The areas of volcanic units on Venus have been measured on the 1:5000000 geological maps published by NASA/USGS. These data were used to obtain a frequency-area distribution. The cumulative frequency-area distribution of 1544 specific occurrence of units cover six orders of magnitude from the largest unit (30 × 10 6 km 2) to the smallest (20 km 2). The probability distribution function has been calculated. The medium and large volcanic units correlate well with a power-law (fractal) relation for the dependence of frequency on area with a slope of -1.83. There are fewer small units than the expected values provided by the power-law relation. Our measurements cover 21.02% of the planetary surface, 3.59% of the study area was found to be tessera terrain and is excluded from this study of volcanism. The measurements were restricted to areas where geological maps have been published. The analysis was performed on two independent areas of the planet, with a complete coverage of published maps. In both areas the largest volcanic unit covers a significant portion of the surface (58.75% and 63.64%, respectively). For the total measured volcanic units (excluding tessera), these two largest units (that could correspond to the same unit or not) cover the 61.18% and they are stratigraphically superimposed on older volcanic units which cover 3.37% of the area. The remaining area (35.45%) is occupied by younger volcanic units stratigraphically superimposed on the large volcanic unit(s). These results are based on the independent mapping of a large number of geologists with different ideas about the geodynamical evolution of Venus and different criteria for geological mapping. Despite this fact, the presence of these very large units is incompatible with the equilibrium resurfacing models, because their generation at different ages would destroy the crater randomness. Our frequency-area distribution of the mapped volcanic units supports a catastrophic resurfacing due to the emplacement of the largest unit(s) followed by a decay of volcanism. Our data for the frequency-area distribution of volcanic units provide new support for catastrophic resurfacing models. It is difficult to make our observations compatible with equilibrium, steady-state resurfacing models.