Think about a factory conveyor belt that keeps producing widgets and dropping them into a shredder at the end of the line. The belt runs continuously, widgets form on it, grow as they accumulate material, and then the belt's own motion drags them into the shredder. The only widgets that survive are the ones being made right as someone finally hits the OFF switch. That's the Canup & Ward model for the Galilean satellites: Jupiter's circumplanetary disk kept forming moons and migrating them inward to destruction, and Io, Europa, Ganymede, and Callisto are simply the last generation — the ones on the belt when the gas inflow stopped. The committed claim: the Galilean satellites formed slowly (over more than 10^5 years) in a low-pressure, low-gas-density circumjovian accretion disk during the final stages of Jupiter's gas inflow, and gravitational satellite-disk interactions drove inward orbital migration that destroyed earlier generations of moons. The system we see today is the surviving residue, not the cumulative product. This is not a first — Canup and Ward developed this 'gas-starved disk' model in earlier papers — but this 62-page review (for the University of Arizona Press 'Europa' volume) represents the mature synthesis, integrating their model with current constraints on giant planet formation. The ladder: the model this replaces is the 'minimum mass sub-nebula' (MMSN) approach, which treated the circumplanetary disk as a static, isolated structure analogous to the minimum-mass solar nebula. The MMSN model produces satellites too quickly (in too dense an environment) and cannot explain why the total satellite system mass relative to planet mass (~10^-4) is suspiciously similar across Jupiter, Saturn, and Uranus. The Canup & Ward gas-starved disk model explains this ratio as a natural outcome: satellite-disk interactions set a maximum satellite mass, and the total system mass converges to ~10^-4 of the planet mass regardless of specific formation history. That's the model's strongest prediction — it explains a ratio that the MMSN approach treats as coincidence. Architecturally, this is semi-analytic planetary science modeling: the disk is described by steady-state inflow models with parameterized viscosity and gas surface density, satellite growth is tracked through accretion timescales, and orbital migration is computed from Type I torques (the gravitational interaction between a satellite and the gas disk). The method leans on the physical constraints of gas giant formation — specifically, that Jupiter's gas accretion was throttled in its final stages by the supply rate from the solar nebula, producing a 'gas-starved' rather than gas-rich circumplanetary environment. This is the key architectural choice that distinguishes the model: the disk is actively fed and drained, not a closed system. Integrity check: this is theory constrained by observation, not a laboratory experiment. The 'data' are the observed properties of the Galilean system — four satellites, their masses, compositions (ice-to-rock ratios decreasing inward), orbital spacing, and the ~10^-4 mass ratio. The model is validated by whether it can reproduce these properties from physically motivated initial conditions. The circularity risk is moderate: the model parameters (inflow rate, viscosity, disk temperature profile) have some freedom to be tuned, though they are bounded by independent constraints from giant planet formation models. The strongest test is the mass ratio prediction across multiple planet systems (Jupiter, Saturn, Uranus), which the model explains without per-system tuning. The weakest point is that we have exactly three data points for that ratio. The milestone to track: Europa Clipper (launched October 2024, arriving at Jupiter ~2030) will measure Europa's ice shell thickness, ocean depth, and bulk composition with unprecedented precision. The specific number: if Europa's ice-to-rock ratio can be pinned down to within a few percent, it directly constrains the temperature and solid-to-gas ratio of the circumjovian disk at Europa's formation distance — converting a model parameter into a measured quantity. That's ~4 years away. Longer term, a Uranus orbiter (proposed but not funded) would test whether Uranus's satellite system mass ratio holds at the same ~10^-4 value with better precision than we currently have. The obvious successor experiment the authors didn't run: full 3D hydrodynamic simulations of satellite formation within a circumplanetary disk with realistic inflow boundary conditions, tracking multiple satellite generations forming, migrating, and being lost. My read: (a) compute constraints. In the mid-2000s when this model was developed, 3D hydro simulations of circumplanetary disks were just becoming feasible, and running them long enough to track multiple generations of satellite formation and loss was prohibitively expensive. The semi-analytic approach was the right tool for the available compute. Since then, groups like Szulágyi and others have run 3D hydro simulations of circumplanetary disks, but the full multi-generation satellite formation problem remains computationally demanding.
