betaFeedback

WorksMathematics

On a generalization of result due to Peskine and Szpiro-II

Abstract

Let (R,m)(R,\mathfrak{m}) be a regular local ring and let II be a Cohen-Macaulay ideal of RR. When Char⁡R=p>0\operatorname{Char} R=p>0, a result of Peskine and Szpiro proves that HIi(R)=0H^i_I(R)=0 for i≠height⁡Ii\neq\operatorname{height} I. This result is not true when Char⁡R=0\operatorname{Char} R=0. In this article, we prove that the Bass numbers of HIheight⁡I(R)H^{\operatorname{height} I}_I(R) completely determines the vanishing of HIi(R)H^i_I(R) for i≠height⁡Ii\neq \operatorname{height} I. This result was previously established by the second author in the equicharacteristic case and more generally for Gorenstein local rings by Hellus and Schenzel. In contrast to the spectral sequence methods employed by Hellus and Schenzel, our proof is elementary and self-contained.

Introduction

Let (R,m)(R,\mathfrak{m}) be a regular local ring. We say that an ideal II of RR is Cohen–Macaulay if R/IR/I is Cohen–Macaulay local ring. We recall the definition of Peskine–Szpiro ideal.

Definition 1.1. An ideal II of RR is said to be a Peskine–Szpiro ideal of RR if the following hold.

  1. II is a Cohen–Macaulay ideal of RR.

  2. HIi(R)=0H^i_I(R)=0 for all i≠height⁡Ii\ne\operatorname{height} I.

Note that as height⁡I=grade⁡I\operatorname{height} I=\operatorname{grade} I, we have HIi(R)=0H^i_I(R)=0 for i<height⁡Ii<\operatorname{height} I. Thus a Cohen–Macaulay ideal II is Peskine–Szpiro ideal if HIi(R)=0H^i_I(R)=0 for i>height⁡Ii>\operatorname{height} I. The next result is due to Peskine and Szpiro [9 Proposition III.4.1].

Theorem 1.2. Let RR be a regular domain of characteristic p>0p>0. If II is a Cohen–Macaulay ideal of RR, then HIi(R)=0H^i_I(R)=0 for all i≠height⁡Ii\ne\operatorname{height} I.

The above result does not hold over regular rings of characteristic zero, for instance, see [6 Example 21.31]. This motivated the second author to prove the following surprising result.

Theorem 1.3. [10 Theorem 1.2] Let (R,m)(R,\mathfrak{m}) be a regular local ring of dimension dd containing a field KK. Let II be a Cohen–Macaulay ideal of height gg. The following conditions are equivalent.

(a) II is a Peskine–Szpiro ideal of RR.

(b) If P∈V(I)P\in V(I),

μi(P,HIg(R))={1if i=height⁡P−g,0,otherwise.\mu_i(P,H^g_I(R))= \begin{cases} 1 & \text{if } i=\operatorname{height} P-g,\\ 0, & \text{otherwise}. \end{cases}

The proof of Theorem 1.3 uses the fact that injdim⁡RHIi(R)≤dim⁡Supp⁡HIi(R)\operatorname{injdim}_R H^i_I(R)\leq\dim\operatorname{Supp}H^i_I(R) which was prove by Huneke and Sharp [4] when char⁡K=p>0\operatorname{char} K=p>0 and by Lyubeznik [7] when char⁡K=0\operatorname{char} K=0. Unfortunately, when the regular ring does not contain a field, the bound on injective dimension of local cohomology modules are different. In [13], Zhou proved that injdim⁡RHIi(R)≤dim⁡R(Supp⁡RHIi(R))+1\operatorname{injdim}_{R} H_{I}^{i}(R) \le\dim_{R}(\operatorname{Supp}_{R} H_{I}^{i}(R)) + 1 when RR is an unramified regular local of mixed characteristic and II is an ideal of RR. The aim of this article is to establish Theorem 1.3 without the assumption that RR contains field. More precisely, we prove the following.

Theorem 1.4. (Theorem 3.4 below) Let (R,m)(R,\mathfrak{m}) be a regular local ring of dimension dd. Let II be a Cohen–Macaulay ideal of height gg. The following conditions are equivalent.

(a) II is a Peskine–Szpiro ideal of RR.

(b) If P∈V(I)P \in V(I),

μIi(P,HIg(R))={1if i=height⁡P−g,0,otherwise.\mu_{I}^{i}(P,H_{I}^{g}(R)) = \begin{cases} 1 & \text{if } i = \operatorname{height} P - g,\\ 0, & \text{otherwise}. \end{cases}

We remark that the above result follows as a special case of a more general result due to Hellus and Schenzel [3 Theorem 3.2]. However, our proof is elementary and avoids a spectral sequence argument.

Cohen-Macaulay ideals of small dimensions

Let RR be a Noetherian ring and let II be an ideal of RR. The local cohomological dimension of II, denoted as cd⁡(I)\operatorname{cd}(I) is defined as follows:

cd⁡(I)=sup⁡{j∣HIj(R)≠0}.\operatorname{cd}(I) = \sup\{j \mid H_{I}^{j}(R) \ne0\}.

Let us first recall the first vanishing theorem of local cohomology which is commonly known as Hartshorne-Lichtenbaum Vanishing Theorem (HLVT for short) [2].

Theorem 2.1. (Hartshorne-Lichtenbaum vanishing theorem) Let (R,m)(R,\mathfrak{m}) be a complete local domain. If II is an ideal of RR which is not m\mathfrak{m}-primary, then cd⁡(I)≤dim⁡R−1\operatorname{cd}(I) \le\dim R - 1.

Let us now recall the second vanishing theorem of local cohomology modules.

Theorem 2.2. Suppose (R,m)(R,\mathfrak{m}) is a complete regular local ring of dimension dd with separably closed residue field. Then,

cd⁡(R/I)≤d−2⟺dim⁡(R/I)≥2 and Spec⁡(R/I)∖{m} is connected.\operatorname{cd}(R/I) \le d - 2 \Longleftrightarrow\dim(R/I) \ge2 \text{ and } \operatorname{Spec}(R/I) \setminus\{\mathfrak{m}\}\text{ is connected}.

The equal characteristic zero and positive characteristic cases were established by Ogus [8] and Peskine–Szpiro [9], respectively. Later, Huneke and Lyubeznik [5] gave a characteristic-free proof of the second vanishing result. The unramified mixed characteristic case was proved by Zhang [12] and the remaining ramified mixed characteristic case was recently settled by Scheffelin [11]. We use the second vanishing theorem to prove the following proposition.

Proposition 2.3. Let (R,m)(R,\mathfrak{m}) be a regular local ring of dimension dd and let II be a Cohen–Macaulay ideal of RR of height gg with d−g≤2d-g \le2. Then II is a Peskine–Szpiro ideal.

Proof. When char⁡R=p>0\operatorname{char} R = p > 0, the result follows from [9 Proposition III.4.1]. Therefore we assume that char⁡R=0\operatorname{char} R = 0.

If g=dg=d, then II is m\mathfrak{m}-primary and therefore by Grothendieck vanishing theorem HIi(R)=0H_{I}^{i}(R)=0 for i>height⁡Ii > \operatorname{height} I. So, II is a Peskine–Szpiro ideal. Now consider g=d−1g=d-1. Therefore II is not m\mathfrak{m}-primary and hence IR^I\widehat{R} is not mR^\mathfrak{m}\widehat{R}-primary. By Hartshorne–Lichtenbaum theorem, HIR^d(R^)=0H^{d}_{I\widehat{R}}(\widehat{R})=0. By faithful flatness we get HId(R)=0H^{d}_{I}(R)=0.

Now let g=d−2g=d-2. Choose a faithfully flat extension (S,n)(S,\mathfrak{n}) of RR such that mS=n\mathfrak{m}S=\mathfrak{n}, SS is complete and S/nS/\mathfrak{n} is algebraically closed. We note that (S,n)(S,\mathfrak{n}) is regular local, S/ISS/IS is Cohen–Macaulay and dim⁡S/IS=2\dim S/IS=2. Therefore Spec⁡(S/IS)∖{n}\operatorname{Spec}(S/IS)\setminus\{\mathfrak{n}\} is connected by [6 Theorem 15.7]. By the second vanishing theorem we get HISd−1(S)=0H^{d-1}_{IS}(S)=0 and this gives HId−1(R)=0H^{d-1}_{I}(R)=0. By a similar argument to the previous case we also get HId(R)=0H^{d}_{I}(R)=0.

