Efficient and accurate measurements of uranium (U) and U-series radionuclides in earth's materials are needed to assess its environmental impact, reconstruct geochemical histories, and quantify heat production in the crust. To date, measurements of ²³⁴Th and ²³⁸U by gamma spectrometry have relied on the quantification of ²³⁴Th gamma emissions at ∼63 keV (absolute intensity = 3.7%) and the ^234mPa gamma at 1001 keV (absolute intensity = 0.84%). However, use of the 63 keV emissions can be hampered by ²³²Th interferences and self absorption, and the 1001 keV photon has a very low yield. Here we describe the effective use of the ²³⁴Th doublet gamma emission at ∼92.5 keV (total absolute intensity = 4.8%) for ²³⁴Th and ²³⁸U measurements. This doublet has been largely ignored because of its proximity to the Th K_α1 (93.3 keV) and thus its vulnerability to a Th self-fluorescence interference. We demonstrate that additions of U and ⁴⁰K to a Th ore sample do not increase the count rate at 92–93 keV above that which would be expected from the associated additions of U and ²³⁴Th. We also show that the Th self-fluorescence interference appears to be an anomaly associated only with the analysis of relatively thick (>1 mm) Th minerals, and suggest that fluorescence will not complicate the 92–93 keV region in most environmental samples. A review of decay data reveals that Th K_α1 X-rays associated with the decay of ²³⁵U and ²²⁸Ac can significantly interfere with quantification of the 92.5 keV ²³⁴Th doublet. We show that simple experimentally-derived correction factors for these X-rays can be used to accurately determine ²³⁴Th using its strongest gammas, resulting in higher count rates and smaller self-absorption corrections relative to the traditional analytical lines. Total 1σ analytical error associated with U measurements at 92.5 keV ranged from 1 to 9% and is dominated by the relative size of the ²²⁸Ac interference. Detection limits for U in environmental samples using this technique are on the order of 50 ppb.