Pre-Registration v2.0: Plant Architecture Inversion via Coherence Gradient Field Specification — Five-Scenario Two-Chamber Cone Membrane System (Operative Protocol v2.3)

This document is the version 2.0 pre-registration of the Plant Architecture Inversion experiment, superseding v1.0 (DOI: 10.5281/zenodo.19040399, 2026-03-13). The primary prediction is unchanged from v1.0: a radish seed (Raphanus sativus) placed in an inverted coherence gradient field — hydroponic foam above, UV-A light embedded in reticulated foam below, complete external light deprivation — will germinate with root growing upward and shoot growing downward. Architecture Inversion Score (AIS) = 2 confirmed at T=72h and maintained at T=168h. The experimental apparatus has been substantially upgraded from the agar-cup system of v1.0 to a two-chamber opaque PVC pipe system with a Mylar-wrapped cone membrane signal separator. The cone membrane physically separates the S1 (top) and S2 (bottom) substrate chambers, prevents liquid and UV-A signal migration between chambers, and orients the germinating seed at the signal boundary. The number of experimental scenarios has been expanded from four to five, forming a structured comparison space over two variables: UV signal direction and signal separation vs co-location. Three substrate types and one free UV condition are defined:— Free UV: UV-A LED free in S1 air space, zero penetration cost. Scenario 1 S1 only.— Type A: Reticulated foam 20ppi + hydroponic mineral solution. Water and minerals only.�� Type B: Reticulated foam 20ppi + UV-A LED embedded 10–20mm depth. UV only, no hydroponic.— Type C: Reticulated foam 20ppi + hydroponic solution + UV-A LED embedded. All signals co-located. The five scenarios are:— Scenario 1 (true normal control): Free UV above, Type A hydroponic below. AIS=0 required. Validity gate for every run.— Scenario 2: Type B UV-only foam above, Type A hydroponic below. Normal architecture + active shoot penetration upward through UV foam.— Scenario 3 (PRIMARY CLAIM): Type A hydroponic above, Type B UV-only foam below. Full inverted coherence field. AIS=2 predicted.— Scenario 4: Type C combined foam above, void below. All signals above.— Scenario 5: Void above, Type C combined foam below. Maximum inversion signal, no competing attractor. The central design principle is signal deprivation: Scenarios 2–5 are in complete external light deprivation. The embedded UV-A LED is the only photonic signal in the coherence field. The seedling develops in skotomorphogenic state — maximum photoreceptor sensitivity, zero ambient noise — making the embedded UV-A source effectively infinite signal-to-noise. All component mechanisms are independently confirmed in published literature: root upward toward water in radish (Takahashi et al. 2003, Plant Physiology); MIZ1-mediated gravitropism suppression under drought (Zhang et al. 2025, PNAS 122:e2427315122); root negative phototropism to UV-A (Silva-Navas et al. 2016, PNAS); root positive phototropism unmasked in microgravity (Kiss et al. 2012, NASA TROPI ISS); shoot positive phototropism toward buried light (Christie 2007); photosynthesis from buried geothermal light (Beatty et al. 2005, PNAS); radiotropism toward embedded radiation (Dadachova et al. 2007, PLoS ONE). No mechanism is novel in isolation. The novel claim is: the combination of these mechanisms under a pre-specified inverted coherence gradient field produces simultaneous full architectural inversion of both root and shoot, predicted from a geometric attractor framework before any experiment is run, confirmed across a five-scenario structured comparison that eliminates alternative explanations. This pre-registration is part of the OrganismCore research program testing the geometric claim: Genome + Coherence Gradient Field → Organism. Change the field — the genome executes a different program. Same genome. Different field. Different organism. Linked LECA pre-registration DOI: 10.5281/zenodo.18986790Repository: github.com/Eric-Robert-Lawson/attractor-oncologyOperative Protocol: v2.3 (2026-03-18)No experimental data exists at time of upload.