Plant–animal mutualisms such as seed dispersal can play an important role in enabling some species to become invasive. For example, an introduced species could become invasive because birds prefer its fruits to those of native plants. To investigate this possibility, we compared various measures of fruit quality of 22 of the most common native and invasive woody species on the oceanic island Mahé (Seychelles, Indian Ocean).Individual measures of food quality tended to vary much more amongst invasive species than amongst native species; thus, whereas for particular traits the fruits of some invasive species had higher values than any native species, others had relatively low values. However, invasive species consistently produced fruits with a lower water content, resulting in a higher relative yield (i.e. dry pulp weight to total wet fruit weight ratio), and a higher energy content. The fruits of the most abundant invasive tree Cinnamomum verum (Lauraceae) were of particularly high nutritional quality, with individual berries containing 3.5 times more protein and 55 times more lipid than the median values of the native species.We suggest that our results may reflect a general tendency for island plants to produce fruits of low energy content, perhaps reflecting reduced competition for dispersal agents on isolated islands. In addition, we argue that inconsistent results on the relevance of fruit quality for plant invasions reported in the literature may be resolved by comparing the widths of trait spectra for native and alien floras rather than average values. Gaps in the native fruit trait spectrum may provide opportunities for particular invasive species, and weaken the resistance of regional floras to invasions. Such empty niche opportunities may occur for several reasons, including generally broader trait spectra in globally assembled alien than regional native floras (especially on oceanic islands), or the loss of native species due to human activities. More generally, a focus on trait variation rather than average trends may help to advance generalisation in invasion biology.