Define $x$~$y$ if $xy$ are the sum of $2^n$ squares in a field $K$. Is this an equivalence relation? The only non-trivial thing to verify is transitivity, and I have no clues of whether it is true or not, let alone proving it or providing a counter-example. Could somebody give a hint? Thanks in advance.
P.S. I've verified it in the case of $n=1$ using complex number and $n=2$ using quaternion, but don't know how to continue. I think the difficulty is to define multiplication in $2^n$-dimension vector space over K.
Hint: You'll want to use Pfister's Theorem: In any field, the set of sums of $2^n$ squares is closed under multiplication.
Full Solution: