Search

We prove the universal triviality of the third unramified cohomology group of a very general complex cubic fourfold containing a plane. The proof uses results on the unramified cohomology of quadrics due to Kahn, Rost, and Sujatha., Comment: 20 pages, comments welcome

F. Campana had asked whether a certain threefold is rational. In arXiv:1310.3569v1 [mathAG], this variety was shown to be birational to a specific conic bundle and then to be unirational. We prove that this conic bundle is rational., Comment: in French

We construct a conic bundle over an elliptic curve over a real quadratic field that is a counterexample to the Hasse principle not explained by the \'etale Brauer-Manin obstruction. We also give simple examples of threefolds with the same property that are families of 2-dimensional quadrics, and discuss some other examples and general properties of the Brauer-Manin obstruction., Comment: 22 pages, to appear in Mathematische Zeitschrift

Connections between the second Chow group of a smooth projective variety and its third unramified cohomology group, with coefficients the roots of unity twisted twice, feature in several recent works. In this note we revisit a 1996 paper by B. Kahn and specialize it to various types of rationally connected varieties., Comment: In French. Final version, to appear in Documenta math. Section 5 contains a discussion of the universal unramified third cohomology group for Fano hypersurfaces, and in particular for cubic hypersurfaces

Let C be the complex field and K=C((x,y)) or K=C((x))(y). Let G be a connected linear algebraic group over K. Under the assumption that the K-variety G is K-rational, i.e. that the function field is purely transcendant, it was proved that a principal homogeneous space of G has a rational point over K as soon as it has one over each completion of K with respect to a discrete valuation. In this paper we show that one cannot in general do without the K-rationality assumption. To produce our examples, we introduce a new type of obstruction. It is based on higher reciprocity laws on a 2-dimensional scheme. We also produce a family of principal homogeneous spaces for which the refined obstruction controls exactly the existence of rational points., Comment: Final version, in French, to appear in the Transactions of the American Mathematical Society

Let k be a field, X a smooth, projective k-variety. If X is geometrically rational, there is an injective map from the quotient of Brauer groups Br(X)/Br(k) into the first Galois cohomology group of the lattice given by the geometric Picard group. In this note, where the main attention is on smooth compactifications of homogeneous spaces of algebraic k-tori, we show how under some hypotheses the map is onto, and how one may in some special case exhibit concrete generators in Br(X). This is applied to the analysis of counterexamples to the local-global principle for norms in biquadratic extensions of number fields.

Let $k$ be a number field and $X$ a smooth integral affine variety equipped with a morphism $f : X \to A^1_k$ to the affine line. Assume that all fibres of $f$ are split, for instance that they are geometrically integral. Assume that the generic fibre of $f$ is a homogeneous space of a simply connected, almost simple, semisimple group $G/k(t)$, and that the geometric stabilizers are connected reductive groups. Let $v$ be a place of $k$ such that the fibration $f$ acquires a rational section over the completion $k_v$ at $v$. Assume moreover that at almost all points $x \in A^1(k_v)$ the specialized group $G_x$ is isotropic over $k_v$. If the Brauer group of $X$ is reduced to the Brauer group of $k$, then strong approximation holds for $X$ away from the place $v$., Comment: (July 16th, 2013) The summary is correct, but as Dasheng Wei pointed out to us, there was a mistake in a proof in the previous versions of the paper. Section 3 has been rewritten. Alongside with a technique from earlier papers by the second named author, a new technical tool is Proposition 3.4, which combines strong approximation with a specific proof of Hilbert's irreducibility theorem

Let k be a field, G a finite group embedded in the k-group SL(n). For k an algebraically closed field, Bogomolov gave a formula for the unramified Brauer group of the quotient SL(n)/G. We develop his method over any characteristic zero field. This purely algebraic method enables us to recover and generalize results of Harari and of Demarche over number fields, such as the triviality of the unramified Brauer group for k=Q and G of odd order. --- Soient k un corps et G un groupe fini plong\'e dans le k-groupe SL(n).Pour k alg\'ebriquement clos, Bogomolov a donn\'e une formule pour le groupe de Brauer non ramifi\'e du quotient SL(n)/G. On examine ce que donne sa m\'ethode sur un corps k quelconque (de caract\'eristique nulle). Par cette m\'ethode purement alg\'ebrique, on retrouve et \'etend des r\'esultats obtenus par Harari et par Demarche au moyen de m\'ethodes arithm\'etiques, comme la trivialit\'e du groupe de Brauer non ramifi\'e pour k= Q et G d'ordre impair., Comment: 10 pages, in French

