Skip to main content

Timeline for Flatness and intersection of fibers

Current License: CC BY-SA 3.0

9 events
when toggle format what by license comment
Mar 31, 2015 at 17:57 comment added Allen Knutson Let $X_1 = \{(x,0)\}$, $X_2 = \{(x,x)\}$, $f = $projection to the first factor. Then $X_1,X_2,X_1\cup X_2$ are all flat over $Y$, and their generic fibers do not intersect. But the intersection is supported over $0\in Y$. (Maybe this was Matthieu's example.)
Mar 31, 2015 at 16:40 comment added Karl Schwede You always have a short exact sequence$$ 0 \to O_{X_1 \cup X_2} \to O_{X_1} \oplus O_{X_2} \to O_{X_1 \cap X_2} \to 0. $$ This should tell you exactly what the obstruction is. Indeed, it immediately follows that if $X_1, X_2$ and $X_1 \cap X_2$ are flat, so is $X_1 \cup X_2$, but as you can see from the long exact sequence of Tor, you don't get to deduce that $\text{Tor}^1_{f^{-1} O_Y}(O_{X_1 \cap X_2}, N)$ is zero.
Mar 31, 2015 at 16:06 history undeleted Ron
Mar 31, 2015 at 16:06 history deleted Ron via Vote
Mar 31, 2015 at 14:57 history edited Ron
edited tags
Mar 31, 2015 at 14:06 history edited Ron CC BY-SA 3.0
deleted 6 characters in body; edited tags; edited title
Mar 31, 2015 at 13:56 comment added Ron @Romagny: I wanted to add the condition that $Y_1 \cup Y_2$ is also flat over $X$. I will edit the question. Is there a counterexample in that case as well?
Mar 31, 2015 at 13:22 comment added Matthieu Romagny Certainly not. Geometrically $Y_1=Spec(B_1)$ and $Y_2=Spec(B_2)$ are closed subvarieties in some affine space $\mathbb{A}^{n+1}_X$ with $X=Spec(A)$ and you are asking: is it true that $Y_1,Y_2$ flat over $X$ imply $Y_1\cap Y_2$ flat over $X$? A counterexample is e.g. $X=\mathbb{A}^1_{\mathbb{C}}$ and $Y_1,Y_2=$ the coordinate axes in $\mathbb{A}^2_{\mathbb{C}}$.
Mar 31, 2015 at 13:01 history asked Ron CC BY-SA 3.0