Suppose a finite group $G$ has more than one Sylow-2 subgroup and any two intersects trivially. Show $G$ contains exactly one conjugacy class of involutions.
Here are some of my thoughts: Suppose not, then there exist involutions $s,t$ such that they are not conjugate.
$s,t$ generates a dihedral group $D_{2n}$. $n$ is even since $s,t$ are not conjugate in $D_{2n}$.
If $n$ is not a power of 2, then $D_{2n}$ has more than one Sylow 2 subgroups and $D_{2n}$ are generated by them. But any two of the Sylow 2 subgroups of $D_{2n}$ has a nontrival intersection, and any two Sylow-2 subgroups of G intersects trivally, so all Sylow 2 subgroups of $D_{2n}$ must lie in the same Sylow-2 subgroup of $G$. So $D_{2n}$ lie in a Sylow-2 subgroup of $G$, this leads a contradiction.
But what if $n$ is a power of 2?
Thanks to a comment, I have solved it as follows.
My argument shows that any two non-conjugate elements generates a dihedral group whose order is a power of 2. Thus $s,t$ and $s,gtg^{—1}$ lies in a same Sylow-2 subgroup of G, for all $g\in G$. Thus it has only one Sylow-2 subgroup, contradiction.