We study equations in four variables (x,y,z,t) of the shape q(x,y,z)=P(t), where q(x,y,z) is an indefinite ternary quadratic form over the integers and P(t) is a polynomial in one variable with integral coefficients. If P(t) is not the product of a constant and the square of a polynomial, strong approximation holds for integral solutions (x,y,z,t). In the general case, we show that the integral Brauer-Manin conditions are the only obstructions to strong approximation. We actually study the analogous situation over an arbitrary number field. --- Nous \'etudions les \'equations \`a quatre variables (x,y,z,t) \`a coefficients entiers du type q(x,y,z)=P(t), o\`u q(x,y,z) est une forme quadratique enti\`ere ternaire ind\'efinie sur les r\'eels, et P(t) un polyn\^ome \`a coefficients entiers en une variable. Lorsque le polyn\^ome n'est pas le produit d'une constante et d'un carr\'e de polyn\^ome, nous \'etablissons l'approximation forte pour les solutions de ces \'equations en entiers (x,y,z,t). Dans le cas g\'en\'eral, nous montrons que l'obstruction de Brauer-Manin enti\`ere est la seule obstruction \`a l'approximation forte. Nous \'etudions la situation sur un corps de nombres quelconque., Comment: 26 pages, in English

The third unramified cohomology group is shown to vanish on certain varieties equipped with a pencil of quadrics or of smooth complete intersections of two quadrics. Over the complex field, this shows that the integral Hodge conjecture in degree 4 holds for such varieties. --- On \'etablit la nullit\'e du troisi\`eme groupe de cohomologie non ramifi\'ee pour certaines vari\'et\'es munies d'un pinceau de quadriques ou d'intersections compl\`etes lisses de deux quadriques. Sur les complexes, ceci permet d'\'etablir la conjecture de Hodge enti\`ere en degr\'e 4 pour de telles vari\'et\'es.

For a smooth and projective variety X over a field k of characteristic zero we prove the finiteness of the cokernel of the natural map from the Brauer group of X to the Galois-invariant subgroup of the Brauer group of the same variety over an algebraic closure of k. Under further conditions on k, e.g. over number fields, we give estimates for the order of this cokernel. ---- Soit X une vari\'et\'e projective et lisse sur un corps k de caract\'eristique z\'ero. Le groupe de Brauer de X s'envoie dans les invariants, sous le groupe de Galois absolu de k, du groupe de Brauer de la m\^eme vari\'et\'e consid\'er\'ee sur une cl\^oture alg\'ebrique de k. Nous montrons que le quotient est fini. Sous des hypoth\`eses suppl\'ementaires, par exemple sur k un corps de nombres, nous donnons des estimations sur l'ordre de ce quotient., Comment: 33 pages, in French

Let $X$ be a smooth projective variety over a finite field $\F$. We discuss the unramified cohomology group $H^3_\nr(X,\Q/\Z(2))$. Several conjectures put together imply that this group is finite. For certain classes of threefolds, $H^3_\nr(X,\Q/\Z(2))$ actually vanishes. It is an open question whether this holds true for arbitrary threefolds. For a threefold $X$ equipped with a fibration onto a curve $C$, the generic fibre of which is a smooth projective surface $V$ over the global field $\F(C)$, the vanishing of $H^3_\nr(X,\Q/\Z(2))$ together with the Tate conjecture for divisors on $X$ implies a local-global principle of Brauer--Manin type for the Chow group of zero-cycles on $V$. This sheds a new light on work started thirty years ago. ----- Soit $X$ une vari\'et\'e projective et lisse sur un corps fini $\F$. On s'int\'eresse au groupe de cohomologie non ramifi\'ee $H^3_\nr(X,\Q/\Z(2))$. Un faisceau de conjectures implique que ce groupe est fini. Pour certaines classes de solides, on a $H^3_\nr(X,\Q/\Z(2))=0$. Savoir si c'est le cas pour tout solide est un probl\`eme ouvert. Lorsqu'un solide $X$ est fibr\'e au-dessus d'une courbe $C$, de fibre g\'en\'erique une surface projective et lisse $V$ sur le corps global $\F(C)$, la combinaison de $H^3_\nr(X,\Q/\Z(2))=0$ et de la conjecture de Tate pour $X$ a pour cons\'equence un principe local-global de type Brauer--Manin pour le groupe de Chow des z\'ero-cycles de la fibre g\'en\'erique $V$. Ceci \'eclaire d'un jour nouveau des investigations commenc\'ees il y a trente ans., Comment: Slightly revised version of arXiv:1104.3350v2 [math.AG]. To appear in the Journal of K-Theory

