Markedly unstable count rates were observed using a toluene-Triton (2:1, v v ) scintillant during counting of water-soluble radioactive compounds when < 5% ( v v ) water was present, because of the separation of phases. Efficiency correction in these instances could not be made by using 3H 2O as internal standard, because under the same conditions count rates with tritiated water were stable. Increasing water to ≥6% stabilized the count rates. With toluene-Triton (2:1, v v ) scintillant, the water level should preferably be maintained between either 6 and 12 or 18 and 24% for 14C- and 3H-labeled compounds for counting at 6°C or at ambient temperature (but only between 6 and 12% for 3H counting at room temperature). With a “Tritosol” ( Anal. Biochem. 63, 555 (1975) modified to contain 35 ml of ethylene glycol, 140 ml of ethanol, 250 ml of Triton X-100, 575 ml of xylene, 3 g of PPO, and ±200 mg of POPOP, water levels of up to 23% were acceptable for 14C and 3H for counting at room temperature or at 6°C. Within these limitations, with the toluene-Triton or with the modified Tritosol as scintillant, both polar and apolar radioactive compounds exhibited similar efficiencies and gave quench-correction curves, based on the external standard ratio, that were linear for both 14C and 3H-labeled compounds.