The proof of the following result is identical to that of [10 Proposition 2.1] and is omitted to avoid repetition.

Proposition 2.4. Let (R,m)(R,\mathfrak{m}) be a regular local ring. Let II be a Peskine–Szpiro ideal of RR of height gg. Then, the following statements are true.

(a) Let PP be an ideal in RR containing II. Then IPI_P is a Peskine–Szpiro ideal of RPR_P.

(b) Assume g<dim⁡Rg<\dim R. Let x∈m−m2x\in\mathfrak{m}-\mathfrak{m}^2 be R/IR/I-regular. Then the ideal (I+(x))/(x)(I+(x))/(x) is a Peskine–Szpiro ideal of R/(x)R/(x).

Proof of the main theorem

The following fact is well known, but we include a proof for completeness, as we are unable to find a suitable reference.

Lemma 3.1. Let (R,m)(R,\mathfrak{m}) be a Noetherian local ring and let MM be an RR-module. Assume that Supp⁡M={m}\operatorname{Supp} M=\{\mathfrak{m}\} and x∈mx\in\mathfrak{m}. Then ker⁡(M→xM)≠0\operatorname{ker}(M\xrightarrow{x}M)\ne0.

Proof. Let z∈Mz\in M be non-zero. Then Supp⁡(Rz)=V(ann⁡(Rz))={m}\operatorname{Supp}(Rz)=V(\operatorname{ann}(Rz))=\{\mathfrak{m}\}. This implies ann⁡(Rz)=m\sqrt{\operatorname{ann}(Rz)}=\mathfrak{m} and hence mtz=0\mathfrak{m}^{t}z=0 for some t≥1t\ge1. In particular, xtz=0x^{t}z=0. Choose a smallest tt with this property. Then y=xt−1z≠0y=x^{t-1}z\ne0 and y∈ker⁡(M→xM)y\in\operatorname{ker}(M\xrightarrow{x}M).

We also note the following remark.

Remark 3.2. Let RR be a regular ring and let JJ be an ideal of height gg of RR. Then Ass⁡HJg(R)={P∣P⊇J and height⁡P=g}\operatorname{Ass} H^g_J(R)=\{P\mid P\supseteq J\text{ and }\operatorname{height}P=g\}. Also for any such prime PP we have μ0(P,HJg(R))=1\mu_0(P,H^g_J(R))=1 since

(HJg(R))P=HJRPg(RP)=HPRPg(RP)=ERP(RP/PRP).(H^g_J(R))_P=H^g_{JR_P}(R_P)=H^g_{PR_P}(R_P)=E_{R_P}(R_P/PR_P).

Let us also recall the following result due to Rees [1 Lemma 3.1.16].

Lemma 3.3. Let SS be a commutative ring and let MM and NN be SS-modules (not necessarily finitely generated). Let x∈Sx\in S be S⊕MS\oplus M regular with xN=0xN=0. Then

Ext⁡Si(M,N)≅Ext⁡S/xSi(M,N/xN).\operatorname{Ext}^{i}_{S}(M,N)\cong\operatorname{Ext}^{i}_{S/xS}(M,N/xN).

We are now in a position to prove the main theorem. While the direction (a)⟹(b)(a)\Longrightarrow(b) is analogous to [10 Theorem 1.2], we provide the complete proof here to ensure the paper is self-contained. Note that the proof of the converse direction (b)⟹(a)(b)\Longrightarrow(a) does not follow from [10 Theorem 1.2] and requires modification.

Theorem 3.4. Let (R,m)(R,\mathfrak{m}) be a regular local ring of dimension dd. Let II be a Cohen–Macaulay ideal of height gg. The following conditions are equivalent.

(a) II is a Peskine–Szpiro ideal of RR.

(b) If P∈V(I)P \in V(I),

μi(P,HIg(R))={1if i=height⁡P−g,0,otherwise.\mu_i(P,H_I^g(R)) = \begin{cases} 1 & \text{if } i=\operatorname{height} P-g,\\ 0, & \text{otherwise}. \end{cases}

Proof. We first prove (a) ⟹\Longrightarrow (b). Let II be a Peskine-Szpiro ideal. We prove the result by induction on d−gd-g. If d−g=0d-g=0, then I=m\sqrt{I}=\mathfrak{m} and hence HId(R)=Hmd(R)=ER(R/m)H_I^d(R)=H_{\mathfrak{m}}^d(R)=E_R(R/\mathfrak{m}).

