I am trying to solve the following problem in Allen Hatcher's Algebraic Topology book :

  • Let $A_1, A_2, A_3$ be compact sets in $\mathbb R^3$. Use the Borsuk-Ulam theorem to show that there is one plane $P \subset \mathbb R^3$ that simultaneously divides each $A_i$ into two pieces of equal measure.

I managed to find the good idea, but I am stuck with this detail that I cannot find a way to explain. At some point in my proof, I am given a continuous function $f : S^2 \times [a,b] \to [0,1]$ such that for every $v \in S^2$, there exists $t \in [a,b]$ with $f(v,t) = 1/2$. Furthermore, I know that the map $f(v,-) : [a,b] \to [0,1]$ is decreasing and surjective (maybe this is not necessary, but it is how I have shown that $t$ with $f(v,t) = 1/2$ existed...).

I want to show the existence of a continuous map $\varphi : S^2 \to S^2 \times [a,b]$ such that $f \circ \varphi$ is the constant map at $1/2$ (I expect $\varphi(v) = (v,t(v))$ for some function $t : S^2 \to [a,b]$). If I can do this then my proof is complete and using Borsuk-Ulam's theorem I have proved the statement.

I must add that I expected $t(v)$ to be as follows : if $f(v,-)$ is such that $f(v,-)^{-1}(\{1/2\}) = [t_{min}(v),t_{max}(v)]$, I thought $$ t(v) = \frac{t_{min}(v) + t_{max}(v)}2. $$ This choice of $t(v)$ has the more elegant property that $f(-v,t(-v)) = f(v,-t(v))$ because of the way $f$ is defined, but I don't know how to show it will be continuous, even though I expect it to be.

Any ideas?


Here is an idea: the space $S^2 \times [a, b]$ can be thought of as a solid spherical shell of thickness $[a,b]$. Take your function $f:S^2 \times [a, b] \to [0,1]$, and define a function $g : S^2 \to S^2 \times [a, b]$ in the following way: for a given $v \in S^2$, $g(v) = (v, t_v)$ where $t_v$ is the smallest value of $t$ with $f(v, t_v) = 1/2$. Let us put $E=g(S^2)$. By construction, $E$ contains exactly one point in each fibre of the projection $S^2\times [a,b] \to [a, b]$. It is like a "slice" of the solid spherical shell $S^2 \times [a, b]$. The projection $S^2 \times [a, b] \to S^2$ restricts to a map $E \to S^2$. This map is an isomorphism with inverse $g$. By construction, the function $f$ is identically $1/2$ on $E$, so the inclusion $E \to S^2\times [a,b]$ is the map $\varphi$ you are looking for (once we identify $E$ with $S^2$).

  • $\begingroup$ Yes, that I figured! But why should $g$ be continuous? That is precisely where I am stuck and it is why I am asking this on MSE. $\endgroup$ – Patrick Da Silva Sep 24 '13 at 16:55
  • $\begingroup$ @PatrickDaSilva It was far from clear from your post that you had gotten this far. ;) Continuity of $g$ should be a consequence of the fact that $f(v, -)$ is decreasing. I'll let you figure it out! $\endgroup$ – Bruno Joyal Sep 24 '13 at 16:59
  • $\begingroup$ I agree that it wasn't clear from my question, I guess I should've explained it. But I really can't figure out why $g$ should be continuous... $\endgroup$ – Patrick Da Silva Sep 24 '13 at 17:00

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.