During the mid-Cretaceous, angiosperms diversified from several nondiverse lineages to their current global domination [1], replacing earlier gymnosperm lineages [2]. Several hypotheses explain this extensive radiation [3], one of which involves proliferation of insect pollinator associations in the transition from gymnosperm to angiosperm dominance. However, most evidence supports gymnosperm-insect pollinator associations, buttressed by direct evidence of pollen on insect bodies, currently established for four groups: Thysanoptera (thrips), Neuroptera (lacewings), Diptera (flies), and now Coleoptera (beetles). Each group represents a distinctive pollination mode linked to a unique mouthpart type and feeding guild [4-9]. Extensive indirect evidence, based on specialized head and mouthpart morphology, is present for one of these pollinator types, the long-proboscid pollination mode [10], representing minimally ten family-level lineages of Neuroptera, Mecoptera (scorpionflies), and Diptera [8, 10, 11]. A recurring feature uniting these pollinator modes is host associations with ginkgoalean, cycad, conifer, and bennettitalean gymnosperms. Pollinator lineages bearing these pollination modes were categorized into four evolutionary cohorts during the 35-million-year-long angiosperm radiation, each defined by its host-plant associations (gymnosperm or angiosperm) and evolutionary pattern (extinction, continuation, or origination) during this interval [12]. Here, we provide the first direct evidence for one cohort, exemplified by the beetle Darwinylus marcosi, family Oedemeridae (false blister beetles), that had an earlier gymnosperm (most likely cycad) host association, later transitioning onto angiosperms [13]. This association constitutes one of four patterns explaining the plateau of family-level plant lineages generally and pollinating insects specifically during the mid-Cretaceous angiosperm radiation [12].