Phébus FP studies the phenomenology of severe accidents in water-cooled nuclear reactors. Tests cover fuel rod degradation and behaviour of fission-products released via the coolant system into the containment. Analysis using computer codes aims to identify modelling weaknesses. Regarding fission-product behaviour in the coolant circuit, analyses of tests FPT0 and FPT1 are presented that used a standard version of a code, ASTEC/SOPHAEROS, with input data based solely on measured boundary conditions. Disagreements between calculated and experimental results are explored and plausible explanations presented. It is shown that in laminar conditions for a supersaturated vapour with competing homogeneous nucleation, heterogeneous nucleation and condensation on structures, codes can significantly underestimate the latter if entrance effects are ignored. Where thermophoresis dominates in hydrodynamically developed, weakly turbulent flow, codes can overestimate deposition; the likely explanation is underestimated mechanical resuspension. Concerning the transport of vapour species, it is shown that observations are compatible with passage of non-negligible amounts of the gas hydrogen iodide through the circuit. The final aspect of the exercise concerns deposit remobilization where this was different in the two tests and the understanding of which remains more speculative. Explanation invokes vibration of the apparatus producing mechanical resuspension in FPT0 and steam reacting with caesium deposits producing caesium hydroxide in FPT1.