The cosmic star formation history peaks at z of 2 to 3 and misbehaves at both ends: at high redshift JWST finds efficient star formation before the model's halos can assemble the fuel, with extreme systems demanding near-50 percent conversion efficiency against the roughly 20 percent observed ceiling (Madau and Dickinson 2014; Xiao 2024), while the post-noon decline outruns halo-assembly rates.
Star formation in the model waits on hierarchical assembly: no halo, no gas reservoir, no stars. The high-z end therefore cannot start early without impossible efficiencies, and the low-z end cannot stop without tuned quenching. Feedback prescriptions absorb both failures, recalibrated epoch by epoch, predicting neither.
SCT changes both boundary conditions at once. The high-z end: collision-seeded proto-structures carry their gas reservoirs from deposition (P22, P25), with mass set by the cascade geometry rather than accumulated by mergers, so star formation begins when the gas cools, not when hierarchical assembly catches up. The observed early efficiency is then ordinary physics on early fuel: the apparent near-50 percent efficiencies are a geometric signature of collision-seeded initial mass, not a real conversion rate, the same dissolution that carries the JWST mass crisis (recid 107) and the chemical-maturity timeline. The seeding mechanics are in Paper 4, From Chaos To Collisothermal Cosmogenesis.
The low-z end: as the parent-frame mesh dissipates, Λ_eff grows more dominant (P14, P17, P18), suppressing the large-scale gas inflow that feeds disks, a global throttle the model's halo-assembly rates do not contain. The peak at cosmic noon is the crossover: deposited reservoirs draining down while the Λ_eff throttle closes, two registered mechanisms whose intersection is the curve's shape rather than a feedback fit. The expansion-side machinery is in Paper 7, From Chaos To Cosmic Expansion.
Keystone economy: P25 fuels the early end, P18 closes the late end, and the peak falls out of their crossing. No epoch-by-epoch feedback retuning required.
Both ends are exposed. If deeper JWST and Roman censuses show the high-z star formation history collapsing onto hierarchical-assembly expectations with standard efficiencies, the seeded-reservoir reading is unnecessary. At the late end, SCT requires the post-noon decline to correlate with environment through the Λ_eff field: star formation should persist longer in dense, mesh-shielded regions than the global average, an environment-split signature tuned feedback does not predict; its absence in DESI and Euclid star-formation environment splits would refute the M5 throttle.