Some confusion here!
Where an electromag wave travels through an aperture smaller than its wavelength, then the far field beyond the aperture is much decreased, and the wave cannot propogate. Propogation isn't actually zero (which would need an infinitely small aperture, in effect a solid surface), but it becomes so close to zero as makes no difference.
With a sat dish, if the energy can't propogate through, then it must be reflected back instead (conservation of energy!).
As suggested, the lower efficiency of a mesh dish is due to surface irregularity. If the holes were perfect apertures, they would have to be razor sharp, so instead they're made rounded, so reflection becomes more diffuse, and also the apertures are no longer perfect.
See:
http://www.physlink.com/Education/AskExperts/ae176.cfm .
(Added, rather than just quote equations, a better way to follow this is via more familiar Fourier optics, which is exactly the same physics, except the wavelengths are much smaller!
The far field pattern, beyond an aperture, is an Airy disc:
http://en.wikipedia.org/wiki/Airy_disk .
As the aperture size decreases, the central spot also decreases in intensity (reduced propogation), corresponding to a smaller value of I in the Frauenhoffer intensity. But, when k becomes very large, effectively you get just a single very bright spot, corresonding to "straight through" propogation).
PS, here's a nice interactive demo, but your web browser will need Java:
http://micro.magnet.fsu.edu/primer/java/mtf/airydisksize/index.html .
PPS, filling in dish holes won't work! But, for entirely different reasons. Where two different surfaces meet, there's a boundary, hence an impedence mismatch.
