The first phase of the Uranium-Series Intercomparison Project (USIP) consisted of the analysis of six natural carbonate samples by 14 laboratories in five countries. Measured 234U/238U ratios show a higher precision than do 230Th/234U ratios or U abundances. For 234U/238U ratios the standard deviation of the mean approaches the mean counting uncertainty, whereas for 230Th/234U ratios and U abundances the standard deviations greatly exceed the mean analytical uncertainty. Much of this variation is likely explained by systematic errors in calibration, spike activity and activity ratio determinations, and spectral resolution.
On the basis of USIP results presented here, “best-estimate” 230Th/234U ages for the coral samples analyzed are as follows: Barbados II (Ventnor) terrace, 105,000 ± 5,000 yr; Barbados III (Rendezvous Hill) terrace, 118,000 +9,000−8,000 yr; Barbados (Kendal Hill) terrace, 210,000 +2,500−2,100 yr; Curaçao +6 m terrace, 124,000 +1,500−1,400 yr; Key Largo Limestone (Windley Key Quarry), 139,000 +1,900−1,400 yr. The Barbados III, Curaçao +6 m, and Key Largo samples are thought to represent the last interglacial high sea stand; thus, observed isotopic differences probably result from slightly different diagenetic modification under varying climatic conditions. The results presented here suggest that geologists should place little confidence in single, random U-series age determinations unaccompanied by stratigraphic control.