       = {Pith},
  title        = {Pith review of: On the K-theory of algebraic tori},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UOPQ4K3B}},
  note         = {Machine review of arXiv:2507.12954}
}

abstract

Given an algebraic torus $T$ over a field $F$, its lattice of characters $\Lambda$ gives rise to a topological torus $\mathfrak{T}(T)=\Lambda_{\mathbb R}/\Lambda$ with a continuous action of the absolute Galois group $G$. We construct a natural equivalence between the algebraic $K$-theory $K_{\ast}(T)$ and the equivariant homology $H^{G}_{\ast}(\mathfrak{T}(T);K_G(F))$ of the topological torus $\mathfrak{T}(T)$ with coefficients in the $G$-equivariant $K$-theory of $F$. This generalizes a computation of $K_0(T)$ due to Merkurjev and Panin. We obtain this equivalence by analyzing the motive $\mathbb{K}_{F}^{T}$ in the stable motivic category $\mathrm{SH}(F)$ of Voevodsky and Morel, where $\mathbb{K}_{F}$ is the motivic spectrum representing homotopy $K$-theory. We construct a natural comparison map $\mathfrak{F}\colon \mathbb{K}_{F}[B\Lambda] \to \mathbb{K}_{F}^{T}$ from the $\mathbb{K}_{F}$-homology of the \'etale delooping of $\Lambda$ to $\mathbb{K}_{F}^{T}$ as a special case of a motivic Fourier transform and prove that it is an equivalence by using a motivic Eilenberg--Moore formula for classifying spaces of tori.

Figures

Figures reproduced from arXiv: 2507.12954 by the authors.

**Figure 1.** The equivariant cell structure of T(Ta) We thus have a cell-attachment cofiber sequence G/G0 ⊗ SG −−→ SG ⊕ SG −−→ SG ⊗ T(Ta) ∈ SpG. Tensoring with the equivariant K-theory spectrum of F we get a cofiber sequence KG(F) ⊗ G/G0 −−→ KG(F) ⊕ KG(F) −−→ KG(F)[T(T − a)] ∈ SpG. By Theorem 5.8, the G-fixed points of the last term are given by K(Ta). Since taking fixed points preserves cofiber sequences and (KG(F) ⊗ G/G0) G ≃ … view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

13 extracted references · 10 canonical work pages

[5]

arXiv:2503.09928 [math.AG]. [HA] J. Lurie. Higher algebra

work page Pith review arXiv
[7]

Standard $t$-structures

arXiv:2504. 07473 [math.CT]. url: https://arxiv.org/abs/2504.07473. [HTT] J. Lurie. Higher topos theory. Vol

work page Pith review arXiv
[9]

[Mou21] T

arXiv:2103.17141 [math.CT]. [Mou21] T. Moulinos. The geometry of filtrations. In: Bull. Lond. Math. Soc. 53.5 (2021), pp. 1486–1499. 44 [MP97] A. S. Merkurjev and I. A. Panin. K-theory of algebraic tori and toric varieties. In: K-Theory 12.2 (1997), pp. 101–143. [MV99] F. Morel and V. Voevodsky. A1-homotopy theory of schemes. In:Inst. Hautes Études Sci. P...

arXiv 2021
[10]

[Qui72] D

arXiv: 2209.05103 [math.CT]. [Qui72] D. Quillen. On the cohomology andK-theory of the general linear groups over a finite field. In:Ann. of Math. (2)96 (1972), pp. 552–586. [Qui73] D. Quillen. Higher algebraicK-theory. I. In:AlgebraicK-theory, I: HigherK-theories (Proc. Conf., Battelle Memorial Inst., Seattle, Wash., 1972). Vol. Vol

arXiv 1972
[88]

Progr. Math. Birkhäuser Boston, Boston, MA, 1990, pp. 247–435. 45

work page 1990
[96]

[Bar17] C. Barwick. Spectral Mackey functors and equivariant algebraicK-theory (I). In:Adv. Math. 304 (2017), pp. 646–727. [BGT13] A. J. Blumberg, D. Gepner, and G. Tabuada.A universal characterization of higher algebraicK-theory. In:Geom. Topol.17.2 (2013), pp. 733–838. [BH21] T. Bachmann and M. Hoyois. Norms in motivic homotopy theory. In:Astérisque 425...

work page 2017
[170]

Princeton University Press, Princeton, NJ, 2009, pp

Annals of Mathematics Studies. Princeton University Press, Princeton, NJ, 2009, pp. xviii+925. [KR24] A. A. Khan and C. Ravi. Generalized cohomology theories for algebraic stacks. In: Adv. Math.458 (2024), Paper No. 109975, 104.issn: 0001-8708,1090-2082. [Mar22] L. Martini. Yoneda’s lemma for internal higher categories

work page 2024
[341]

