Current hypotheses on the origin of the fungus–attine ant (Hymenoptera: Formicidae) symbiosis propose, as an ancestral first step in the development of fungal cultivation, fortuitous feeding on fungi growing adventitiously on substrates such as rotting wood, insect parts, seeds stored by ants in nests, regurgitated infrabuccal pellets, free-living soil fungi, or mycorrhizae. However, feeding-deterrent fungi regularly colonize these substrates. De novo feeding on these fungi by the attine ancestor is unlikely because the almost universal presence of mycotoxins on adventitious fungi is a formidable barrier to mycophagy. In addition, there is no evolutionary history of mycophagy in the Hymenoptera. Instead, I propose that attine mycophagy began from opportunistic, selective feeding on wood-colonizing fungi previously “domesticated” by other insects: ambrosia beetles: (Coleoptera: Curculionidae: Scolytinae or Platypodinae); and less likely, woodwasps (Hymenoptera: Siricoidea). Attine ancestors foraged for beetle brood and fungal biomass in the galleries of those insects, which provided the attine ancestor with fungi that were nutritious and nontoxic to insects. The invading ants’ debris (fecal spots, exuviae, meconia, and saliva) possibly allowed the growth of the “good” domesticated fungi in galleries, while ants mechanically eliminated undesirable ones. Feeding on beetle fungi first allowed the development of broad mycophagy in ants, and, later, of the gardening habits. Subsequently, more restricted, specific mycophagy evolved. Only after serious barriers were overcome did incipient mycophagy develop, followed by cultivation. The ambrosia fungi-lined beetle galleries in wood provide one location favorable for this sequence. An analogous progression in mycophagy and incipient fungus manipulation, departing from the omnivore diet has occurred in Megalomyrmex ants that feed parasitically on brood and/or fungus gardens of attines. Colonies of the ancestral attine nested in wood and adopted mated foundress queens after mating flights, and colonies reproduced by budding. Queen adoption by established colonies allowed the chronological continuation of the incipient symbiosis. Mycophagy preadapted the ants to test free-living fungi in wood and soil as cultivars. The fortuitous finding of more adequate lepiotaceous strains in soil allowed the adaptive radiation exhibited in present-day attines. The cultivation, by the primitive attine Apterostigma, of wood-colonizing Basidiomycotina is possibly of ancestral significance.