measure of a frequency selective surface

Hi, I have a question about a measurement I did on some frequency selective surfaces. I measured their transparency and found a perfect match with respect the simulated behavior in CST Microwave Studio: there is a transparency window around several frequency (the fundamental frequency f0 and the grating lobes). I also measured their reflection in order to compare it with the reflection by a metal plate of the same size: the FSS reflects at any frequency like a metal plate and obviously there is no reflection at the frequency f0 (actually there is just a very weak reflection). There is an unexpected behavior around the grating lobes, because at that frequencies the FSS is transparent, but it reflects just like a metal plate. Can anyone explain me this? Thank you!