In a series of publications, L. A. Paul has defended a version of the bundle theory according to which material objects are nothing but mereological sums of ‘their’ properties. This ‘mereological’ bundle theory improves in important ways on earlier bundle theories, but here I present a new argument against it. The argument is roughly this: Material objects occupy space; even if properties have spatial characteristics, they do not quite occupy space; on no plausible construal of mereological composition does a mereological (...) sum of non-space-occupying entities occupy space; therefore, material objects are not mereological sums of properties. (shrink)