For a smooth, projective family of homogeneous varieties defined over a number field, we show that if potential density holds for the rational points of the base, then it also holds for the total space. A conjecture of Campana and Peternell, known in dimension at most 4 and for certain higher dimensional cases, would then imply potential density for the rational points of smooth projective varieties over number fields whose tangent bundle is nef., Comment: 10 p

For a smooth and projective variety over a number field with torsion free geometric Picard group and finite transcendental Brauer group we show that only the archimedean places, the primes of bad reduction and the primes dividing the order of the transcendental Brauer group can turn up in the description of the Brauer-Manin set., Comment: 17 pages, in English

Building upon the Bloch-Kato conjecture in Milnor K-theory, we relate the third unramified cohomology group with Q/Z coefficients with a group which measures the failure of the integral Hodge conjecture in degree 4. As a first consequence, a geometric theorem of the second-named author implies that the third unramified cohomology group with Q/Z coefficients vanishes on all uniruled threefolds. As a second consequence, a 1989 example by Ojanguren and the first named author implies that the integral Hodge conjecture in degree 4 fails for unirational varieties of dimension at least 6. For certain classes of threefolds fibered over a curve, we establish a relation between the integral Hodge conjecture and the computation of the index of the generic fibre. En nous appuyant sur la conjecture de Bloch-Kato en K-th\'eorie de Milnor, nous \'etablissons un lien g\'en\'eral entre le d\'efaut de la conjecture de Hodge enti\`ere pour la cohomologie de degr\'e 4 et le troisi\`eme groupe de cohomologie non ramifi\'e \`a coefficients Q/Z. Ceci permet de montrer que sur un solide unir\'egl\'e le troisi\`eme groupe de cohomologie non ramifi\'e \`a coefficients Q/Z s'annule, ce que la K-th\'eorie alg\'ebrique ne permet d'obtenir que dans certains cas. Ceci permet \`a l'inverse de d\'eduire d'exemples ayant leur source en K-th\'eorie que la conjecture de Hodge enti\`ere pour la cohomologie de degr\'e 4 peut \^etre en d\'efaut pour les vari\'et\'es rationnellement connexes. Pour certaines familles \`a un param\`etre de surfaces, on \'etablit un lien entre la conjecture de Hodge enti\`ere et l'indice de la fibre g\'en\'erique., Comment: Final version, 48 pages, to appear in Duke Mathematical Journal

Let F be a finite field of characteristic p. We consider smooth surfaces over F(t) defined by an equation f+tg=0, where f and g are forms of degree d in 4 variables with coefficients in F, with d prime to p. We prove : For such surfaces over F(t), the Brauer-Manin obstruction to the existence of a zero-cycle of degree one is the only obstruction. For d=3 (cubic surfaces), this leads to the same result for rational points. -- Soit F un corps fini de caract\'eristique p. Pour une surface lisse sur F(t) d\'efinie par une \'equation f+tg=0, o\`u f et g sont deux formes de degr\'e d sur F en 4 variables, avec d premier \`a p, nous montrons que l'obstruction de Brauer-Manin au principe de Hasse pour les z\'ero-cycles de degr\'e 1 est la seule obstruction. Pour d=3 (surfaces cubiques), on en d\'eduit le m\^{e}me \'enonc\'e pour les points rationnels., Comment: The title has changed and the text has been split into sections. In the statement of Theorems 1 and 3, "degree prime to char(F)" was corrected to "degree a power of char(F)."

Ceci est un rapport sur l'article "A finiteness theorem for zero-cycles over p-adic fields" (arXiv:math/0605165) de Shuji Saito et Kanetomo Sato. ----- This is a survey on the paper "A finiteness theorem for zero-cycles over p-adic fields" (arXiv:math/0605165) by Shuji Saito and Kanetomo Sato., Comment: 29 pages, Expos\'e au S\'eminaire Bourbaki (14 Novembre 2009)

