I apologize in advance as this is not a research level question but rather one which could benefit from expert attention but is potentially useful mainly to novice mathematicians.

I'm trying to compile a list of non-obvious theorems about morphisms of schemes which are useful for general intuition but whose proofs are not easy/technical. Here are some examples:

Zariski's Main Theorem: Let $Y$ be quasi-compact, separated and $f:X \to Y$ be separated, quasi-finite, finitely presented. Then there is a factorization $X \to Z \to Y$ where the first map is an open immersion and the second is finite. Mnemonic: (quasi-finite) $\sim$ (finite) $\circ$ (open immersion)

Nagata's compactification theorem: Let $S$ be qcqs and $f:X \to S$ be separated, finite type. Then $X$ densely embeds into a proper $S$-scheme.Mnemonic: non-horrible schemes have compactifications

Temkin's factorization theorem: Let $Y$ be qcqs and $f: X \to Y$ be separated, quasi-compact. Then there's a factorization $X \to Z \to Y$ with the first being affine and the second proper. Mnemonic: (separated + quasicompact) = (proper) $\circ$ (affine).

Chow's lemma: Let $S$ be noetherian and $f: X \to S$ separated finite type. Then there exists a projective, surjective $S$-morphism $\bar{X} \to X$ which is an isomorphism on a dense subset and where $\bar{X} \to S$ is quasi-projective. Moreover $X$ is proper iff $\bar{X}$ is projective, and if $X$ is reduced $\bar{X}$ can be chosen to be so as well. Mnemonic: reasonable schemes have quasi-projective "replacements" and proper schemes have projective "replacements"

Hopefully it's clear now what I'm looking for. All theorems above have very weak assumptions and very satisfying conclusions. These are what I'm after.

  • $\begingroup$ How about something like Chevalley's theorem? It says that a quasi-compact morphism which is locally of finite type preserves locally constructible sets. $\endgroup$ Mar 25, 2016 at 1:49
  • 3
    $\begingroup$ You could add Stein factorization to the list: (proper) = (integral) $\circ$ (proper with connected geometric fibers). stacks.math.columbia.edu/tag/03GX $\endgroup$ Mar 25, 2016 at 2:47
  • 11
    $\begingroup$ Open immersion = etale + monic (EGA IV$_4$, 17.9.1+epsilon), integral morphism = affine + universally closed (EGA IV$_4$, 18.12.8), closed immersion = proper monomorphism (EGA IV$_4$, 18.12.6), qcqs = relatively affine over finite type Z-scheme (Thomason-Trobaugh, Theorem C.9), relative ampleness for a proper fp map is fibral (EGA IV$_3$, 9.6.5), lfp map is unramifed iff etale-locally a closed immersion (EGA IV$_4$, 17.7.1, 18.8.3), qc separated map is immersion iff monic + Mochizuki's valuative criterion (Ch. I, sec. 2.4 of his book on p-adic Teichmuller theory in noetherian case). $\endgroup$
    – nfdc23
    Mar 25, 2016 at 7:46
  • 7
    $\begingroup$ Finitely presented flat map is finite iff quasi-finite & separated with locally constant fiber rank (II, 1.19 Deligne-Rapoport in noetherian case), lfp = commutes with "expected" limits (EGA IV$_3$, 8.14.2), valuative criterion for flatness over a reduced base (EGA IV$_3$, 11.8.1 in noetherian case with dvr's, 4.2.10 in Part I of Raynaud-Gruson in general), universal schematic dominance/density is fibral under "expected" flatness hypotheses (EGA IV$_3$, 11.10.9, 11.10.10), any flat lfp cover admits a section over a locally quasi-finite & flat lfp cover (EGA IV$_4$, 17.16.2). $\endgroup$
    – nfdc23
    Mar 25, 2016 at 8:35
  • 1
    $\begingroup$ Somoth/étale morphisms can be lifted (Zariski-locally on the source) along closed immersions (EGA IV_4, 18.1.1). $\endgroup$
    – AAK
    Mar 25, 2016 at 16:48

1 Answer 1


Here's a theorem I find useful:

Theorem. Let $\phi \colon X \to Y$ be a smooth morphism of schemes of relative dimension $d$. Then there exists an open cover $X = \bigcup U_i$ of $X$ such that each $U_i \to Y$ factors as $$U_i \stackrel \pi \to \mathbb A^d_Y \to Y,$$ with $\pi$ étale. Mnemonic: smooth morphisms have étale coordinates.

See Tag 054L.

  • 6
    $\begingroup$ EGA 4 vol. 4 Corollary 17.11.4 expresses this as a necessary and sufficient criterion for a map to be smooth at a point. $\endgroup$
    – S. Carnahan
    Mar 25, 2016 at 6:53
  • $\begingroup$ Shouldn't the $d_i$'s be constant here? $\endgroup$ Mar 25, 2016 at 12:41
  • $\begingroup$ @SaalHardali: they are (automatically) locally constant. $\endgroup$ Mar 25, 2016 at 21:49
  • $\begingroup$ @R.vanDobbendeBruyn Yes that's what I meant. Just wanted to make sure. Thanks. $\endgroup$ Mar 25, 2016 at 21:50
  • $\begingroup$ @R.vanDobbendeBruyn Under the same conditions, how likely would you say is the existence of a factorization $X \to E \to Y$ where the first map is etale and $E$ is some vector bundle over $Y$? $\endgroup$ Mar 27, 2016 at 1:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.