Fine native ground with wider crop coverage

The exact native grass, damp-earth and meadow originals remain unchanged: WOG-A-003576, WOG-A-003578 and WOG-A-003579. This experiment compares the current crop distribution (frequency 4) with wider coverage (frequency 8) at the same private 1.25-hex logical span. Doubling the blend-lattice frequency and halving each source footprint preserves the native coordinate slope. It changes which painted patches are sampled; it does not repaint, rotate, downscale or blur the artist files.

Earlier wider-crop findings remain unresolved. This follow-up tests the newer fine-scale calibration together with that sampler, rather than treating earlier technical parity or a renderer label as artistic approval. The unchanged defaults were separately verified in the lake sampler investigation.

Three existing foreground-profile controls are used: cottage/grasslands, strata/wetlands and courtyard/fields. The read-only effective-world probe verifies two natural biome types, fields and wetlands. The cottage's grasslands profile label is not a third natural biome. Legacy sheet labels say “native shore”; these are center-ground crops and provide no shoreline-contact qualification.

The completed run contains 24 actual private desktop/phone Region/Hex views and 18 pixel-exact candidate pan-return/shared-view reconnect comparisons. Every one of the 12 source/cohort map pairs changes pixels outside UI. The two actual Hex foreground canvases per profile match exactly across cohorts, for six exact foreground comparisons. This is a separate layer check; it is not an isolated proof of unchanged Region foreground. Only three native context sheets are retained. Full maps and pixel comparisons stayed in memory.

The individual decisions are:

• Grass: the cottage-centered crops look very similar. There is real source contribution in all four map pairs, but these views do not establish a clear quality improvement or complete nonrepetition. Keep the original and current sampler, with stronger ground-dominant and canonical-context witnesses still needed.
• Wetland: wider crops disrupt some directional soil grouping at the inspected strata site while retaining fine angular stone, moss and stem detail. Investigate this scoped preference using the unchanged original. Other wetland states, marsh/water contacts, CPU fallback and performance remain open.
• Meadow: clump distribution changes at the courtyard site and the painted marks remain visible, but a clear quality preference across both view scales is unproven. Keep the original and all remaining physical-scale, palette, recurrence and performance gates.

No public renderer, source permission, artist file, geography, gameplay rule or persistent production save changes. No new game-ready approval is granted. All previous approvals, next actions and 43 deferred local runtime declarations remain intact. The three new source-bound findings are queued for Atlas publication independently of the already-building lake-shore release.

See the individual visual decisions visual-review.json, normal-client observations and comparisons normal/results.json, effective read-only probe normal/read-only-effective-probe.json, exact driver derivation normal/driver-derivation.json, owned process closeout normal/owned-resources-terminal.json and registration registration.json.