Ulrich bundles on blowing up (and an erratum) (Q1690938): Difference between revisions

The paper under review is an accompaniment to the second author's previous paper on Ulrich bundles of blowups [C. R., Math., Acad. Sci. Paris 354, No. 12, 1215--1218 (2016; Zbl 1388.14125); erratum ibid. 355, No. 12, 1291--1297 (2017)]. The authors combine Theorem 0.1 of [loc. cit.] with \textit{I. Bauer}'s analysis of inner-projection surfaces [J. Reine Angew. Math. 460, 1--13 (1995; Zbl 0841.14041)] to prove that except possibly for some special \(K3\) surfaces, any surface of degree at least 4 in \(\mathbb{P}^4\) which results from inner projection is Ulrich-wild, i.e. admits families of non-isomorphic indecomposable Ulrich bundles having arbitrarily large dimension. They also prove that if \(\rho : \widetilde{X} \to X\) is a blowup of a smooth projective variety \(X\) at a point, \(E\) is the exceptional divisor, \(\mathcal{O}(h)\) is a very ample line bundle on \(X\) whose strict transform \(\mathcal{O}(\widetilde{h})\) via \(\rho\) is very ample, and \(\mathcal{E}\) is a vector bundle of rank \(r\) on \(\widetilde{X}\) which is Ulrich with respect to \(\mathcal{O}(\widetilde{h})\) and satisfies \(\mathcal{E} \otimes \mathcal{O}_{E} \cong \mathcal{O}_{E}(1)^{\oplus r},\) then \(\rho_{\ast}\mathcal{E}\) is a vector bundle on \(X\) which is Ulrich with respect to \(\mathcal{O}(h).\) This was erroneously stated without the hypothesis on \(\mathcal{E} \otimes \mathcal{O}_{E}\) in the aforementioned paper of the second author; examples are given in the present paper to show the hypothesis cannot be removed.