Gravitational Manipulation Theory v1.4: A Source-Resolved Screening Framework and Its Consistent, Per-Experiment Observational Mapping

Gravitational Manipulation Theory (GMT) v1.4 develops a source-resolved framework for mapping environment-dependent scalar screening into observational constraints. Starting from a scalar–tensor action with the nonminimal curvature coupling F(χ) = M_Pl² + κ_Dχ² and a minimal double-well plus density-tilt potential, this work fixes the canonical weak-field normalization using the full kinetic factor K, validates the post-Newtonian normalization against the exact Brans–Dicke limit, and derives the single-body screening factor κ_s from a radial boundary-value problem. It further calibrates the deep-screening saturation κ_s,∞ ≃ 2.37 (m_out/m_in)^1.87 and incorporates a numerically calibrated two-body mutual-screening correction R_mutual. The central methodological rule is that each experiment must be mapped through its own physical coordinates, body-screening factors, environmental conditions, and apparatus geometry. No screening factor is transferred between experiments. The experiment-specific factor is G^(i) = κ_s^(s,i) κ_s^(p,i) R_mutual^(i), which is combined with the corresponding published observational bound to obtain the profiled maximum coupling α_DEF,max^(i). The framework is applied to three observational channels: • the Cassini solar-system measurement;• the Eöt–Wash 2020 inverse-square-law experiment;• the HUST-2020 torsion-pendulum experiment. On the adopted 400 × 400 theory-plane grid, the mapped coverage contains 47,904 Cassini cells, 2,537 Eöt–Wash cells, and 2,912 HUST cells. Cassini is disjoint from the two laboratory experiments, while Eöt–Wash and HUST independently upper-bound a common 319-cell region through different experimental geometries. Their strongest profiled reaches in α_DEF,max are approximately 6 × 10⁻³, 0.38, and 0.040, respectively. GMT v1.4 establishes a consistent source-resolved, per-experiment observational mapping and constraint framework. It does not establish full parameter-space exclusion, validity for arbitrary environmental potentials, experimental detection of a new gravitational interaction, experimentally demonstrated gravitational manipulation, or engineering controllability. All quantitative results are conditional on the minimal environmental-potential class and the modeling assumptions documented in the manuscript. This Zenodo deposit includes the complete manuscript, reproducible LaTeX source package, all 17 publication figures, a consolidated vector figure collection, README, and CC BY 4.0 license.