Now assume d−g=1d-g=1. Let P∈V(I)P \in V(I) with height P=d−1P=d-1. By Remark 3.2, μ0(P,HId−1(R))=1\mu_0(P,H_I^{d-1}(R))=1 and μi(P,HId−1(R))=0\mu_i(P,H_I^{d-1}(R))=0 for i≥1i \ge1. Now let P=mP=\mathfrak{m}. Then by Remark 3.2, μ0(m,HId−1(R))=0\mu_0(\mathfrak{m},H_I^{d-1}(R))=0. Choose x∈m∖m2x \in\mathfrak{m}\setminus\mathfrak{m}^2 which is R/IR/I-regular. Let R‾=R/(x)\overline{R}=R/(x), n=m/(x)\mathfrak{n}=\mathfrak{m}/(x) and J=IR‾=(I+(x))/(x)J=I\overline{R}=(I+(x))/(x). Note that JJ is n\mathfrak{n}-primary. The short exact sequence 0→R→xR→R‾→00 \to R \xrightarrow{x} R \to\overline{R} \to0 induces the following long exact sequence on the local cohomology

0→HId−1(R)→xHId−1(R)→HJd−1(R‾)→0.0 \to H_I^{d-1}(R) \xrightarrow{x} H_I^{d-1}(R) \to H_J^{d-1}(\overline{R}) \to0.

Therefore,

μiR(m,HId−1(R))=μi−1R‾(n,HJd−1(R‾))={1if i=1,0,otherwise.\mu_i^R(\mathfrak{m},H_I^{d-1}(R))=\mu_{i-1}^{\overline{R}}(\mathfrak{n},H_J^{d-1}(\overline{R}))= \begin{cases} 1 & \text{if } i=1,\\ 0, & \text{otherwise}. \end{cases}

Thus the result follows in this case.

Now consider the case d−g≥2d-g \ge2. Now let P∈V(I)P \in V(I) be such height P=g+rP=g+r with P≠mP \ne\mathfrak{m}. Then by Proposition 2.4, IPI_P is a Peskine-Szpiro ideal of RPR_P. Note that dim⁡RP−g<d−g\dim R_P-g<d-g. Therefore by the induction hypothesis

μi(P,HIg(R))=μi(PRP,HIPg(RP))={1if i=r,0otherwise.\mu_i(P,H_I^g(R))=\mu_i\left(P R_P,H_{I_P}^g(R_P)\right)= \begin{cases} 1 & \text{if } i=r,\\ 0 & \text{otherwise}. \end{cases}

We now consider the case P=mP=\mathfrak{m}. By Remark 3.2, μ0(m,HIg(R))=0\mu_0(\mathfrak{m},H_I^g(R))=0. Choose x∈m∖m2x \in\mathfrak{m}\setminus\mathfrak{m}^2 which is R/IR/I-regular. Let R‾=R/(x)\overline{R}=R/(x), n=m/(x)\mathfrak{n}=\mathfrak{m}/(x) and J=IR‾=(I+(x))/(x)J=I\overline{R}=(I+(x))/(x). By Proposition 2.4, JJ is a height gg Peskine-Szpiro ideal in R‾\overline{R}. The short exact sequence 0→R→xR→R‾→00 \to R \xrightarrow{x} R \to\overline{R} \to0 induces the following long exact sequence on the local cohomology

0→HIg(R)→xHIg(R)→HJg(R‾)→0.0 \to H_I^g(R) \xrightarrow{x} H_I^g(R) \to H_J^g(\overline{R}) \to0.

The result now follows from the induction hypothesis (dim⁡R‾−g=d−1−g\dim\overline{R}-g=d-1-g) and using the fact μi(m,HIg(R))=μi−1(n,HJd−1(R‾))\mu_i(\mathfrak{m},H_I^g(R))=\mu_{i-1}(\mathfrak{n},H_J^{d-1}(\overline{R})).

