A thick sequence of felsic volcanic rocks in the Cerberean Cauldron of Victoria, Australia, is Upper Devonian as evidenced from a fossil fish fauna found in intercalated sedimentary rocks (Hills, 1958). Radiometric dating of the lavas was undertaken to provide further control as to the age of the Devonian-Carboniferous boundary.
Potassium-argon dates on biotites from a nevadite and a biotite rhyodacite range from 340 to 365 million years; from this data, we estimate a minimum value of 362 ± 6 m.y. for their true age. Rubidium-strontium dates on several of these biotites are systematically lower at 352–357 m.y. (λRb87 = 1.47 × 10−11yr−1), and this difference seems best resolved by decreasing the value selected for λRb87 to 1.43 × 10−11yr−1.
The Rb-Sr isochron for the basal rhyolite in the Cerberean Cauldron, based on whole-rock and feldspar samples, is significantly older at 367 ± 22 m.y. than the Rb–Sr biotite dates and older than an (apparent) isochron for the nevadite. Its relationship to the K–Ar date of 362 m.y. is obscured by the uncertainty in the Rb87 decay constant. We conclude that either the extrusion of the lavas occurred over an unexpectedly long period or that loss of radiogenic strontium has occurred from the biotites and that the nevadite Rb–Sr data do not correctly define an isochron. In either case, these results show that the radiometric age for the top of the Devonian should be increased by 5 per cent from 345 ± 10 m.y. (Kulp, 1961) to 362 ± 6 m.y. and possibly by more than 10 per cent, depending on the correct value for the decay constant for Rb87.
Rb–Sr data for the Barjarg Granite and the Ryans Creek Rhyolite from the nearby Mansfield-Tatong area yield an age of 369 ± 11 m.y., using λRb87 = 1.47 × 10−11yr−1. Although these rocks are thought to be the same age as the lavas of the Cerberean Cauldron, their stratigraphic position within the Devonian is not conclusively established.