Can we recover an integral quadratic form by representing all its subforms?

Abstract

Let $\mathfrak o$ be the ring of integers of a totally real number field. If $f$ is a quadratic form over $\mathfrak o$ and $g$ is another quadratic form over $\mathfrak o$ which represents all proper subforms of $f$, does $g$ represent $f$? We show that if $g$ is indefinite, then $g$ indeed represents $f$. However, when $f$ is positive definite and indecomposable, then there exists a $g$ which represents all proper subforms of $f$ but not $f$ itself. Along the way we give a new characterization of positive definite decomposable quadratic forms over $\mathfrak o$ and a number-field generalization of the finiteness theorem of representations of quadratic forms by quadratic forms over $\mathbb Z$ which asserts that given any infinite set $\mathscr S$ of classes of positive definite integral quadratic forms over $\mathfrak o$ of a fixed rank, there exists a finite subset $\mathscr S_0$ of $\mathscr S$ with the property that a positive definite quadratic form over $\mathfrak o$ represents all classes in $\mathscr S$ if and only if it represents all classes in $\mathscr S_0$.