The system limit is reached when the superficial vapor velocity in the tower exceeds the settling velocity of large liquid droplets. At higher vapor velocities, ascending vapor lifts and carries over much of the tray liquid, causing the tower to flood. This flood cannot be debottlenecked by improving the internals or by increasing tray spacing. The system limit represents the ultimate capacity limit of the vast majority of the existing trays and of all the existing packings. In some applications, where very open packings or trays are used, such as refinery vacuum towers, the system limit is the actual capacity limit. Until recently no published methods were available to predict it. The factors that affect it are poorly understood. In the late 1950s and early 1960s, Fractionation Research Inc. (FRI) systematically studied this limit in a 1.22 m diameter column and developed several different correlations. Publication of this first commercial scale 'system limit' data by FRI last year permitted us to explore this system limit, shedding much-needed new light on this phenomenon. The Stupin correlation, based on a dispersed liquid phase falling through an ascending vapor phase, was shown to give excellent prediction of experimental data on the ultimate capacity in the liquid rate range of 40-140 m(3) h(-1) m(-2) tower cross section area. However, the correlation tends to predict high at lower liquid rates. The current paper extends Stupin's previous model into the low liquid rate region. This correlation provides a useful means to check the limiting capacity that could be expected for a given column diameter and set targets for the capacities of new devices. The challenge is to develop devices that exceed this capacity limit.