In 2008, Blackiston, Silva Casey, and Weiss published the first conclusive demonstration: fifth-instar Manduca sexta caterpillars conditioned with ethyl acetate paired to a mild shock retained the aversion as adult moths, 77 percent choosing ambient air in the Y-maze, with controls excluding chemical carryover, odor applied to naive pupae changed nothing, washing trained pupae removed nothing. The paper's second documented finding is the sharper one: caterpillars trained at the third instar kept the memory through two molts but lost it in pupation, meaning the memory survives metamorphosis only if formed late enough, consistent with the adult brain being built from structures laid down in late larval life, the mushroom bodies. The mechanism remains open: which synapses survive the soup is unresolved. Nagai's 2024 work extended the effect to butterflies and to the next generation, filed in this wing's own entries.
The folk force of the finding is the image: the chrysalis as total dissolution, the popular account's goo, followed by a butterfly that remembers the smell of danger, the persistence of memory through the dissolution of the rememberer being the strongest documented biological metaphor for every survival question the archive files, the soul-trap and Bardo shelves reading it as the animal kingdom's own proof that something outlasts the restructuring. The tradition's caution is documented in the paper itself: the soup model is partly a popular exaggeration, many larval neurons persist, remodeled, into the adult, so the finding may be less miracle than plumbing, some of the pipes never melted.
The claim: the study's least-quoted result is its deepest: early memories survive the molts but not the metamorphosis, the caterpillar's childhood being erased at the exact boundary the late memory crosses, so that the insect has, documented, a larval amnesia wall, what you learn too early is lost at the transformation, and the archive's whole harvest wing runs on the human version of the same wall. Verdict: the archive files the parallel as more than decoration: in both the insect and the traditions, what persists is what was laid down last and deepest before the dissolution, and the remembering, per the entomologists, lives in structures built late. The butterfly avoids the lavender. The moth avoids the ethyl acetate. The soup kept the warning. Nobody, per the literature, knows how. The archive's standing note applies: a memory with no known substrate is not a mystical memory, it is an unfunded one. Ask the caterpillar what it knows. It was ten years old, per Kobe, and it wrote back.
If the memory survives the dissolution of the brain that stored it, where was it stored, and what does memory actually consist of?