Conventional K-Ar and 40 Ar/ 39 Ar analyses on whole-rock samples are reported for basaltic samples retrieved on the Central and Southern Kerguelen plateaus during Ocean Drilling Program Leg 120. Sites 747, 749, and 750 recovered basalts from the plateau basement, whereas Site 748 drilled a lava flow interbedded with sediments of probable Albian age.The freshest core basalts from the basement yielded dates falling in the 110-100 m.y. interval. Sample 120-749C-15R-3 (26-31 cm) gave conventional K-Ar, total fusion, and plateau ^Ar/^Ar ages that are closely concordant: 111.5 ± 3.2 m.y., 109.9 ± 1.2 m.y., and 109.6 ± 0.7 m.y., respectively. Sample 120-750B-15R-5 (54-60 cm), when taking into account the analytical uncertainties, yields conventional K-Ar and 40 Ar/ 39 Ar plateau ages that can be considered similar: 101.2 ± 7.5 and 118.2 ± 5 m.y., respectively. Inspection of the 39 Ar/ 40 Ar vs. 36 Ar/ 40 Ar diagram does not reveal the occurrence of an initial argon component of radiogenic composition in the two samples. Accordingly, our results suggest that the formation of the basement of the Central Kerguelen Plateau was closed at 110 m.y.. Furthermore, these results are in agreement with a K-Ar age of 114 ± 1 m.y. mentioned in the literature for a basalt dredged in the 77°E Graben.The still scant amount of data indicates that the outpourings of the Central Kerguelen Plateau correspond rather well with widespread continental magmatism in Gondwanaland that is believed to mark the incipient opening of the eastern Indian Ocean. This implies a huge head for the mantle plume at the source of these liquids. Nevertheless, on land and at sea the exact duration of magmatism remains unknown. Therefore, a catastrophic pattern similar to that currently invoked for the Deccan Traps at the end of the Cretaceous, though possible, is not yet required by present geochronologic data.