Not yet reviewed by Pith; the record is open.

The opacity due to grains in the envelope of a protoplanet regulates the accretion rate of gas during formation, thus the final bulk composition of planets with primordial H/He is a function of it. Observationally, for exoplanets with known mass and radius it is possible to estimate the bulk composition via internal structure models. We first determine the reduction factor of the ISM grain opacity f_opa that leads to gas accretion rates consistent with grain evolution models. We then compare the bulk composition of synthetic low-mass and giant planets at different f_opa with observations. For f_opa=1 (full ISM opacity) the synthetic low-mass planets have too small radii, i.e., too low envelope masses compared to observations. At f_opa=0.003, the value calibrated with the grain evolution models, synthetic and actual planets occupy similar mass-radius loci. The mean enrichment of giant planets relative to the host star as a function of planet mass M can be approximated as Z_p/Z_star = beta*(M/M_Jup)^alpha. We find alpha=-0.7 independent of f_opa in synthetic populations in agreement with the observational result (-0.71+-0.10). The absolute enrichment level decreases from beta=8.5 at f_opa=1 to 3.5 at f_opa=0. At f_opa=0.003 one finds beta=7.2 which is similar to the observational result (6.3+-1.0). We thus find observational hints that the opacity in protoplanetary atmospheres is much smaller than in the ISM even if the specific value of the grain opacity cannot be constrained here. The result for the enrichment of giant planets helps to distinguish core accretion and gravitational instability. In the simplest picture of core accretion where first a critical core forms and afterwards only gas is added, alpha=-1. If a core accretes all planetesimals inside the feeding zone, alpha=-2/3. The observational result lies between these values, pointing to core accretion as the formation mechanism.