We now prove (b) ⟹\Longrightarrow (a). By Theorem 1.2, we may assume char⁡R=0\operatorname{char} R=0. We apply induction on d−gd-g. If d−g≤2d-g \le2, the result follows from Proposition 2.3. Therefore, we assume that d−g≥3d-g \ge3. Let P∈V(I)P \in V(I) with P≠mP \ne\mathfrak{m}. Then, IPI_P is a Cohen-Macaulay ideal of RPR_P of height gg and satisfies (b). As dim⁡RP−g<d−g\dim R_P-g<d-g, induction hypothesis implies that IPI_P is a Peskine-Szpiro ideal of RPR_P. Therefore, HIPi(RP)=0H_{I_P}^i(R_P)=0 for all i>gi>g. It follows that Supp⁡RHIi(R)⊆{m}\operatorname{Supp}_R H_I^i(R) \subseteq\{\mathfrak{m}\} for all $i>g.

Choose x∈m∖m2x \in\mathfrak{m}\setminus\mathfrak{m}^2 which is R/IR/I-regular. Let R‾=R/(x)\overline{R}=R/(x), n=m/(x)\mathfrak{n}=\mathfrak{m}/(x) and J=IR‾=(I+(x))/(x)J=I\overline{R}=(I+(x))/(x). The short exact sequence 0→R→xR→R‾→00 \to R \xrightarrow{x} R \to\overline{R} \to0 induces the following long exact sequence on the local cohomology

0→HIg(R)→xHIg(R)→HJg(R‾)→HIg+1(R)→xHIg+1(R)→⋯(1)0 \to H_I^g(R) \xrightarrow{x} H_I^g(R) \to H_J^g(\overline{R}) \to H_I^{g+1}(R) \xrightarrow{x} H_I^{g+1}(R) \to\cdots \tag*{(1)}

Let Vr:=coker⁡(HIr(R)→xHIr(R))V^r := \operatorname{coker}(H_I^r(R) \xrightarrow{x} H_I^r(R)) and Lr:=ker⁡(HIr(R)→xHIr(R))L^r := \ker(H_I^r(R) \xrightarrow{x} H_I^r(R)) for all rr. Therefore we have the following two short exact sequences

0→HIg(R)→xHIg(R)→Vg→0(2)0 \to H_I^g(R) \xrightarrow{x} H_I^g(R) \to V^g \to0 \tag*{(2)}

and

0→Vg→HJg(R‾)→Lg+1→0.(3)0 \to V^g \to H_J^g(\overline{R}) \to L^{g+1} \to0. \tag*{(3)}

Note that Supp⁡RLg+1⊆Supp⁡RHIg+1(R)⊆{m}\operatorname{Supp}_{R} L^{g+1} \subseteq\operatorname{Supp}_{R} H_I^{g+1}(R) \subseteq\{\mathfrak{m}\}. We first claim that HIg+1(R)=0H_I^{g+1}(R) = 0.

If possible let HIg+1(R)≠0H_I^{g+1}(R) \ne0. Then By Lemma 3.1, Supp⁡RHIg+1(R)=Supp⁡RLg+1={m}\operatorname{Supp}_{R} H_I^{g+1}(R) = \operatorname{Supp}_{R} L^{g+1} = \{\mathfrak{m}\}. Since R/IR/I is Cohen–Macaulay and R‾/J=R‾/I+(x)=(R/I)/x(R/I)\overline{R}/J = \overline{R}/I + (x) = (R/I)/x(R/I), we see that JJ is Cohen–Macaulay ideal of R‾\overline{R}. Also height⁡J=dim⁡R‾−dim⁡R‾/J=height⁡(I)=g\operatorname{height} J = \dim\overline{R} - \dim\overline{R}/J = \operatorname{height}(I) = g. By Remark 3.2, n∉Ass⁡HJg(R‾)\mathfrak{n} \notin\operatorname{Ass} H_J^g(\overline{R}) as d−1≠gd - 1 \ne g and hence μnR‾(n,HJg(R‾))=0\mu_{\mathfrak{n}}^{\overline{R}}(\mathfrak{n}, H_J^g(\overline{R})) = 0. Therefore, the short exact sequence (3) gives

μ1R‾(n,Vg)≥μ0R‾(n,Lg+1)>0.\mu_1^{\overline{R}}(\mathfrak{n}, V^g) \ge\mu_0^{\overline{R}}(\mathfrak{n}, L^{g+1}) > 0.

