Because the success of the VOF technique is based on the ability to numerically advect a step-function distribution (F) without numerical smoothing, it is worthwhile to investigate the sensitivity of SOLA-VOF to changes in the F-advection algorithm. A good problem for this purpose is the nonlinear development of a Rayleigh—Taylor instability. During the early stages of the instability the fluid surface moves normal to itself, but during the later stages there are regions along the sides of the growing liquid fingers where the flow is mostly tangential to the surface. Thus, this problem offers a good test of the particular combination of donor- and acceptor-cell fluxing used in the code.