The flowering plants (angiosperms) are the dominant and defining group of the Earth ecosystems today. However, from which group and by what way flowers, especially their gynoecia (the key characteristic organs of angiosperms), are derived have been key questions in botany, and have remained unanswered despite botanists' efforts over centuries. Such an embarrassing situation can be attributed to the lack of plants with partially enclosed ovules, which are supposed fill a position between gymnosperms and angiosperms. Here, we report a fossil plant that has apparent coniferous vegetative and reproductive characters but has a single seed partially wrapped by the subtending bract. Such a morphology suggests that a carpel of some angiosperms is equivalent to a lateral appendage (a bract plus its axillary seed) of this fossil. Such a non-traditional interpretation of the homology of angiosperm carpels is compatible with various new progresses made in botany and is in line with Tomlinson's recent hypothesis. Together with other fossil evidence reported recently, it appears that gynoecia in angiosperms are derived in multiple ways.