The amounts of atmospheric depositions were measured by two types of bulk deposit-samplers (BDS). One of them was a commonly used dry-type BDS mounted with a conical polyethylene (PE) funnel on a PE-reservoir (d-BDS). The second BDS was a wet-type BDS which was devised to evaluate the nutrient loads deposited directly from the atmosphere onto lakes, and mounted with a cylindrical funnel filled with 0.005-M H2SO4 at 5 cm in depth (w-BDS). The water sample in the reservoirs was collected after an appropriate time interval. As the funnel sizes of d-BDSs decreased from 20 to 9 cm in diameter, the trapping efficiencies of TN and TP increased, as measurement accuracies decreased. The efficiencies measured by w-BDSs were not affected by the funnel size in the tested range from 20 to 43 cm. Based on the data continuously measured by d-BDS20 and w-BDS20 mounted each with a 20 cm funnel for one year, the atmospheric deposition of TN and TP measured by d-BDS20 were 18.1 and 0.745 kg ha(-1) y(-1), which were apparently smaller than 20.7 and 0.921 kg ha(-1) y(-1) measured by w-BDS20. It is recommended from the results that the direct nutrient-loads from atmosphere into lakes should be evaluated on the base of the more reliable unit loads measured by a w-BDS mounted with a funnel larger than around 30 cm in diameter.