Springer, Berlin-New York, 1973, pp

Lecture Notes in Math. Springer, Berlin-New York, 1973, pp. 85–147. [SAG] J. Lurie. Spectral algebraic geometry

work page 1973

Show all 13 references

[2017]

[Hoy17] M

url: https://www.math.ias.edu/~lurie/papers/ HA.pdf. [Hoy17] M. Hoyois. The six operations in equivariant motivic homotopy theory. In:Advances in Mathematics305 (2017), pp. 197–279. [Hoy20] M. Hoyois. Cdh descent in equivariant homotopyK-theory. In:Doc. Math.25 (2020), pp. 457...

2017
[2018]

url: https : / / www . math . ias . edu / ~lurie/papers/SAG-rootfile.pdf. [Swa85] R. G. Swan. K-theory of quadric hypersurfaces. In:Ann. of Math. (2)122.1 (1985), pp. 113–153. [TT90] R. W. Thomason and T. Trobaugh. Higher algebraic K-theory of schemes and of derived categories...

1985
[2022]

[CLL25] B

arXiv:2010.10325 [math.AT]. [CLL25] B. Cnossen, T. Lenz, and S. Linskens. Partial parametrized presentability and the universal property of equivariant spectra

2010 arXiv
[2024]

[Bar+24] T

arXiv:2408.09390 [math.AG]. [Bar+24] T. Barthel, S. Carmeli, T. M. Schlank, and L. Yanovski. The chromatic Fourier transform. In:Forum Math. Pi12 (2024), Paper No. e8,

2024 arXiv
[2025]

url: https://arxiv.org/abs/2307.11001

arXiv:2307.11001 [math.AT]. url: https://arxiv.org/abs/2307.11001. [EKS25] E. Elmanto, D. Kubrak, and V. Sosnilo. On filtered algebraicK-theory of stacks I: characteristic zero

arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [5]

arXiv:2503.09928 [math.AG]. [HA] J. Lurie. Higher algebra

work page Pith review arXiv

[2] [7]

Standard $t$-structures

arXiv:2504. 07473 [math.CT]. url: https://arxiv.org/abs/2504.07473. [HTT] J. Lurie. Higher topos theory. Vol

work page Pith review arXiv

[3] [9]

[Mou21] T

arXiv:2103.17141 [math.CT]. [Mou21] T. Moulinos. The geometry of filtrations. In: Bull. Lond. Math. Soc. 53.5 (2021), pp. 1486–1499. 44 [MP97] A. S. Merkurjev and I. A. Panin. K-theory of algebraic tori and toric varieties. In: K-Theory 12.2 (1997), pp. 101–143. [MV99] F. Morel and V. Voevodsky. A1-homotopy theory of schemes. In:Inst. Hautes Études Sci. P...

arXiv 2021

[4] [10]

[Qui72] D

arXiv: 2209.05103 [math.CT]. [Qui72] D. Quillen. On the cohomology andK-theory of the general linear groups over a finite field. In:Ann. of Math. (2)96 (1972), pp. 552–586. [Qui73] D. Quillen. Higher algebraicK-theory. I. In:AlgebraicK-theory, I: HigherK-theories (Proc. Conf., Battelle Memorial Inst., Seattle, Wash., 1972). Vol. Vol

arXiv 1972

[5] [88]

Progr. Math. Birkhäuser Boston, Boston, MA, 1990, pp. 247–435. 45

work page 1990

[6] [96]

[Bar17] C. Barwick. Spectral Mackey functors and equivariant algebraicK-theory (I). In:Adv. Math. 304 (2017), pp. 646–727. [BGT13] A. J. Blumberg, D. Gepner, and G. Tabuada.A universal characterization of higher algebraicK-theory. In:Geom. Topol.17.2 (2013), pp. 733–838. [BH21] T. Bachmann and M. Hoyois. Norms in motivic homotopy theory. In:Astérisque 425...

work page 2017

[7] [170]

Princeton University Press, Princeton, NJ, 2009, pp

Annals of Mathematics Studies. Princeton University Press, Princeton, NJ, 2009, pp. xviii+925. [KR24] A. A. Khan and C. Ravi. Generalized cohomology theories for algebraic stacks. In: Adv. Math.458 (2024), Paper No. 109975, 104.issn: 0001-8708,1090-2082. [Mar22] L. Martini. Yoneda’s lemma for internal higher categories

work page 2024

[8] [341]

Springer, Berlin-New York, 1973, pp

Lecture Notes in Math. Springer, Berlin-New York, 1973, pp. 85–147. [SAG] J. Lurie. Spectral algebraic geometry

work page 1973