In conclusion, the study by Rosenberger et al. (1997) has shown that it is now possible to realistically model vapour transport in real (three-dimensional) enclosures, at least at low combined Grashof numbers.
There have been several numerical studies of plane vertical square cavities.
The predicted overall transport rates, with some real exceptions, are
reasonably correlated by simple scale analysis.
Results for taller cavities are much
scarcer; the largest value of 
uncovered in the literature is 7 (Wee et al. 1989; §3.3.10). I have not found any results for
three-dimensional transport in cuboids. Experimental confirmation of the
results for squares, rectangles or cuboids is also notably lacking.
The available theoretical results for horizontal nonisothermal vapour transport in a gravitational field are restricted to two-dimensional geometries and cases where the mass transfer rate factor is vanishingly small. The present work supplies some results for finite mass transfer rates (ch. 4) and three-dimensional geometries (chapters 7-8).
Many of the studies reviewed did not consistently treat the related mass transfer effects of transpiration, interdiffusion and a variable mixture specific heat. This is treated in great detail in chapters 2, 4 and 6. It is shown that, in general, these should either all be included or all be neglected.