M. Amer and A. Brumer have shown that, for two homogeneous quadratic polynomials f and g in at least 3 variables over a field k of characteristic different from 2, the locus f=g=0 has non-trivial solution over k if and only if, for a variable t, the equation f+tg=0 has a non-trivial solution over k(t). We consider a modified version of this result, and show that the projective variety over k defined by f_0=...=f_r=0, where the f_i are homogeneous polynomials over k of the same degree d\ge2 in n+1 variables (with n+1\ge r+2) , has a 0-cycle of degree 1 over k if and only if the generic hypersurface f_0+t_1f_1+...+t_rf_r=0 has a 0-cycle of degree 1 over k(t_1,...,t_r)., Comment: revised final version. 8 pages (originally 6 pages)

Let a be a nonzero integer. If a is not congruent to 4 or 5 modulo 9 then there is no Brauer-Manin obstruction to the existence of integers x, y, z such that x^3+y^3+z^3=a. In addition, there is no Brauer-Manin obstruction to the existence of integers x, y, z such that x^3+y^3+2z^3=a., Comment: 24 pages; minor changes only

This partly expository paper investigates versions of the Tate conjecture on the cycle map for varieties defined over finite fields with values in 'etale cohomology with Z_\ell-coefficients. The bulk of the paper is an exposition of a 1998 result of C. Schoen which shows that an integral version of the conjecture holds for 1-cycles provided the usual conjecture is true for divisors on surfaces. In a last section we then derive from Schoen's theorem new results on the existence of degree one zero-cycles on varieties defined over the function field of a smooth proper curve C over the algebraic closure of a finite field. In particular, we show that if the variety in question is a smooth projective complete intersection of dimension at least 3 and of degree prime to the characteristic, then a zero-cycle of degree one exists if the Tate conjecture is true for divisors on surfaces and the variety extends to a proper fibration over C all of whose fibres possess a component of multiplicity one., Comment: 16 pages, to appear in proceedings of Columbus conference on algebraic cycles (Clay Institute Publications) v2: Discussion of Atiyah-Hirzebruch counterexample added, following suggestions of Burt Totaro

Let $k$ be a field of characteristic zero, let $G$ be a connected reductive algebraic group over $k$ and let $\mathfrak{g}$ be its Lie algebra. Let $k(G)$, respectively, $k(\mathfrak{g})$, be the field of $k$-rational functions on $G$, respectively, $\mathfrak{g}$. The conjugation action of $G$ on itself induces the adjoint action of $G$ on $\mathfrak{g}$. We investigate the question whether or not the field extensions $k(G)/k(G)^G$ and $k(\mathfrak{g})/k(\mathfrak{g})^G$ are purely transcendental. We show that the answer is the same for $k(G)/k(G)^G$ and $k(\mathfrak{g})/k(\mathfrak{g})^G$, and reduce the problem to the case where $G$ is simple. For simple groups we show that the answer is positive if $G$ is split of type ${\sf A}_{n}$ or ${\sf C}_n$, and negative for groups of other types, except possibly ${\sf G}_{2}$. A key ingredient in the proof of the negative result is a recent formula for the unramified Brauer group of a homogeneous space with connected stabilizers. As a byproduct of our investigation we give an affirmative answer to a question of Grothendieck about the existence of a rational section of the categorical quotient morphism for the conjugating action of $G$ on itself. The results and methods of this paper have played an important part in recent A. Premet's negative solution (arxiv:0907.2500) of the Gelfand--Kirillov conjecture for finite-dimensional simple Lie algebras of every type, other than ${\sf A}_n$, ${\sf C}_n$, and ${\sf G}_2$., Comment: Final version, 37 pages. To appear in Compositio Mathematica.

This is the final version, to appear in Commentarii Mathematici Helvetici., Comment: 17 pages, in English

This survey, which contains very few proofs, addresses the general question: Over a given type of field, is there a natural class of varieties which automatically have a rational point? Fields under consideration here include: finite fields, p-adic fields, function fields in one or two variables over an algebraically closed field. Classical answers are given by the Chevalley-Warning theorem and by Tsen's theorem. More general answers were provided by a theorem of Graber, Harris and Starr and by a theorem of Esnault. The latter results apply to rationally connected varieties. We discuss these varieties from various angles : weak approximation, R-equivalence, Chow group of zero-cycles. Ongoing work on `rationally simply connected' varieties over function fields in two variables is also mentioned. A common thread in this report is the study of the special fibre of a scheme over a discrete valuation ring: if the generic fibre has a simple geometry, what does it imply for the special fibre?, Comment: Lecture notes for a CIME summer school (Cetraro, September 2007), 56 pages, in French