But μ1R‾(n,Vg)=μ2R(m,HIg(R))\mu_1^{\overline{R}}(\mathfrak{n}, V^g) = \mu_2^R(\mathfrak{m}, H_I^g(R)) by Lemma 3.3 and therefore by condition (b), d−g=2d-g=2 which is a contradiction. Consequently, HIg+1(R)=0H_I^{g+1}(R) = 0.

Since HIg+1(R)=0H_I^{g+1}(R) = 0, from (1) we get the following short exact sequence

0→HIg(R)→xHIg(R)→HJg(R‾)→0.0 \to H_I^g(R) \xrightarrow{x} H_I^g(R) \to H_J^g(\overline{R}) \to0.

Note that JJ is a Cohen–Macaulay ideal of R‾\overline{R} and satisfies condition (b) of our theorem. Hence by the induction hypothesis, HJi(R‾)=0H_J^i(\overline{R}) = 0 for i>gi > g. Therefore, from (1), we get Li=0L^i = 0 for all i≥g+2i \ge g + 2. This implies HIi(R)=0H_I^i(R) = 0 for i≥g+2i \ge g + 2 by Lemma 3.1 as Supp⁡RHIi(R)={m}\operatorname{Supp}_{R} H_I^i(R) = \{\mathfrak{m}\} for i≥g+2i \ge g + 2. □

References

Indian Institute of Technology Bombay, Mumbai, India

Email address: [email protected]

Indian Institute of Technology Bombay, Mumbai, India

Email address: [email protected]

References

  1. [1]W. Bruns, J. Herzog, Cohen–Macaulay rings, Camb. Stud. Adv. Math. 39, Cambridge University Press, Cambridge, 1993
  2. [2]R. Hartshorne, Cohomological dimension of algebraic varieties, Ann. of Math. (2) 88 (1968), 403–450. MR0232780DOI
  3. [3]M. Hellus and P. Schenzel, On cohomologically complete intersections, J. Algebra 320 (2008), no. 10, 3733–3748, MR 2457720.arxiv.org/abs/0804.2558
  4. [4]C. Huneke and R. Sharp, Bass numbers of local cohomology modules, Trans. Amer. Math. Soc. 339 (1993), 765–779. MR 1124167DOI
  5. [5]C. Huneke and G. Lyubeznik, On the vanishing of local cohomology modules, Invent. Math. 102 (1990), no. 1, 73–93. MR1069240DOI
  6. [6]S. B. Iyengar, G. J. Leuschke, A. Leykin, C. Miller, E. Miller, A. K. Singh, and U. Walther, Twenty-Four Hours of Local Cohomology, Grad. Stud. Math., vol. 87, American Mathematical Society, Providence, RI, 2007, MR 2355715.DOI
  7. [7]G. Lyubeznik, Finiteness properties of local cohomology modules (an application of D-modules to commutative algebra), Invent. Math. 113 (1993), no. 1, 41–55. MR 1223223DOI
  8. [8]A. Ogus, Local cohomological dimension of algebraic varieties, Ann. of Math. (2) 98 (1973), 327–365. MR0506248DOI
  9. [9]C. Peskine and L. Szpiro, Dimension projective finie et cohomologie locale, Inst. Hautes Études Sci. Publ. Math. 42 (1973), 47–119, MR 374130.DOI
  10. [10]T. J. Puthenpurakal, On a generalization of a result of Peskine and Szpiro, Rend. Semin. Mat. Univ. Padova 151 (2024), 77–83, MR 4712808.arxiv.org/abs/1506.03614
  11. [11]A. Scheffelin, The Second Vanishing Theorem in Ramified Mixed Characteristic, arXiv e-prints, June 2026.arxiv.org/abs/2606.26867
  12. [12][Zha25] W. Zhang, The second vanishing theorem for local cohomology modules, Math. Res. Lett. 32 (2025), no. 6, 2039–2062, MR 5026248. 2arxiv.org/abs/2102.12545
  13. [13][Zho98] C. Zhou, Higher derivations and local cohomology modules, J. Algebra 201 (1998), no. 2, 363–372, MR 1612378. 2DOI

Paper details

vlnc:W_73qv9xsu
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents