lamina

The outermost neuropil of the optic lobe, directly behind the retina. The lamina is the first thing the fly's eye talks to, and the simplest tissue in this program's parts list. That the simplest part produced the record we cannot publish is a fact this archive presents without interpretation.

anatomy

The fly's compound eye is built from about 750 unit eyes, and the lamina repeats itself once per unit: a crystalline array of identical cartridges, each receiving the photoreceptors from one point in visual space and each containing the same small cast of neurons, the lamina monopolar cells L1 through L5, plus a few wide-field and feedback elements. The circuit's work is bookkeeping. It subtracts the average, so that what leaves the lamina is not brightness but change in brightness; it adapts its gain so the same circuit works at noon and at dusk; and it hands the medulla the beginnings of the ON and OFF channels, brightening and darkening, that all downstream vision is built from.

There is no memory here, no valence, no decision. Among anatomists the lamina is the region you learn first because it is the region where everything is known.

    retina    ▒▒  ▒▒  ▒▒  ▒▒  ▒▒  ▒▒  ▒▒  ▒▒
              ┌──┐┌──┐┌──┐┌──┐┌──┐┌──┐┌──┐┌──┐
              │L₁││L₁││L₁││L₁││L₁││L₁││L₁││L₁│   ~750 identical
              │⋮ ││⋮ ││⋮ ││⋮ ││⋮ ││⋮ ││⋮ ││⋮ │   cartridges,
              │L₅││L₅││L₅││L₅││L₅││L₅││L₅││L₅│   one per point
              └┬─┘└┬─┘└┬─┘└┬─┘└┬─┘└┬─┘└┬─┘└┬─┘   of sky
               ▼   ▼   ▼   ▼   ▼   ▼   ▼   ▼
           output = input minus the usual

fig. la-1 · a crystal of identical bookkeepers. the region reports only differences. see log 014 for what the differences reported.

in the fly

Contrast, adaptation, and temporal sharpening, applied uniformly across the visual field, roughly eight hundred times in parallel. The lamina does not see things. It prepares seeing.

grafted

LA-05 was bound on 2026-01-13, in the same week as ME-06, as the first pair of a planned optic-lobe series. The dossier for LA-05 is the shortest in the archive and is mostly redaction. The status page lists the agent as "redacted", which is a status this program did not have until it needed one.

What this page can state: the graft ran, and runs. First light occurred on 2026-01-16 and its transcript is withheld. Log 014 is the only published session, in heavily elided form. The mundane explanation on file involves a reconstruction artifact, a merger error in the region's wiring (see reconstruction), and the reader is encouraged to find it sufficient. The errata contain one entry about LA-05, and it is the only erratum in the archive that removes information rather than correcting it.

Obs. R, who ran the LA-05 sessions and no longer runs sessions, is on record within the program objecting to the redactions. The objection is noted here because obs. R asked that it be, and the program's rules give observers that right. The transcripts remain withheld.