B. Poonen recently produced smooth threefolds over a number field which do not have a rational point but have no Brauer-Manin obstruction even after descent to a finite 'etale cover. In this note I show that the varieties he produces have zero-cycles of degree 1. ----- B. Poonen a r'ecemment exhib'e des exemples de vari'et'es projectives et lisses de dimension 3 sur un corps de nombres qui n'ont pas de point rationnel et pour lesquelles il n'y a pas d'obstruction de Brauer-Manin apr`es rev^etement fini 'etale. Dans cette note, je montre que les vari'et'es qu'il construit poss`edent des z'ero-cycles de degr'e 1., Comment: 8 pages

An integer may be represented by a quadratic form over each ring of p-adic integers and over the reals without being represented by this quadratic form over the integers. More generally, such failure of a local-global principle may occur for the representation of one integral quadratic form by another integral quadratic form. We show that many such examples may be accounted for by a Brauer-Manin obstruction for the existence of integral points on schemes defined over the integers. For several types of homogeneous spaces of linear algebraic groups, this obstruction is shown to be the only obstruction to the existence of integral points. ----- Une forme quadratique enti\`ere peut \^etre repr\'esent\'ee par une autre forme quadratique enti\`ere sur tous les anneaux d'entiers p-adiques et sur les r\'eels, sans l'\^etre sur les entiers. On en trouve de nombreux exemples dans la litt\'erature. Nous montrons qu'une partie de ces exemples s'explique au moyen d'une obstruction de type Brauer-Manin pour les points entiers. Pour plusieurs types d'espaces homog\`enes de groupes alg\'ebriques lin\'eaires, cette obstruction est la seule obstruction \`a l'existence d'un point entier., Comment: 53 pages, in English

The "canonical dimension" of an algebraic group over a field by definition is the maximum of the canonical dimensions of principal homogenous spaces under that group. Over a field of characteristic zero, we prove that the canonical dimension of the projective linear group PGL_6 is 3. We give two distinct proofs, both of which rely on the birational classification of rational surfaces over a nonclosed field. One of the proofs involves taking a novel look at del Pezzo surfaces of degree 6.

Let $k$ be a field of characteristic zero and ${\bar k}$ an algebraic closure of $k$. For a geometrically integral variety $X$ over $k$, we write ${\bar k}(X)$ for the function field of ${\bar X}=X\times_k{\bar k}$. If $X$ has a smooth $k$-point, the natural embedding of multiplicative groups ${\bar k}^*\hookrightarrow {\bar k}(X)^* $ admits a Galois-equivariant retraction. In the first part of the paper, over local and then over global fields, equivalent conditions to the existence of such a retraction are given. They are expressed in terms of the Brauer group of $X$. In the second part of the paper, we restrict attention to varieties which are homogeneous spaces of connected but otherwise arbitrary algebraic groups, with connected geometric stabilizers. For $k$ local or global, for such a variety $X$, in many situations but not all, the existence of a Galois-equivariant retraction to ${\bar k}^*\hookrightarrow {\bar k}(X)^* $ ensures the existence of a $k$-rational point on $X$. For homogeneous spaces of linear algebraic groups, the technique also handles the case where $k$ is the function field of a complex surface., Comment: To appear in Duke Mathematical Journal. An appendix on the Brauer-Manin obstruction for homogeneous spaces has been added

Let K be a global field, T a K-torus and S a finite set of places of K. Let K_v be the completion at a place v. Denote by T(O_v) the maximal compact subgroup of the group T(K_v) of K_v-points of T. We show that the diagonal map from T(K) to the product for v in S of the T(K_v)/T(O_v) need not be onto. As a corollary, for suitable v, the group T(O_v) does not cover all R-equivalence classes in T(K_v). While studying the height zeta functions of an algebraic torus over a function field in one variable over a finite field, D. Bourqui showed that Peyre's constant from the number field case should be multiplied by an integer. The same type of torus as constructed for the above problem enables us to show that this integer need not always be 1. ----- Soient K un corps global, T un K-tore, S un ensemble fini de places de K. On note K_v le compl\'et\'e de K en une place v. Soit T(K), resp. T(K_v), le groupe des points K-rationnels, resp. K_v-rationnels, de T. Notons T(O_v) le sous-groupe compact maximal de T(K_v). Nous montrons que pour T et S convenables l'application diagonale de T(K) vers le produit pour v dans S des T(K_v)/T(O_v) n'est pas surjective. Cela implique que pour v convenable le groupe T(O_v) ne couvre pas forc\'ement toutes les classes de R-\'equivalence de T(K_v). Lorsque K est un corps de fonctions d'une variable sur un corps fini, en utilisant le m\^eme type de tore, nous montrons que le facteur suppl\'ementaire par lequel D. Bourqui doit, lors son \'etude de la fonction z\^eta des hauteurs des vari\'et\'es toriques, multiplier la constante de Peyre, n'est pas toujours \'egal \`a 1., Comment: 12 pages, in French, to appear in Annales de l'institut Fourier

