In 2005, it was found that the fluorescence of crystals of the major light-harvesting complex LHCII of green plants is significantly quenched when compared to the fluorescence of isolated LHCII (A. A. Pascal et al., Nature, 2005, 436, 134-137). The Raman spectrum of crystallized LHCII was also found to be different from that of isolated LHCII but very similar to that of aggregated LHCII, which has often been considered a good model system for studying nonphotochemical quenching (NPQ), the major protection mechanism of plants against photodamage in high light. It was proposed that in the crystal LHCII adopts a similar (quenching) conformation as during NPQ and indeed similar changes in the Raman spectrum were observed during NPQ in vivo (A. V. Ruban et al., Nature, 2007, 450, 575-579). We now compared the fluorescence of various types of crystals, differing in morphology and age. Each type gave rise to its own characteristic mono-exponential fluorescence lifetime, which was 5 to 10 times shorter than that of isolated LHCII. This indicates that fluorescence is not quenched by random impurities and packing defects (as proposed recently by T. Barros et al., EMBO Journal, 2009, 28, 298-306), but that LHCII adopts a particular structure in each crystal type, that leads to fluorescence quenching. Most interestingly, the extent of quenching appears to depend on the crystal morphology, indicating that also the crystal structure depends on this crystal morphology but at the moment no data are available to correlate the crystals' structural changes to changes in fluorescence lifetime. © the Owner Societies 2011.