Ax gave examples of fields of cohomological dimension 1 which are not C_1-fields. Kato and Kuzumaki asked whether a weak form of the C_1-property holds for all fields of cohomological dimension 1 (existence of solutions in extensions of coprime degree rather than existence of a solution in the ground field). Using work of Merkur'ev and Suslin, and of Rost, D. Madore and I produced examples which show that the answer is in the negative. In the present note, I produce examples which require less work than the original ones. In the original paper, some of the examples were given by forms of degree 3 in 4 variables. Here, for an arbitrary prime p>3, I use forms of degree p in p+1 variables., Comment: 5 pages, in English

Let X be a smooth compactification of a connected linear algebraic group over a field k. The Chow group of degree nought zero-cycles on X is a torsion group. When k is a p-adic field, we show that the prime-to-p component of this group is finite. ----- Sei X eine glatte Kompaktifizierung einer zusammenhaengenden linearen Gruppe ueber einem Koerper k. Die Chowgruppe der nulldimensionalen Zyklen von X vom Grad Null ist eine Torsionsgruppe. Wir zeigen : wenn k ein p-adischer Koerper ist, dann ist der prim-zu-p Anteil dieser Gruppe endlich., Comment: 14 pages, in French, with a German summary. To appear in Inventiones mathematicae

Let $X$ be an algebraic variety, defined over the rationals. This paper gives upper bounds for the number of rational points on $X$, with height at most $B$, for the case in which $X$ is a curve or a surface. In the latter case one excludes from the counting function those points that lie on lines in the surface. The bounds are uniform for all $X$ of a given degree. They are best possible in the case of curves. As an application it is shown that if $F$ is an irreducible binary form of degree 3 or more then almost all integers represented by $F$ have essentially one such representation., Comment: 46 pages, published version; appendix by J.-L. Colliot-Th\'el\`ene

For each integer d=2,3,4, there exists a field F with cohomological dimension 1 and a del Pezzo surface of degree d over F having no rational point. Proofs use the theorem of Merkur'ev and Suslin, the Riemann-Roch theorem on a surface and Rost's degree formula. ----- Pour chaque entier d=2,3,4, il existe un corps F de dimension cohomologique 1 et une surface de del Pezzo de degre d sur F sans point rationnel. Les demonstrations utilisent le theoreme de Merkur'ev et Suslin, le theoreme de Riemann-Roch sur une surface et la formule du degre de Rost., Comment: 21 pages, in French; to appear in the Journal de l'Institut Mathematique de Jussieu; second version adds a new section refining earlier results

Let k be a p-adic field. Some time ago, D. Harbater [9] proved that any finite group G may be realized as a regular Galois group over the rational function field in one variable k(t), namely there exists a finite field extension $F/k(t)$, Galois with group G, such that F is a regular extension of k (i.e. k is algebraically closed in F). Moreover, one may arrange that a given k-place of k(t) be totally split in F. Harbater proved this theorem for k an arbitrary complete valued field. Rather formal arguments ([10, \S 4.5]; \S2 hereafter) then imply that the theorem holds over any `large' field k. This in turn is a special case of a result of Pop [15], hence will be referred to as the Harbater/Pop theorem. We refer to [10], [16], [6] for precise references to the literature (work of D\`ebes, Deschamps, Fried, Haran, Harbater, Jarden, Liu, Pop, Serre, and V\"olklein). Most proofs (see [10], [19, 8.4.4, p.~93] and Liu's contribution to [16]; see however [15]) first use direct arguments to establish the theorem when G is a cyclic group (here the nature of the ground field is irrelevant), then proceed by patching, using either formal or rigid geometry, together with GAGA theorems. In the present paper, where I take the case of algebraically closed fields for granted, I show how a technique recently developed by Koll\'ar [12] may be used to give a quite different proof of the Harbater/Pop theorem, when the `large' field k has characteristic zero. This proof actually gives more than the original result (see comment after statement of Theorem 1)., Comment: 15 pages